Ptolemy’s Celtic tribes: Part 1


[Note:     Most of the text below was drafted some time ago and concerns the tribes in Britain listed in the Geography of Ptolemy. The text includes several references to the Deceangli of North Wales, which is slightly embarrassing for the writer as the Deceangli are not in fact included in Ptolemy’s list. However, there seems little sense in deleting those references now and so, in the hope that the references to the Deceangli will be of interest to some readers, the writer will leave the text as it is. The other tribes referred to below are indeed included in Ptolemy’s list.      D.G.   29.03.2020]



1     Introduction

1.1     Other pages on this website explain how topographical Celtic place-names include one or more of the letters r, m, n, l and s, each used to represent ‘hill‘, the letter b or t representing ‘high’, c or g representing ‘steep’, d meaning ‘summit’ and v meaning ‘slope’. These basic building-blocks of Celtic place-names are explained in Chapter 1 of the Home menu. Many place-names are compound names comprising two or more of the hill-letters. The structure of these names and the categories into which they fall are explained in Chapter 2.  Other chapters of the Home menu discuss the Romano-British place-names in different regions of Britain. The present page studies the correlation of names in particular hill-letters and the Celtic tribes referred to in the Geography of Ptolemy. The place-names used in the study are taken from Ptolemy’s Geography, the Ravenna Cosmography, the Antonine Itinerary and the Notitia Dignitatum.

1.2     The different hill-letters appear to have been used by different waves of settlers arriving in Britain from the Continent at different times. It would appear that the normal chronological order of the hill-letters was n, s, m, r, l. Amongst the earliest place-name elements are bindo, seen in names such as Bindogladia, Vindomora and Vindogara (in the last two examples the initial B was changed to V, apparently by the Romans rather than by the Celts) and cindo, seen in Cindocellum. These name-elements, bindo (meaning ‘high hill summit’) and cindo (meaning ‘steep hill summit’) are never qualified by an earlier element but do themselves qualify later elements, as in all of the examples referred to above. But there is a complication – each of the hill-letters n and l appears to have been used by two different groups of settlers who arrived in Britain at different times. So far as the writer has been able to ascertain the normal chronological order of the hill-letters used in Celtic place-names in Britain was thus n1, s, m, r, l1, n2, l2, where n1 to n2 were used by settlers who arrived before about 150BC and l2 by settlers who may have arrived after that date. The above chronological order of the hill-letters appears to apply to the island of Britain as a whole, but there is some regional variation. Where the order of the hill-letters in a particular old-style place-name is not consistent with the order indicated above this may in many cases be explained as a consequence of inter-tribal warfare leading to some tribes gaining, and others losing, territory.

1.3     All of the Celtic place-names used in this study are assumed to have existed at the date of the Roman invasion, even where the ancient source in which we find some of those names was produced at a considerably later date. Where a place-name element comprises a hill-letter accompanied by a qualifying b or t meaning ‘high’ or c or g meaning ‘steep’, or by a d meaning ‘summit’ or a v meaning ‘slope’, it is normally possible to identify that element as being either old-style (qualifier before the generic hill-letter or hill-letter followed by a d or v) or inversion-type (generic hill-letter before the qualifier or hill-letter after a d or v). Where this is not possible and the order of the hill-letters in a compound name is consistent with the normal order indicated above, then the name is assumed to be old-style.

1.4     It is possible to identify regions within Britain where names with a particular hill-letter as the latest hill-letter clearly outnumber names in which some other hill-letter is the latest hill-letter. Some such regions coincide with areas said by Ptolemy to be occupied by particular tribes. In such cases it is reasonable to conclude that a particular tribe used a particular hill-letter. In other areas there is a patchwork of names with different hill-letters as the latest hill-letter so that it is not possible to deduce whether any particular tribe was dominant and so difficult to see any correlation with the tribes mentioned by Ptolemy. But we will start with those regions in which a particular hill-letter is clearly the most common latest hill-letter in place-names.  


2     The Iceni

2.1     We will start with East Anglia, the Celtic topographical place-names in this region being shown on the map below.










The earliest name in this region appears to be Ptolemy’s Sidumanis. Ptolemy gives this as a river-name, but it is in fact a land-name (a place-name comprising one or more hill-letters) transferred by the Romans to a river. The old-style Sid element refers to a location on the summit of a hill and the place concerned appears to have been in the Halesworth/Holton area, on the river Blyth. (That the s-people did settle in this region is borne out by the river-name Deben, the b in this name being the river-letter corresponding to the hill-letter s). But this settlement was then taken over by people who used the hill-letter m. These were the same people who founded Camborico, apparently at Thetford, Combredovio (original form of Combretovio) in the Baylham House area and Camuloduno at Colchester. The Colchester area was later taken over by people who used the hill-letter l, these people also being active around Romford (Durolito), London (Londinium) and St Albans (Virolamium). But the northern part of East Anglia was taken over by people who used the hill-letter r. We see this in Camborico and Combredovio. But we see this hill-letter also in Branodunum at Brancaster. The dunum element in this name appears to have been applied by the Romans themselves, presumably just with the meaning ‘fort’, but the Brano element is earlier, the initial Br being an old-style element meaning ‘high hill’. And there is one more name – Ptolemy’s Gariennus river. This was probably earlier Gariserna, where serna is just a river-element (comprising the river-letter s  and r) used as a suffix to a place-name, in this case Gari. The Gar of Gari is an old-style element meaning ‘steep hill’ and so the most likely location of the place called Gari was Norwich, this being the first place where one sees steep slopes when one sails up the river Yare from the coast. Thus, leaving on one side both Venta and Sidumanis – the latter name may have become fossilized in this form or the n might actually have been an r originally – all of the place-names available to us indicate that at the date of the Roman invasion the northern part of East Anglia was controlled by people who used the hill-letter r. These people must have been Ptolemy’s Iceni and they must have been in control of that region for some considerable time since all of the names are old-style names in the hill-letter r, though Camborico might possibly include the transitional element boric meaning ‘high hill steep’. Ptolemy’s Venta of the Iceni is a later name (it is inversion-type) applied to the new town built by the Romans at Caistor St. Edmund, presumably replacing Gari  at nearby Norwich. The name Venta, in the hill-letter n, is not only alien to this region but also is inappropriate for the site – it means ‘slope of hill high’. Presumably the Romans looked around for a suitable name for a new tribal capital and simply copied Venta from Venta Velgarom (Belgarum in Iter VII of the Antonine Itinerary) at Winchester.

2.2   It may be noted in passing that the tribal name Iceni, like most British Celtic tribal names, is based on a topographical place-name, presumably the name of the residence of the tribal chieftain. We have to start with the form ECENI since this appears in coin inscriptions (see the entry for Iceni  in the Alphabetical List of Rivet and Smith, 1979). And the topographical place-name should include the hill-letter r, the hill-letter used by the Iceni. The topographical place-name may thus have been Recenion. But this form is unlikely to be complete. We have seen above that there must have been a place on the river Yare, apparently at Norwich, with a name including the element Gar (meaning ‘steep hill’), so the first century BC name of the tribal centre was probably Garirecenion, this being a name of the kind having an old-style element in one hill-letter followed by an inversion-type element in that same hill-letter, the two elements having exactly the same meaning (cf. Banvaventa and Bannatia). For some reason the tribal chief or chiefs of the Iceni used the eceni part of the place-name in their coin inscriptions. The people who were ruled from Garirecenion would be called the Garireceni. Deletion of Gar and re then yields the tribal name Iceni, but who effected this modification of the tribal name, whether Celt or Roman, is not clear.


3     The Silures

3.1     Going over now to southeast Wales and the adjoining part of England we see a number of place-names in the hill-letter l1Gleva at Gloucester, Alvinundo (apparently the Celtic form of Albinumno) at Welshbury Camp, Nemedonbala (apparently the original form of Metambala) at Lydney, Velesedio (apparently the original form of Blestio) at Little Doward Camp near Monmouth, Bulgaeum/Buldaeum (possible original Celtic forms of Bullaeum) at Usk, and possibly Luba (apparently the original form of Iupania) at Chepstow. Alvinundo  was lost to n-people, but the other places mentioned were in the hands of the l1people at the date of the Roman invasion.






3.2     But note that at the date of the Roman invasion the territory of the l-people was very confined, for to the northwest Bannio (Abergavenny) was in the hands of n-people, Ariconio (Weston under Penyard) in the hands of r-people and Macatonion (Dymock) in the hands of m-people.  The l-people, confined to a small area on the northern side of the Severn estuary, must then have been Ptolemy’s Silures.

3.3     As to Isca Augusta at Caerleon there is in fact a hill-fort just northwest of the Roman fortress, in Lodge Wood, and that hill-fort does indeed stand at the top of a steep hill. Isca means ‘hill steep’ so it is possible that the Romans transferred the name of the hill-fort to their new fortress. It seems much more likely, however, that the Legio II Augusta, which had earlier been based at Isca at Exeter, simply applied the name of its old base to its new fortress at Caerleon, this then being known as Isca Augusta. Turning now to Venta Silurum at Caerwent, Venta is an inversion-type name meaning ‘slope of hill high’ and is clearly inappropriate for the site. It seems likely that the Romans, after defeating the Silures, built a new administrative centre for the tribe at Caerwent and gave it a name which they thought appropriate for a tribal centre – they simply copied Venta from Venta Velgarom at Winchester, just as they did in the case of Venta Icenorum at Caistor St. Edmund. It follows from the above that Isca and Venta, though Celtic topographical names, were applied to Caerleon and Caerwent by the Romans themselves and not by the Celts. These two names do not therefore affect what is written above regarding the Silures being a people who used the hill-letter l1.

3.4   The tribal name Silures appears quite clearly to be based on a topographical place-name of the form Silunion, the name indicating that the Silures had taken the place over from a people who used the hill-letter s. No suggestion can be made here as to the identification of Silunion. The only name discussed above which uses the hill-letters s and l in that chronological order is Velesedio (suggested as the original form of Blestio at Monmouth), though this is a later name since the Vel element is inversion-type.


4     The Deceangli

4.1     Moving up now to North Wales, we can recognize the tribe generally known as the Deceangli, presumably the Ceangi mentioned by Tacitus, as a people who used the hill-letter s. We see this in Seguntio at Caernarfon and in Stuccia. This latter name is given as a river-name by Ptolemy but it is in fact a land-name. The Celtic name was probably Stuccla, this being a compound comprising the inversion-type element stuc (meaning ‘hill high steep’) qualified by the old-style element cla (meaning ‘steep hill’). We thus have two place-names in which s is the latest hill-letter in the compound. But the river-names in that region also indicate that that region was inhabited by people using the hill-letter s, since the river-letter b corresponding to the hill-letter s appears in Toesobis (the Dee), Tuerobis (the Conwy), Tobius (the Dovey) and Ratostabius (the Ystwyth).  




Note that it is not always clear whether bis/bius at the end of a river-name is just a name-ending or is in fact a name-element including the river-letter b. But the above-mentioned river-names ending with bis/bius are all close together in Ptolemy with no river-name of a different form amongst them - Stuccia is a land-name transferred to a river – and there are no other river-names ending with bis/bius anywhere else in Wales or up the west coast of England. River-names with a bis/bius ending thus seem to be peculiar to one particular region in northwest and north Wales. It thus seems safe to conclude that the b in those names is indeed the river-letter b corresponding to the hill-letter s. We can thus see that at one time the s-people – the Deceangli – controlled a coastal strip from Aberystwyth up to Caernarfon and along to the Dee, though it is not clear how much of that land they still held at the date of the Roman invasion. It seems clear that they had lost the area around Aberystwyth since Ratostabius is a river-name of the kind having a river-suffix, here stabius, attached to a place-name, here Rat, and Rat is an inversion-type element in the hill-letter r, the element meaning ‘hill high’. Note also that the river-letter s corresponding to the hill-letter r is also applied in an inversion-type manner, i.e. it comes at the beginning of stabius. We can thus conclude that the area around Aberystwyth had been controlled by s-people but later fell into the hands of r-people.  

4.2     Note that the name Canubio applied to the Roman fort at Caerhun is not inconsistent with the above. Canubio is an alternative name for the Tuerobis (the modern Conwy) and was transferred by the Romans to their fort at Caerhun. Canubio is a shortened form of the Celtic river-name Descecanglubena  (place-name Descecanglion + river-suffix ubena). Descecanglion  will have been the name of one of the hill-forts along the course of the Conwy. The can element uses the hill-letter n1 and is the same name as the earliest element, Conion, of Elconionemedo at Launceston. There are other examples where a river has two names, one a pure river-name, i.e. comprising only river-letters, and the other a name with a river prefix or suffix added to a place-name. For example the Velox, the modern Brit, comprises the river-letter b changed to a v and used as a prefix to the place-name Lox (=Locs). But the pure river-name also existed since it has survived in Brit, where the b and t are the river-letters corresponding to the hill-letters s and l of Locs.

4.3     It may be noted in passing, with regard to the river-name Toesobis, that the T changed to D, the s was omitted (a fairly common change in Romano-British names) and the b changed to v to yield the river-name Deva. The Romans then simply transferred the river-name to the fortress which they built at Chester. The place-name Deva thus has no connection with any goddess, as is commonly supposed.

4.4    The tribal name Deceangli, assuming this form is substantially correct, appears clearly to be based on a topographical place-name somewhat of the form Descecanglion, comprising the inversion-type element Desc meaning ‘summit of hill steep’ and two old-style elements can and gl both meaning ‘steep hill’. The writer has put an s in the first element because the Deceangli used the hill-letter s. One sees the can element in the river-name Canubio discussed above in paragraph 4.2. The people ruled from Descecanglion will have been called the Descecangli and this form, with loss or omission of the s and the second c, yields the tribal name Deceangli.


5     The Damnoni and Dumnoni

5.1     In central Scotland we see evidence that the area was populated by a variety of different people over time. Thus we see l1-names (i.e. names in which l1 is the latest hill-letter) in Cindocellum at Dumbarton/Dumbuck, and possibly in Subredobiladon (probable original form of Subdobiadon) at Kirkintilloch, an r-name in Cibra at Bearsden and s-names in Pecsa (original form of Pexa) at Mumrills (though originally the Iron-Age hillfort on Cockleroy hill) and Begsesse (probable original form of Begesse) at Seabegs (though originally the Iron-Age hillfort on Bowden Hill). It is not clear whether the l in Litana and Lodone is l1 or l2. But the initial L in Lindinonaco (probable original form of Lintinomago) at Yett’s o’ Muckhart on the river Devon is apparently l1 since the corresponding river-letter t appears, changed to d, at the beginning of the river-name Devon. However, overwhelmingly the commonest latest hill-letter in place-names in central Scotland is m – one sees it in the old-style names Cerma (Dalginross), Cermium (on the Water of May) and Nedionemedon (probable original form of Medionemeto at Bar Hill) and in the inversion-type names Marcotacson (original form of Marcotaxon, on the river Ruthven in Strathearn), Demerosessa (original form of Demerosesa at Drumquhassle), Mugulesde (apparently the original form of Ugueste at Stirling) and Melantvrum (apparently the original form of Memanturum at Dunfermline). In addition it is likely that Maulion at Malling originally had a form such as Matulion or Maculion, so that m is also the latest hill-letter in this name. It seems quite clear, then, that these people, who used the hill-letter m, were the Damnoni of Ptolemy.




5.2     But note that there is not one single old-style place-name in Scotland where m is the earliest hill-letter of the name (at least not in the place-names known to us from the ancient sources). In other words the people who used the hill-letter m appear not to have settled in Scotland when they came to Britain. But they were in control of central Scotland at the date of the Roman invasion. They must therefore have moved to central Scotland from some other region of Britain. And it appears reasonably clear where they came from, since Ptolemy refers to a tribe called the Dumnoni in southwestern Britain.







If one looks at the names in that region there is a concentration of old-style names in which m is the earliest hill-letter – Moriduno (Hembury), Omirededertis (probable original form of Omiretedertis at Ham Hill), Morionio (Norton Fitzwarren), Milidunum (South Molton) and presumably Masona (unknown). Lelamon (Celtic form of Ravenna’s Melamoni) is another – it appears to have been close to either the West Dart or the East Dart river between Tavistock and Exeter. In addition the m-people took over Launceston from the n1-people, since m is the second earliest hill-letter in the name Elconionemedo, after n1. The m-people thus did settle in that region when they came to Britain. But all of those places were lost to other people. By the date of the Roman invasion the only place in that region which appears clearly to have been in the hands of the m-people was Mestevia at Tiverton. It thus seems reasonable to conclude that some of the Dumnoni moved north to what is now central Scotland, and they did so at a time when they were still coining old-style place-names – Cerma, Cermium, Nedionemedon. Later they switched to using inversion-type names – Marcotacson, Melantvrum, Mugulesde, Demerosessa and probably Matulion/Maculion at Malling, perhaps even Matagea (Ravenna’s Tagea) at Strageath if the form of this name is correct. Mestevia at Tiverton is a name of the same kind. It is thus clear that Ptolemy’s reference to the Dumnoni  is earlier in date than his reference to the Damnoni, since these were apparently one and the same people and they presumably did not control Somerset/Devon and central Scotland at the same time. 

5.3   The tribal name Dumnoni/Damnoni seems quite clearly to be based on a topographical place-name, presumably the name of the tribal centre, which was presumably the name of the place where the tribal chieftain lived at the time the tribal name was coined. The topographical place-name will have been somewhat of the form Dumnonion or Damnonion, where Dum/Dam means 'summit of hill'. The vowel in the first element is of no importance – indeed it appears as e in Demerosesa (Drumquhassle). The first element of Dumnonion/ Damnonion is qualified by the hill-letter n, which may be n1 or n2. Note that the chronologically later hill-letter in the place-name is m, the hill-letter used by the Dumnoni/Damnoni. The writer can make no suggestion as to the identification of Dumnonion, but it was presumably a hillfort in southwest England. And of course the people of Dumnonion were called the Dumnoni.

5.4   But a word of caution. It seems clear that the hill-letter m is the latest element in Maridunum at Merlin’s Hill/Carmarthen since Ptolemy assigns Maridunum to the Demetae tribe and the Demetae used the hill-letter m (‘Ptolemy’s Celtic tribes: Part 2’, 16). Similarly it may be that the hill-letter m is the latest element in Moriduno at Hembury since Hembury is fairly near Exeter, apparently the administrative centre of the Dumnoni, and the Dumnoni also used the hill-letter m. It may likewise be the case that m is the latest hill-letter in Milidunum (South Molton) and Morionio (Norton Fitzwarren, west of Taunton) and possibly in Masona (as yet unidentified, but apparently somewhere east of Ilfracombe). If this is correct then the Dumnoni clearly occupied a substantial area of the southwest of England after 120BC, so there seems no good reason to assume that some Dumnoni  migrated to what is now central Scotland prior to 130BC. Moreover, it would appear that at one time people who used the hill-letter m occupied the whole area along the north side of the Solway Firth all the way along to the Rhinns of Galloway, but were displaced from most of that area (except for the extreme southwest corner of Scotland) by the Selgovae/Seglovae at some date prior to 130BC (Home/Chapter 25, 3). It is thus possible that some of the displaced m-people moved north to central Scotland. It is thus probably safer to conclude that the tribal name Damnoni of a tribe in central Scotland developed wholly independently of the tribal name Dumnoni in southwest England. The place-name Damnonion (the basis for the tribal name Damnoni) will have been an Iron Age hillfort or settlement which had been occupied by an n-people, presumably the Vacomagi, Venicones or Epidi (‘Ptolemy’s Celtic tribes: Part 2’, 18 and 19), but was taken over by the m-people who came to be called the Damnoni.


6     The Venicones

Moving on now, the Celtic/Romano-British place-names south of the Firth of Forth are shown on the map below.

Olcaclavis at Elginhaugh comprises the inversion-type element Olc in the hill-letter l2 qualified by the old-style element cla in the hill-letter l1, this name indicating that that area was at one time occupied by people who used the hill-letter l1 but was taken over by people who used the hill-letter l2. In addition the river Tyne in East Lothian appears as Novitia in Ravenna, the river-letter t indicating that the area around that river was also occupied by people who used the hill-letter l2. Further west the hill-letter l in Mugulesde (Ravenna’s Ugueste) at Stirling is l2, corresponding to the river-letter t, changed to d, appearing in Bdora/Boderia, the then name of the river Forth. And the late pre-Roman name of Doune appears to have been Lecilodanum (Ravenna’s Leviodanum), the inversion-type element Lec (meaning ‘hill steep’) using l2. The l in the middle of the name is l1. The river-letters corresponding to l1 and the n of danum are t and m (for major rivers), this yielding a river-name somewhat of the form tem. As usual the people who used the hill-letter l2 placed their river-letter t at the end of the existing river-name, so that the river-name became temit. This form, with omission of intervocalic m (omission of intervocalic consonants in a name is a fairly common change) and change of final t to th, yielded the modern river-name Teith.  Note that the change of c in the Celtic name to v in Ravenna’s form of the name is seen also in RacatoniumRavatonium and Lacobrinda Lavobrinta. The other Celtic topographical names in that part of Britain are Lecidensca (Ravenna’s Evidensca) at Inveresk, Lamond/Almond at Cramond, Velunia at Carriden and Volitanio at Kinneil. In all of these names the hill-letter l  is used in an inversion-type manner as in the Olc element of Olcaclavis, as in Mugulesde and as in the Lec element of Lecilodanum. It is thus most probable that Lecidensca, Lamond/Almond, Velunia and Volitanio all use the hill-letter l2. What this means is that a swathe of land stretching from Doune on the river Teith right along the south side of the Firth of Forth to the North Sea was occupied by people who used the hill-letter l2. Now, in Chapter 16 of the Home menu (Roman names in Scotland) the present writer identifies the Roman fort at Camelon as Ravenna’s Poreoclassis. If this is correct, if we are right in seeing Ptolemy’s (H)orrea as being the same place as Ravenna’s Poreoclassis, and if Ptolemy was right in assigning (H)orrea to the Venicones, then Camelon lay in the territory of the Venicones. But Camelon lies in that swathe of land occupied by people who used the hill-letter l2, so it follows that the Venicones used the hill-letter l2 and that they lived south of the Firth of Forth. This is contrary to current belief, which has it that the Venicones lived in Fife, though some scholars have placed them further north still, in Angus. Note, too, that the conclusion reached above means that Traprain Law, if it was a tribal capital, was a capital of the Venicones and not of the Votadini. This may surprise those who believe that the Votadini occupied all the land between the Forth and Tyne, though the place-names available to us indicate that that region is highly unlikely to have been occupied by one and the same tribe. If we move south of Lothian, for example, there was a  Lenda (given as a river-name in Ravenna but it is in fact a land-name transferred by the Romans to a river) which appears to have been on or near the river Till, the l of Lenda being l1. In addition, it is suggested elsewhere on this website that the Celtic forms of Itucodon (Thockrington) and Maromago (Low Learchild) were respectively Litucodon and Lacerocomaguve. If these suggested forms are correct the initial l of Litucodon could be l1 or l2  (though l1 is more likely), whereas that of Lacerocomaguve is l1. It is also suggested elsewhere on this website that the Celtic form of Habitanci was Gamblidandi, the l in this form being l1. We also have Bindobala/Vintovala (possible Celtic forms of Vindovala) at Rudchester, the l in this name also being l1. Alavna at Alnwick is a river-name transferred to a Roman-fort, so the l in this name is a river-letter, not a hill-letter. But there are in this region several names which do not include the hill-letter l. We see Certisnassa at Berwick or Tweedmouth and Venutio, apparently somewhere around the Kielder reservoir. In addition we see Serduno at Wallsend, Condecor at Benwell and Onno at Halton Chesters. It is thus clear that in the region between the Tweed and the Tyne as a whole no hill-letter is predominant as the only or the latest hill-letter of Celtic/Romano-British place-names, though north of the river Coquet the hill-letter l1 is predominant. This is in complete contrast to Lothian, where l2 is the only or the latest hill-letter in each and every place-name known to us (leaving aside Mugulesde at Stirling which had been occupied by the l2-people but was taken over by the Damnoni, who used the hill-letter m, and of course leaving aside Latin Poreoclassis/(H)orrea). It thus seems quite reasonable to regard the l2-people of Lothian as being quite distinct from the people who lived further south and so not unreasonable to conclude, as was done above, that the l2-people who lived in Lothian were in fact the Venicones.

6.1     But a word of caution. The above conclusion as to the location of the territory of the Venicones is based on the common assumption that Ptolemy’s (H)orrea and Ravenna’s Poreoclassis were one and the same place. But there is no pressing need for them to have been one and the same place and there are good reasons for taking them to be two quite different places. The chief reason concerns the very name Venicones. As explained in ‘The Celtic “Picts”’ and ‘The Boresti and Mons Graupius’, both of which may be accessed from the main menu above, with the exception of the Taxali all of the tribes living north of the line of the Antonine Wall had tribal names derived directly from the topographical place-names of the respective tribal centres. The tribal centre of the Venicones will have had a name such as Veniconion, which comprises the straightforward inversion-type topographical element Venic, which means ‘slope of hill steep’.  But Venic is an element in the hill-letter n, so the l2-people living south of the Firth of Forth cannot have been the Venicones. The Venicones can only have been the n2-people who lived in a swathe of land extending east/southeast from Braco in Perthshire (Ravenna’s Bograndium) in the direction of the river Leven, which includes the river-letter l applied to minor rivers by those who used the hill-letter n. Veniconion must therefore have been somewhere in that area, as must Ptolemy’s Orrea. If Orrea was originally Latin Horrea, as most scholars seem to assume, then Orrea was presumably a storage centre built after the battle of Mons Graupius to collect the taxes of the people living in what is now Clackmannanshire, Kinross and Fife, the taxes being paid in the form of grain and other foodstuffs. Orrea may have been on the coast, but need not have been. The l2-people living south of the Firth of Forth must therefore have had some other name which has simply not come down to us in the ancient texts. 


7     The Parisi and other l-people

Looking at the North of England the area north of the Humber appears to have been populated by people who used the hill-letter l2  at the date of the Roman invasion. We see this in the river-name Derbentione, now the Derwent. The river-letter t, together with an ione ending, was added on to the end of the earlier river-name Derben, this corresponding to the Lucamos of Lucamosessa, apparently the Celtic form of Ravenna’s Camulosessa at Malton. The D at the beginning of Derbentione is a modified river-letter t corresponding to the hill-letter l1 in Lucamos, whereas the t at the end of the full name corresponds to the hill-letter l2 (the people who used the hill-letter l2 seem normally to have added their river-letter t to the end of an existing river-name). We can thus deduce that l2 was the latest hill-letter used in that region even though it is not clearly present in any place-name available to us. We do have Delgovicia, either at Wetwang or a little further west, nearer to Fridaythorpe, but it is not clear whether the l  in that name is l1 or l2. It is also possible that Ravenna’s Decuaria, Ptolemy’s Petuaria at Brough on Humber, may have been Delcutaria, but we have no proof that it was. And in any event the l in Delcutaria could be l1 or l2. All we have to go on, therefore, is the river-name Derbentione, but at least this tells us that people who used the hill-letter l2 lived in that area prior to the arrival of the Romans, and because they used the hill-letter l2 they must have come to Britain at some time between about 120BC and AD43. These people were surely Ptolemy’s Parisi. But note that although Ptolemy states that the Parisi lived on the Opportunum bay, which must have been the Roman name for the Humber, the Celtic name being Ga(m)brantuicorum bay, it is clear that the l-people occupied a sizeable area to the west and southwest.  In the Aire valley, for example, we see Lagentium at Castleford, Alitacenon, apparently at Adel, just north of Leeds, and Loxa at Keighley. The initial L of Lagentium is presumably l2 since Lag is the generic element within the name and is qualified by the earlier element ent, in which n is n2. The initial l of Loxa is clearly l2 since the corresponding river-letter t appears, changed to th, at the end of the river-name Worth, the Worth being the river which flows past Loxa at Exley Head, Keighley. The l of Alitacenon, the name before Loxa in Ravenna, is then presumably also l2In addition it is clear that the l-people occupied the upper valley of the Derbentione  (the Derbyshire Derwent) and the lower reaches of the Trisantona (the Trent) since the river-letter t at the end of Derbentione and Trisantona  corresponds to the hill-letter l2. It is thus probable that the intervening region, around the river Don, was also occupied by the l-people, the river-letter t  changed to d in Don then corresponding to the hill-letter l2. But perhaps only the l-people living north of the Humber remained independent as the Parisi. Those further west and southwest may have been brought under Brigantian rule. This may be what Ptolemy means when he writes, after referring to the Parisi, "below these are the Brigantes, even though he had already referred to the Brigantes earlier.


8     The Brigantes

8.1     Leaving aside the southern Brigantes discussed above, the region apparently occupied by Ptolemy’s Brigantes is made up of a variety of areas in each of which a particular hill-letter is dominant, but different hill-letters are dominant in different areas. In the northeast of modern Yorkshire is a region which seems to have been occupied at one time by people who used the hill-letter s. One sees this s in Lucamosessa at Malton, where s is the earliest hill-letter in the name. But we see also the corresponding river-letter b, changed to v, in river-names such as Dove and Leven. And there was presumably a Celtic settlement with a name of the isca-family at or near the later Roman fort on Lease Rigg, the Romans then transferring the isca part of the name to the river now called the Esk and its tributary the Murk Esk. And further west we see the hill-letter s in Castaractonion (original form of Cataractonium) at Catterick and Deriscoti (apparently the original form of the Dixio element of Ravenna’s Dixiolugunduno) at Roulston Scar, just east of Thirsk. Malton was then taken over by people who used the hill-letter m and Catterick and Roulston Scar by people who used the hill-letter r. We see this latter hill-letter also in Eburacum at York, in Cardadoriton (apparently the original form of Tadoriton) at Tadcaster, and in Maboridon (original form of Maporiton) at Bramham. It would thus seem that at one point northeast Yorkshire was in the hands of people who used the hill-letter s and that the Vale of York was in the hands of r-people. But further west and north we see a different picture.  On the Stainmore pass we have Lavaris at Bowes and Valteris at Brough-under-Stainmore. Both names use the hill-letter l1, as do Caluvio, apparently over the Pennines at Ingleton, Camulodono at Skipton, though this last place was lost to people who used the hill-letter n2 , and Bamvocalia (original form of Pampocalia) at Ilkley. We also see l1 in Galava (Lancaster) and Galluvio (Celtic Gabluvion, Casterton). But one should note another group of names – those using the hill-letter n2. These are Camulodono at Skipton, Smedriadunum (apparently the original form of Smetriadum) at Bainbridge, Cambroduno (original form of Camboduno) at Eccleshill in Bradford, Clindum at Clint, Carbandium (original form of Carbantium) at Harrogate, and, further north, Lugunduno (part of Ravenna’s Dixiolugunduno) at Dinsdale Park on the Tees. In all of these names the hill-letter n2 appears to be the latest hill-letter in the name, and within the region defined by these names is not only the Romano-British capital of the Brigantes, at Aldborough near Boroughbridge, but also the large Celtic fort at Stanwick, held by some scholars to have been the earlier capital of the Brigantes.  It may thus be that these people, who used the hill-letter n2, were the original Brigantes of Ptolemy and that they managed to bring neighbouring tribes under their control, so that their territory stretched from sea to sea, as Ptolemy tells us.

8.2     But if the territory of the Brigantes stretched from sea to sea then that tribe must also have controlled the region west of the Pennines. Here, in addition to Caluvio at Ingleton, Galava at Lancaster and Galluvio at Casterton, all mentioned above, there were some l1 names around Carlisle, for example Gambaglanda (original form of Gabaglanda) at Birdoswald and Ucseludamo (original form of Uxellodunum) at Castlesteads, though both were lost to other people. There was also Corucingolonion (possibly the original form of Croucingo) at Croglin and Bladobulgio (original form of Blatobulgio) at Birrens/Burnswark Hill. And there was Carlisle itself, Lagubalium (the bal element uses l1). But there were also l2 -names. Lagubalium at Carlisle is one (the lag element). It is not clear whether Avalava at Burgh by Sands and Olerica at Maryport use l1 or l2 but the river name Derwent suggests that the area around this river was at one time controlled by l1-people and was later taken over by l-people – the t changed to d at the beginning of Derwent corresponds to the hill-letter l1, whereas the t at the end of the river-name corresponds to l2. This seems thus to have been a region where l1 names at one time predominated, but where at least part of the region was later taken over by people who used the hill-letter l2. Ptolemy gives us no name for these l-people.

8.3     The area on and to the south of the Tyne is somewhat confusing. We see names in which n is the only or the latest hill-letter, for example  Binovia (original form of Vinovia) at Binchester, Bindolande (original form of Vindolande) at Chesterholm, Serduno at Wallsend and Condo (possibly the original form of Onno) at Halton, names in which l  is the only or the latest hill-letter, for example  Lincobigla at Lanchester, Brocoliti at Newborough (the name later being transferred to Carrawburgh), Velurcion at Grindon Hill (the name later being transferred to Housesteads) and Celovion at Chesters on the North Tyne, and names in which r is the latest hill-letter, for example Condecor at Benwell. Of the names in the hill-letter l the l  in Celovion is l1 and the initial l of Lincobigla is l2. In the case of Velurcion and Brocoliti it is not absolutely clear whether the l  is l1 or l2, though l1 seems more likely. There was thus a very mixed population in this region. One cannot say on the basis of the place-names whether one particular tribal group was dominant.

8.4     Overall, then, in the territory of the Brigantes, at the date of the Roman invasion the l-people were dominant in the west (Galluvio, Galava, Caluvio, Pampocalia) and on the Stainmore pass (Lavaris, Valteris), the l-people in the northwest (Lagubalium and the area around the river Derwent) and in part of the Aire valley (Lagentium, Alitacenon, Loxa), r-people in the Vale of York (Eburacum, Castaractonion, Deriscoti, Cardadoriton, Maboridon), and n-people in a central region (Camulodono, Smedriadunum, Cambroduno, Clindum, Carbandium, Lugunduno), with perhaps some s people still surviving in northeastern Yorkshire. The population of the Brigantes was thus very mixed, with a group of n-people controlling the central region – these may have been the original Brigantes

8.5     And now we can see a possible alternative explanation for Ptolemy’s reference to Brigantes living below the Parisi, for there were in Lincolnshire four place-names which include the hill-letter n2. These are Gambrand (incorporated in the name Gabrantuicorum bay) which the present writer identifies as South Ferriby or a place up on the high ground between Barton-upon-Humber and South Ferriby, Banvobalum (Celtic form of Bannovalum) at Caistor-on-the-Wolds, Lindum at Lincoln and Croconcalana (Celtic form of Crococalana) at Ancaster. At some stage Caistor, Lincoln and Ancaster were lost to l-people but it may be that the area south of the Humber was still occupied by n-people at the date of the Roman invasion, if Gambrand was still in existence at that date. Those n-people south of the Humber would then be Ptolemy’s Brigantes living below the Parisi.

8.6   The tribal name Brigantes appears quite clearly to be based on a topographical place-name, perhaps of the form Brigantion. This was presumably the name of the place where the tribal chieftain lived at the time the tribal name was coined. It is not clear whether the place-name comprises old-style Br (meaning ‘high hill’) and transitional gant (meaning ‘steep hill high’) or transitional Brig (meaning ‘high hill steep’) and inversion-type ant (meaning ‘hill high’). It is also possible, even though all recorded forms of the tribal name have a t, that the place-name was earlier Brigandion, where old-style gand means ‘steep hill summit’), the d later changing to t. In all three of the above cases the chronologically later hill-letter in the place-name is n2, the hill-letter used by the Brigantes. And of course the Brigantes were just the people of Brigantion.


9     The Coritani

9.1     Moving south to the Coritani it seems quite clear that this tribe used the hill-letter l1. The place-names in this region in which l1 is the only or the latest hill-letter are shown on the map below.