Secondly, the name after Coguveusuron in Ravenna is Durbis, clearly the river Dour at Dover, so it was not unreasonable of earlier writers to assume that these rivers – Adron to Coguveusuron – were spaced apart down the east coast of England. And yet, apart from Ravenna’s Alavna and Ptolemy’s Alaunus, both referring to the river Aln, there is no name common to the two lists, which is odd. The solution to this little mystery is as simple as it is surprising – most of the names in the Ravenna group are those of tributaries of the river Tweed. Note that there is no indication in Ravenna itself that all of the rivers listed are rivers which reach the sea under the names given in Ravenna. The title ‘East Coast Rivers?’ placed above this group of rivers in the work of Richmond and Crawford is one given by Richmond and Crawford themselves – it does not appear in the Ravenna Cosmography. The Bdora and Novitia, appearing in Ravenna before the Adron, were respectively the Forth and Tyne (in East Lothian), the Alavna was the Aln and the Coguveusuron was the Coquet. All of the other names in this group are those of rivers in the Tweed basin. The Tweed itself was the Certisnassa, this being a land-name (of Berwick-on-Tweed) transferred by the Romans to the river. The Celtic name of the river was Duabs, this appearing in Ravenna’s Duabsissis, another name for Berwick-on-Tweed (this identification is discussed in Chapter 15: Navione to Alavna (187)). One needs to remember that in names with an essa-type ending the part of the name before the essa-ending may be either a land-name or a river-name (this being explained in Chapter 4: place-names with an essa-type ending). In Certisnassa the first part of the name is a land-name, whereas in Duabsissis the first part of the name is a river-name. And the hill-letters r and s in Certisnassa correspond to the river-letters s and b in Duabs. The initial d in Duabs is the river-letter t (changed to d) applied by those who coined place-names using the hill-letter l₁. Presumably one of the names Certisnassa and Duabsissis was used of Berwick in the Flavian period and the other in the Antonine period, but it was the Celtic river-name Duabs which survived in the modern river-name Tweed (Duabs + river-letter t = Duabsit → Tuabid → Tvidi → Tweed). Note that the t of Duabsit will have been added to Duabs by those people who used the hill-letter l₂.

[NB. Detailed information as to the different river-letters and as to how they were combined to form compound river-names, together with information as to the four categories of Celtic river-names, is given in Chapter 19: the rivers of Roman Britain. Detailed information as to the different hill-letters is given in Chapter 1 and information as to how the hill-letters were combined to form compound place-names is given in Chapter 2]

Adron

Identification: the river Whiteadder, Berwickshire

This name is a compound in the river-letters t (changed to d) and r.  The compound is very common in river-names – one sees it in Derventio, in Trisantonis, in Bdora and in Vedra. There are no significant rivers between the East Lothian Tyne (the Novitia) and the Tweed (the Certisnassa), the biggest being perhaps the Ale Water and the Eye Water, so there seems no good reason to doubt that the name Adron has survived in the adder part of the river-name Whiteadder, this being a tributary which joins the Tweed just west of Berwick. The name presumably appears before that of the Tweed in Ravenna’s list because the Whiteadder joins the Tweed from the north and so the compiler of Ravenna, running his eye from north to south on a map, came across the name Adron on his map before he saw Certisnassa.

Intraum and Antrum

These names both appear to include the river-letter combination t and r. One of them probably refers to the Leader Water and the other to the Ettrick, both of which also include the river-letter combination t (changed to d in the case of the Leader) and r, though quite how the old names developed into the modern names is not clear. But both rivers were of interest to the Romans. The Leader was followed by Dere Street heading north from Newstead towards the forts in the Edinburgh area and the Romans had a fort in the Ettrick valley, namely Flavian Oakwood.

Presumably Intraum  had been a Celtic river-name of the kind having a river-suffix attached to a land-name (a place-name comprising one or more hill-letters), the In of Intraum just being the ending of the land-name (probably inion in the land-name but reduced to in in the compound river-name). The Romans shortened the river-name by deleting the significant letters of the land-name, just leaving the in ending and the river-suffix tra(um), comprising the river-letters t and r. The river-name Antrum will have had a similar history, though in this case the ending of the land-name was an (probably anion but reduced to an in the compound river-name). For a similar example of the bizarre modification of a Celtic river-name consider the development of the tribal name Novantae from a river-name of the form Abranobandion, as explained in the notes for the river Abravannus.

Tinea

Identification: the Eden Water, Borders

This river is normally taken to be the Tyne flowing past Newcastle, but, as explained above, the group of names including Tinea appears to relate to tributaries of the river Tweed, and since there is a tributary called the Eden Water this is much more likely to be Ravenna’s Tinea.

Liar

Identification: the Yarrow Water

It would appear that at some stage the initial l of this name has been taken, rightly or wrongly, to be the river-letter l, so that the proper name of the river was taken to be iar. It is entirely possible that such a name would develop into Yarrow. The Yarrow valley would be of interest to the Romans as it provided communication between Trimuntium at Newstead and the forts in Annandale, on the main western route into what is now Scotland.

Lenda

Identification: the river Till

This is a land-name transferred by the Romans to a river. It is a name like Lindum and Londinium. The river-letter corresponding to the hill-letter l is t and the river-letter corresponding to the hill-letter n is m for major rivers and l for minor rivers, including tributaries of major rivers. The tributary which joins the Witham at Lincoln is the Till, which includes both river-letters t and l. The Tweed also has a tributary called the Till, so Lenda may have been a Roman post on that river, the name being transferred by the Romans to the river. That post is likely to have been on a road from Berwick/Tweedmouth to Newstead, since such a road would have been important for bringing supplies to Newstead from a harbour at Berwick/Tweedmouth - the river Tweed itself is not likely to have been of much use for transporting supplies.

Vividin

Identification: the river Teviot

The Teviot is a major tributary of the Tweed and is crossed by Dere Street a little to the north of the fort or fortlet at Cappuck. It would therefore be surprising if the Teviot were not included in a list of tributaries of the Tweed. The initial v of Vividin is probably simply an error for t, as the original spelling was probably Tividin. The v is a modified b, this being the river-letter corresponding to the hill-letter s, and it is clear that the people who used the hill-letter s did settle in that area. This is clear since the hill-letter s is present in Certisnassa, one of the two known names for Berwick-on-Tweed, and the corresponding river-letter b is present in Duabs, the Celtic name of the river Tweed. The initial t of Tividin was apparently applied by the people who coined place-names in the hill-letter l₁, and the d will simply be a modified t, in this case the river-letter applied by those who coined inversion-type place-names in the hill-letter l₂.

Durolavi

Identification: the Lyne Water

Rivet and Smith see this name as being the Antonine Itinerary Durolevo, included by error in Ravenna’s river list. However, Durolevo was a place somewhere in Kent and at the point where Ravenna lists Durolavi it is dealing with tributaries of the river Tweed, so an error of the kind suggested by Rivet and Smith is in the highest degree unlikely. The Duro element of Durolavi is hardly likely to be the same as the Duro element employed in town-names in southeast England. It is more likely to be a compound in the river-letters t (changed to d) and r, like the Der of Derventio. The lavi element will be a place-name element in the hill-letter l₁, the hill-letter used by the Votadini. The Lyne Water was of course of interest to the Romans since they had two forts on it, Flavian Easter Happrew and Antonine Lyne.

Coguveusuron

Identification: the river Coquet, Northumberland

Rivet and Smith, following Richmond and Crawford, see this name as a fusion of two different names, and they suggest Coguve and Usuron as the two names, Coguve referring to the river Coquet in Northumberland and Usuron  referring to the town of Isurium at Aldborough. This seems an unlikely fusion, given the distance between the river and the town in the north-south direction and the fact that there would probably be other names between them on a map. Some confusion is caused by the existence of Ravenna’s place-name Coccuveda, presumably a fort on the river Coquet. The element cog or cocc is thought by scholars to refer to the colour of the silt carried by the river from the red sandstone of the Cheviot hills, but is much more likely to be a modified topographical compound. Indeed it seems quite clear that Coguveusuron is derived from the then Celtic name of the river Coquet, the absurdly long name Lacerocomaguvetusuron. This is a river-name of the kind having a river-suffix, here tusur, attached to a place-name comprising hill-letters, here Lacerocomaguve. Note that the river-letters t, s and r in the river-suffix correspond to the hill-letters l, r and m in the place-name, and they appear in precisely the same chronological sequence as the hill-letters in the place-name. Lacerocomaguve appears quite clearly to have been the Iron Age promontory fort at Brinkburn Priory, within a loop of the river Coquet. The Roman place-name Coccuveda is derived from the Lacerocomaguvet part of the long river-name. If one rewrites this as [La] c [er] oc [oma] guvet and then deletes the letters in brackets one obtains the form Cocguvet. If one then carries out the common changes g → c and t → d and adds the letter a as an ending one obtains the form Coccuveda. If one rewrites the long Celtic river-name as [Lacero] co [ma] guve [t] usuron and then deletes the letters in brackets one obtains Ravenna’s name for the river Coquet, Coguveusuron. It may be noted in passing that the promontory fort name was transferred, modified to Maromago, to the Roman fort at Learchild, a little further along the Roman road known as the Devil’s Causeway. The changes involved were the l/m interchange (cf. Lelamon → Melamoni), the deletion of ce and co and the replacement of the uve ending with a simple o.

[This page was last modified on 08 March 2021]

[Navigation tip: simply close this window to return to Chapter 19, if that is where you came from. Click on Prev below to proceed to the notes for Vedra. Click on Next below to go back to the notes for Novitia. Click here on Romano-British river-names to go to the Contents page.]