And the answer is …

First, perhaps, one should review the questions:

Some time ago, there were claims that Traiectus was Aust where, before the first Severn Bridge was built, a car ferry crossed the Severn over to Beachley, a crossing known as the ‘Old Passage’. This was also claimed to be the route by which the Roman legions crossed into Wales (but is this authenticated?). If Aust was Traiectus, it means that the manuscript of the Antonine Itinerary should have read Venta Silurum-Traiectus-Abone, rather than Venta-Abone-Traiectus. And good luck to whoever tried to make sense of the distances between these stations and Aquae Solis.

It would solve the issue of why there appeared to be no mention of the crossing of the Severn but provokes a second query regarding the necessity to go to Abone/Sea Mills at all from Aust if the destination was Silchester via Bath. By what route would the legions have reached Aust and how did travellers get from Aust to Abone? There is no conjectural Roman road between Aust and Abone. Too many problems for it to be Aust.

There was Haverfield’s half suggestion that Traiectus might be Sea Mills and Abone Bitton (Bitton, he said, being quite close to the river Avon, hence Abone), again a reversal of the two stations. At least there would be no problem with the distances between the stations, if they were not also reversed. But Haverfield moved on to propose that the reading should be ‘Abone traiectus’ with another station, X, on the road to Bath. The noun traiectus, recorded mistakenly as a separate station closer to Bath, ‘extruded’ the name of station X.

Well, ‘appen …

But if the Itinerary is accepted as it is, it still leaves the problem of why no mention is made of the crossing of the Severn, but the rest is clearer and the weight of opinion nowadays seems to identify Traiectus, not as Bitton itself, but as somewhere between Willsbridge and Bitton, including the area just across the river – Keynsham.

The evidence would be stronger if there were signs of a north-south road, crossing the Avon at Keynsham. In fact, Stratford Lane, on the northern slope of the Mendips, runs north perhaps from Charterhouse. It was a Roman road disappearing into the recently man-made Chew Valley Lake, with some signs of its contination just north of the lake. For some archaeologists, that is where the road ends, and there are no clear signs that it went further north to Keynsham.

But, intriguingly, the alignment of Stratford Lane is directly to Keynsham. If the road did continue that far, Keynsham-Traiectus would be on the southern bank of the crossing place. That appears to be the route of Margary’s 540, starting at Charterhouse on Mendip and ending on the east-west Roman road near Willsbridge-Bitton; and 541a runs northwards, starting  from near Hanham or Longwell Green. Both 540 and 541a seems to be largely conjectural, however.

Nevertheless, the verdict is … Traiectus was in or around Keynsham.


A bit of a muddle?

And finally … a look at the suggestion in the Victoria County History for Somerset published back in 1906. The sections on Romano-British Somerset, with the corrupt manuscript theory, were apparently written by the distinguished scholar FJ Haverfield. According to this, ‘Traiectus’ was never intended to be taken as a station. Not having an original manuscript in front of me, it’s hard to imagine how the text was laid out:

This is the text of Iter XIV as it appears in GA Parthey’s edition: ‘Another journey from Isca to Calleva’. Iter XIV is 103 Roman miles in length, and the first ‘station stop’, as they say, is Venta Silurum (Caerwent), which is 9 Roman miles from Isca (Caerleon). The next stop, with no mention of the crossing of the Severn, is Abone (generally assumed to be Sea Mills), 14 Roman miles from Caerwent.

As for Traiectus:

“It would at first sight seem natural to identify Traiectus with Bitton and Abone with Seamills. This solution, however, raises serious difficulties. Seamills is indeed not much more than nine miles from Bitton. But Bitton is ten or eleven, not six, Roman miles west of Bath, and no ‘station’ except Bitton exists on this part of the route.”

There seems to be a mistake here: Bitton is certainly not ‘ten or eleven miles’, be they Roman miles or English, from Bath. Even by the slightly meandering modern road (and we cannot be certain of the Roman route) it is no more than 6.5 Roman miles to the centre of Bath, which would make the Itinerary‘s distance accurate if this Traiectus were Bitton.

It’s impossible to trace the exact Roman route from Sea Mills to Bitton, but taking it to be (impossibly?) straight it would be about 10 Roman miles. That might easily stretch to the Itinerary‘s 14 miles. These, after all, are the distances which “at first sight” led to the identification of Bitton with Traiectus.

What is not at all clear is why Bitton is being identified with a ‘station’ if it wasn’t Traiectus. But let’s read on:

“The remedy most often suggested is to transpose Abone and Traiectus, making Abone the name of the village at Bitton, which is within half a mile of the Avon, and identifying Traiectus with Seamills. Perhaps it would be better to suppose that Abone is Seamills and that Traiectus was put against it in the Itinerary: the double entry then was by error extended into two lines and Traiectus extruded the name corresponding to Bitton.”

In other words, he posited that the original manuscript reading was:

Venta Silurum 9 mpm (from Isca)
Abone traiectus 14 mpm (from Venta, crossing the Severn)
[Bitton?] 9 mpm (from Abone))
Aquae Solis 6 mpm (from Bitton)

Yet in the first quotation the distances are held to be wrong for Traiectus to be at Bitton, so they must also be wrong for Bitton, under whatever name, to have been the station whose name has been ‘extruded’. For that matter, the theory that Sea Mills was Traiectus and Bitton was Abone would also work.

Although the manuscript tradition that we have does contain some obvious errors, transposing two place names does not seem to be one of the errors that occurred.

Next time: the adjudication


Much too long in the doing …

So, a résumé of the appearance of the name/word ‘Traiectus’ in the Antonine Itinerary:

1. Examples where traiectus simply refers to a crossing – the crossing of the Adriatic, the crossing of the Strait of Messina, the crossing from N. Africa to Sicily. Whether or not Traiectus was, in a few cases, also the name given to a particular place (for example, near Reggio di Calabria and the Strait of Messina (Parthey §86 & §98)? – unattested, apart from in the Itinerary itself), it can be said that in these cases the crossing itself was by boat over a substantial body of water (even the Strait of Messina is over 3km wide at its narrowest point). Could that possibly be what the Traiectus between Caerwent and Bath was – not a particular settlement but just ‘the place to cross’ something (river, strait, sea)?

2. ‘Traiectus’ in Aquitania, between Agen and Périgueux: the certainty is that the crossing concerned was over the river Dordogne, close to present-day Lalinde. [A diversion was whether or not Lalinde was the Diolindum of the Tabula Peutingeriana. For what it’s worth, intriguing though the similarity of Lalinde and Diolindum is, I would guess not. For one thing, Peutinger’s distances for Diolindum are a more accurate fit for the road to Cahors than the Itinerary‘s road to Périgueux.] This does leave ‘Traiectus’ as a possible place name, perhaps for modern Lalinde or Pontours. It is the only other possible place name (besides the Gloucestershire example) which appears in the subject form ‘Traiectus’ rather than ‘Traiecto’ or Traiectum’.

3. Utrecht as a place name derived from something like  ‘Ultra Traiectus’: Is this the only example where a modern (and historically attested) name is clearly derived from Roman Latin (i.e. not medieval) traiectus? The crossing here referred to is of the river Rhine. The name is attested in Itinerary §369: Traiecto is 17 Roman miles south of Albinianis (Alphen aan den Rijn) and 15 Roman miles north of Mannaritio (Maurik). As for Maastricht, there is no evidence of the name Traiectum ad Mosam or Traiectum Mosae in Roman times: it is not mentioned in the Itinerary. Modern ‘Maastricht’  seems to be derived from its medieval name of Mosae Traiectum. Somewhat similarly, Traiectum Rhodani was a late description of the town of Beaucaire. It is a crossing place over the river Rhône on the Via Domitia, but the Roman name for Beaucaire was apparently Ugernum; so Traiectum Rhodani seems to be a description rather than a place name: ‘the traiectum Rhodani was at Ugernum’.

Does all this cast any light at all on the ‘Traiectus’ between Bristol and Bath? It certainly suggests that the traiectus would be over a wider body of water than the Boyd at Bitton. Even the Avon is a stream compared with the Dordogne which, at its narrowest, is nearly three times wider (150m) than the Avon at its widest at this point. The Rhône at Beaucaire is at least 200m (the Île du Comté divides the river in two at this point, otherwise the crossing to Tarascon would be 400m wide). The Meuse at Maastricht is 150m wide. The Rhine at Utrecht is over 100m wide. [But see the comments below on the Avon floodplain].

Two possibilities

Neither of the two competing identifications for the Gloucestershire Traiectus is new. There is the ‘corrupt manuscript’ theory which would have the traiectus as being across the Severn estuary. That would answer the puzzle as to why the Itinerary makes no mention of a crossing from (modern) Wales to (modern) England. Such a ‘traiectus’ would be similar to other examples in the Itinerary – just ‘a crossing’ rather than a station. The theory could merit closer investigation to see if this amount of corruption could be plausibly explained.

The second possibility is that the crossing was over the river Avon. But if the Antonine iter Sea Mills > Traiectus > Bath was following the known Roman road north of the river, there would be no need to cross the river at all; nor has a Roman settlement of any importance been found here north of the river. Nevertheless, and there’s nothing new here, any crossing place for a north-south road down to the Mendips looks most likely to have been at Keynsham [NB in Somerset, not Gloucestershire – Ed]: there are at least three areas where Roman remains have been found: i. Durley Hill to the west where the so-called ‘palace’ was discovered ii. beneath the Somerdale factory, and possibly iii. further east close to Wellsway school. Whatever the purpose of the ‘palace’, it was a large enough building to be part of extensive Roman activity in the area. The Somerdale site lays claim to being a small town and already is Traiectus in some sources. It even boasts a Trajectus Way (as well as roads named after Hadrian, Claudius, Augustus, Titus and Tiberius).

Preparatory investigation prior to development in East Keynsham drew a claim that traces of a north-south road on the site east of Wellsway school had been identified. On the developers’ plans the direction  of the road points towards Bitton where, it should be said, there are few Roman remains to excite interest.

But the puzzle here is why does the Itinerary‘s journey [i.e. Abone mpm xiiii, Traiectus mpm viiii, Aquae Solis mpm vi] mark Traiectus, given that the known Roman road, marked on the OS maps between Hanham and Oldland, and between Bitton and Swineford, doesn’t cross the Avon? Was there a mansio at Traiectus/Keynsham where travellers could stay by making a small detour over the river? How likely is that, with a mere 6 Roman miles either to or from the more important  town of Aquae Sulis?

One further thought: the Roman road is a little distance from the river nowadays, but was it in Roman times? Was part of the Avon’s flood plain permanently below water? Interesting thought, but the river level would seem to have been at an impossibly high level for that to be the case. But, if it was (yes, highly unlikely, though the area might have been marshy), the road would have run just alongside the river bank; and Willsbridge [vii mpm from Aquae Solis] would have been the closest crossing point to the extensive settlement known to exist underneath the Somerdale site on Keynsham Hams – which would likewise have been near the marsh or water’s edge; as would the Durley Hill ‘palace’. So, for fun:
In fact, maps of the possible sea/river levels suggest that the area south of the river would have been first to flood, especially Keynsham Hams (though not the eastern side where Somerdale was built, nor the Durley Hill ‘palace’ site). It would need seemingly impossibly high levels before the northern floodplain was covered completely. So, Not This Way. But it leaves the possibility that there was a pause at Willsbridge and a quick traiectus across the river to the Keynsham Hams settlement (Somerdale) for a bit of refreshment ….

Next time: a short look at the suggestions regarding the crossing of the Severn estuary, published in the Victorian County History, Somerset vol I, Romano-British Somerset Part 3, ‘THE ROAD FROM BATH WESTWARDS TO THE SEVERN CROSSING AND SOUTH WALES’.


A bit of messing about …

The Tabula Peutingeriana indicates that, from Agen, the Roman road went to Excisum, from Excisum to Diolindum and from Diolindum to Divona/Cahors. All four places seem to be placed in a line as if situated on the same river, which they aren’t. Toulouse is placed somewhere between Bordeaux and Agen, whereas from Agen it is about the same distance as Bordeaux – but in the opposite direction – Agen equidistant from the two.

It’s clear, as described before, that the section of the map covering modern Brittany should be swivelled clockwise about 90º. Now, take the section showing Tolosa, Aginnum, Excisum and Diolindum, leading to Divona (Cahors):

If this is twisted round anti-clockwise 135º, we have Toulouse in its correct position in relation to Agen, while Excisum and Diolindum lead to Vesunna (Périgueux) rather than Cahors – just as the Itinerary has Agen, Excisum, Traiectus, Vesunna.

Absurd, though, to manœuvre a small pencil of land into the desired position and leave its surroundings untouched. What kind of mistake could explain that?

But … Just One More Thing.

Looking at the figures after the placenames: XIII indicates the distance from Aginnum to Excisum, exactly the same as the distance given in the Itinerary. No mystery there because there’s no argument that Excisum was Eysses so they are recording the identical journey. Something more of a coincidence is that Excisum to Diolindum is XXI (enhanced this is quite clear and accepted by the authorities). This is exactly the same as Excisum to Traiectus in the Itinerary. Then Diolindum to Cahors is XXIIII. Annoyingly, Traiectus to Vesunna/Périgueux is XVIIII. Doubly annoying because XXIIII could so easily have been misread as XVIIII – or vice versa.¹ But it’s weak evidence that is based on, ‘Yes, but perhaps someone made a mistake there’.

But my money is still on Diolindum and Traiectus being at roughly the same place on the Dordogne.

Probably time now to move on from Aquitaine and scour the Itinerary for other examples of the placename Traiectus.

¹ In fact, if the Itinerary’s XVIIII was a misreading for XXIIII, the distance between Traiectus/Lalinde and Vesunna would be much more accurate (‘as the crow flies’  26 Roman miles rather than 18) .

Traiectum, Traiectus, traiectus (3c)

This is a wander round the Tabula Peutingeriana, focusing on the section covering present-day France with the roads and towns of Aquitania – that is, the area surrounding the Traiectus of the Antonine Itinerary. The immediate impression is that there are huge inaccuracies in the location of known towns and cities, and the orientation needs constant adjustment.

The wide inlet resembling a river estuary is marked Sinus Aquitanicus – which is the Bay of Biscay. From the towns that we can identify, it looks as if the upper ‘jaw’ should be prised up 90 degrees and either stretched out east-west or squashed down north-south. Then the placing of some of the towns, at least, would make a bit more sense. Most notably, Pretorium Agrippe (Valkenburg) and Lugdunum (Leyden) in the Netherlands would then be over to the east rather than in north west ‘Brittany’. Portunamnetu (Nantes) would lie due south of Condate (Rennes), Fanomartis (Corseul) would be north west of Rennes, rather than just south of west, and Darioritum (Vannes) would move from south east of Rennes to, correctly, south west.

The lower ‘jaw’ cannot so easily be explained. As it stands, the map depicts Vesonna (Périgueux) as roughly west-north-west of Agen, whereas it ought to be almost due north; and Tolosa (Toulouse) is almost due west of Agen when it should be south-east. This leads one to examine why the road Aginnum-Excisum-Diolindum goes in an easterly direction and leads to Divona (Cahors) when, especially if one wanted to identify Diolindum with Lalinde, the road should go north and lead to Vesunna.

Again, there appear to be two rivers running east to west. Agen and Tolouse both lie on the Garonne, whereas Eysses and Cahors are on the Lot; yet all four towns are depicted in an east-west line on or close to the same river. Is the more northerly river the Dordogne? If so, Périgueux is lying on the southern bank, but Périgueux is not on the Dordogne at all, but further north on the river Isle.

Whether or not Traiectus is Diolindum at Lalinde/Pontours, as Dr Chaume thought¹ is not really relevant. Lalinde is roughly where the ‘traiectus’ was, by whatever other name the place was known. The name ‘Diolindum’ is merely a curiosity of Peutinger. But, in the search for truth: Did Aginnum-Excisum-Diolindum go north to Vesunna, as the Itinerary would have it (Peutinger, contradicting the Itinerary, has no road at all linking Agen with Périgueux); or did it go east to Cahors as Peutinger suggests? Answers on a postcard, please.The first case explains why some consider Lalinde, or the closely surrounding area, to be the site of Diolindum, the second why the towns of Duravel² or even Belvès³ are suggested.

For the record, my tossed coin falls in favour of Lalinde … It is hard to make a strong case based on Peutinger’s geography.

¹ Review, M. Chaume, ‘Le “trajectus” de la Dordogne’, Bulletin of the Société historique et archéologique du Périgord, Périgueux, 1908.

² Conjecture in ‘Les voies romaines en Gaule’, Revue archéologique, n.s. 4º year, vol 8,  p 74. Duravel is a site rich in Roman archaeological remains. But one of many.

³ P. Barrière, ‘A propos des voies antiques des Cadurques. Organisation et circulation’, Revue des Etudes anciennes, 1952, pp. 102-108.

Traiectum, Traiectus, traiectus (3b)

So, the most potentially useful measurements to locate the place between Agen and Périgueux called Traiectus are sadly inaccurate – or are they? Another way of viewing the measurements in the Antonine Itinerary, rather than calculating Roman miles x 1.48 to give the number of kilometers, is to look at the number of Roman miles between the stations as percentages of the entire distance.

If Agen to Périgueux is 52 Roman miles, Agen to Excisum, at 13 miles, would be exactly 25% of the total journey; Périgueux to Traiectus, at 18 Roman miles, is 34.62% of the journey; and Excisum to Traiectus, at 21 Roman miles, is 40.38% of the journey. Measurement of the ‘as the crow flies’ distances ¹ using these percentages had astonishingly accurate results:

The journey Aginnum to Excisum, using the percentage method, ends about one Roman mile north of the archaelogical remains of Excisum in Villeneuve-sur-Lot

In other words, the Itinerary reckoned the distance as 27.3km and it was actually 26.1, a percentage error of about 4%. The figure for Traiectus is even more interesting. The Itinerary‘s percentage would place it at 37.8km south of Périgueux.

Mouleydiers, Couze and Pontours, as well as Lalinde, have all been considered possible sites of the crossing place, the ‘traiectus’. Mouleydiers and Lalinde are almost exactly the distance from Périgueux which the Itinerary specified.

However, the only point relevant to the current study is confirmation that the ‘traiectus’ concerned was across the River Dordogne; though a possibly useful further clue is the discovery at the turn of the last century by the archaeologist and historian, Dr Maurice Chaume, of what he claimed to be the signs of a Gallo-Roman ford,² just down-river from Pontours, visible when the water level was low: a traiectus could be effected by ford, ferry or even a bridge.

The main points to take away are that this ‘Iter’ necessitated the crossing of a river and that it appears in the Itinerary as Traiectus.

A related study is the evidence of the Tabula Peutingeriana, though, if anything, it tends to muddy the already turbid waters. Whether the river crossing was at the place which Peutinger calls Diolindum, or whether Diolindum was somewhere else is not of great importance, but I may as well look at it anyway while I’m in the area. Traiectum, Traiectus, traiectus (3c) follows.

¹ ‘As the crow flies’ will not be exact in terms of Roman miles/kilometers travelled because the length of the actual journey will reflect the twists and turns of the road. If we  presume that the four stations are roughly in line with each other (which they are), the deviations to one side or the other will be to some extent self-compensating. I think.

² Archaeological remains of the ford are apparently preserved in the museum at Périgueux.

Traiectum, Traiectus, traiectus (3a)

The example in the Itinerary which most closely resembles the occurrence of Traiectus, between Bristol and Bath, is a place in Gaul. Item de Aquitania in Gallias is in sections 461- 462 of the Parthey-Pinder edition. The first journey here is A Burdigala Argantomago, that is from Bordeaux to Argenton-sur-Creuse, in Berry – not to be confused with Argentan in Normandy.

The first thing to be noted is that this is not ‘how to get to Argenton from Bordeaux’. It seems more likely to represent a ‘business trip’ which involved visits to Bordeaux, Agen, Périgueux and Limoges (since a journey from Bordeaux to Argenton would not include Agen). That would explain why the journey began travelling south east, then turned in a northerly direction for the rest of the way.

At the start, the journey from Bordeaux to Agen, two intermediate stations  – Sirione and Ussubium – are tentatively identified; and ‘Fines’ (a word frequently occurring throughout the Itinerary) between Ussubium and Agen may denote a border between the tribal lands of the Bituriges Vivisci, centred on Burdigala/Bordeaux, and the Nitobriges who were established at Aginnum/Agen¹.

If Sirione is Cérons, and Le Mas d’Agenais Ussubium, as is conjectured, the distances in the Itinerary from Bordeaux to Agen are quite accurate, with a border between the Bituriges Vivisci and the Nitobriges somewhere around the Lot river. However, the story is very different for the next section of the journey, from Agen to Périgueux, Aginnum to Vesunna.

There are just four stations – Aginnum, Excisum, Traiectus and Vesunna. The beginning and the end are clear: Agen and Périgueux; and we know the Itinerary‘s measured distances between the four stations which should confirm the exact locations of the two intermediate stations, Excisum and Traiectus.

But the distances given are as follows:

Aginnum to Excisum, 13 Roman miles; Excisum to Traiectus, 21 Roman miles; Traiectus to Vesunna, 18 Roman miles, making a total of 52 Roman miles between Aginnum and Vesunna, approximately 77 kilometers. But, as the crow flies, Périgueux is about 110 kilometers from Agen, about 74 Roman miles, so a road journey could not be less than 74 Roman miles, and must be appreciably more. The Itinerary‘s 52 Roman miles is some way out.

Excisum has certainly been identified as the Roman site of Eysses within Villeneuve-sur-Lot. This is some 16 Roman miles north of Agen, rather than 13, but that is not too inaccurate. Measuring 21 Roman miles north from Excisum and 18 Roman miles south from Vesunna should pinpoint Traiectus – but it doesn’t: there is a gap of 18 Roman miles between the two points, which is not surprising in view of the discrepancy between the Itinerary‘s measurements here, and ‘as the crow flies’ which is appreciably longer.

However, Traiectus, as we know, is a crossing place, and approximately mid way along that 18 Roman mile gap runs the Dordogne river and the place suggested by some (and contested by others) for Traiectus. The name of the modern settlement is Lalinde – particularly interesting because on the Tabula Peutingeriana there is a place close to Aginnum and Excisum, otherwise unidentified, called Diolindum.

The debate among French archaeologists and historians seems to have revolved around the question of whether Traiectus was at Lalinde, or a little up river at Pontours, or down river at Couze or Mouleydier. One could draw a small ellipse round an area covering Mouleydier, Lalinde and Pontours and agree that it was somewhere inside it, on the Dordogne river. Drawing a straight line between Agen and Périgueux, one would expect nothing else.

A few further points may be considered, for example: the distances given in the Itinerary, the interpretation of the evidence of the Tabula Peutingeriana; and any relevance to ‘our’ Traiectus, between Bristol and Bath. Traiectum, Traiectus, traiectus (3b) follows.

¹ Fines also occurs between Vesunna (Périgueux) and Augustoritum (Limoges), the respective tribal centres of the Petrocorii and the Lemavices.  The two towns eventually dropped their Roman names and adopted their tribal names.