Explanatory alphabetical list of Romano-British place-names


Part 5: E - G



[For information as to which names are included in this list and which not, and an explanation of the abbreviations employed, click on Alphabetical List menu provided above]




 (Celt) (Rav) (Ptol) (AI)   (Mod)
 Bindo Ebio (176) Epiacum Vindomora   Ebchester
      (Iter I)   (Durham) 

Note 1:

Rivet and Smith equate Ravenna's Ebio with Ptolemy's island Epidium, which they say was probably Kintyre. The name appears to them to mean simply 'place inhabited by the Epidii'.


But the order of names in Ravenna makes it quite clear that Ebio was the fort at Ebchester. Likewise, Epiacum belongs to a group of Ptolemy names which appears to define an Itinerary going from north to south down the eastern side of the Pennines, starting with Epiacum and then going south via Binchester and Catterick, this strongly suggesting that Epiacum was at Ebchester. However, there is equally no doubt that the name Vindomora was applied to the fort at Ebchester. The answer to this little conundrum is that Bindo will have been the name of the Iron Age promontory fort (NGR: NZ 098 560) believed to have been located on the summit of the high hill on the opposite side of the river Derwent from Ebchester and that Bindomora  is actually a river-name of the kind comprising a river-suffix, comprising one or more river-letters, attached to a place-name including one or more hill-letters. The river-suffix here is mora, comprising the river-letters m (corresponding to the hill-letter n in Bindo) and r. Ravenna’s Ebio will just be a modified form of Bindo. The nd has been dropped and an E has been placed at the front of the shortened place-name. Ptolemy’s Epiacum is presumably just an attempt to write Ebio in a more classical Latin manner, though the b has changed to p. The first fort at Ebchester is believed to have been occupied from around AD69 to AD117 (see, for example, the Pastscape website of Historic England). During that period it will have been called Ebio/Epiacum. The fort is said to have been reoccupied around AD150, and during this second phase the fort will have been the Vindomora of the Antonine Itinerary, the fort simply having taken the name of the river in the Bindomora form, with the common b→v change. Note that the Notitia Dignitatum lists a Derventione after Longovico at Lanchester and before starting a list of forts on Hadrian’s Wall with Segeduno at Wallsend. The writer has suggested elsewhere that that Derventione might have been the fort at Washingwells in the western outskirts of Gateshead, but if the Ebchester fort had in fact three occupation phases then the third phase fort may have taken the name of the river in the Derbentione form, again with the b→v change. Note that there are other cases where one and the same river had two names, one a name made up only of river letters (here Derbentione) and the other a name comprising a river-prefix or river-suffix attached to the name of a place on the river concerned (in this case Bindomora). For example, the river Duddon in Cumbria appears to have been known as the Tuton and the Colantia (Ravenna’s Coantia) and the river Brit in Dorset as the Brit (with some ending) and the Velox.


Note 2:

Vindomora shows b → v

Epiacum shows b → p


[The entry for Ebio was last modified on 16 February 2021]




(Celt) (Rav) (Ptol) (AI) (ND) (Mod)
Eburacum Eburacum Eboracum Eburacum leug. VI victrix Sextae York
   (137)   (Iter I)    
      (Iter II)    
      (Iter V, VIII)    


Rivet and Smith mention two possible derivations. Firstly, the ebur element might be derived from hypothetical British eburo, taken to mean 'yew', so that the name as a whole might mean 'place of yews', 'place abounding in yews'. The other possibility, with the aco suffix meaning 'property of, estate of', is a meaning 'estate of Eburos', where Eburos was a personal name known in Gaul.


But the single most important point to make about York is that there is at that point a substantial area of high, dry ground on both sides of the river Ouse, this providing a natural land bridge linking eastern Yorkshire to the high ground further west, nearer the Pennines. It also made York a natural transshipment point for transport by land and water. It is therefore most probable that the ebur element is simply an old-style topographical element meaning 'high hill', the hill-letter r in ebur corresponding to the river-letter s in Ouse. It is not clear whether the acum part of the name is just an ending, which may or may not have had some meaning, or whether the name includes the transitional topographical element burac meaning 'high hill steep'.


And of course the ND's Sextae has no connection with the name Eburacum - it simply refers to the Legio VI Victrix, which was based at York. 


[The entry for Eburacum was last modified on 09 April 2021]




(Celt) (Rav)       (Mod)
Eburoucselum? Eburocaslum (184)       Broomholm
          (Dumfries and Galloway)

Note 1:

Rivet and Smith, following Schnetz, saw the correct form of the name as being Eburo Castellum. They suggested that the word castellum may have been abbreviated to caslum on the map used by the compiler of Ravenna. They see the Eburo element as either being derived from hypothetical British eburo, taken to mean 'yew', or as a personal name Eburos, which was known in Gaul.


The order of names in Ravenna points to Eburocaslum having been at Broomholm, and the fact that the Broomholm fort stands on a high, steep hill adjacent the river Esk suggests that the Celtic name may have been somewhat of the form Eburoucselum. The point is that the rivers of the Axe, Esk, Exe, Usk family all appear to take their names from a place-name including an element sc meaning 'hill steep' or cs meaning 'steep hill' - for example the Exe in Devon takes its name from the name Isca of the fortress at Exeter, the Yscir (west of Brecon) takes its name from the fort at Brecon Gaer, apparently Yposcessa, the Somerset Axe appears to take its name from Ptolemy's Iscalis, the Axe in Devon is derived from the river-name Traxula, which incorporates the place-name Acsula, the Esk in Midlothian takes its name from Evidensca at Inveresk, and the South Esk in Tayside takes its name from Ptolemy's river (Is)caelis, which itself incorporates the place-name Isca. One would therefore expect to see a place with an isca-type name on the Esk in southern Scotland. In addition, the fact that the Broomholm fort stands on a high, steep hill indicates that the bur element of Eburocaslum may be an old-style topographical element meaning 'high hill'. The caslum element may in fact be correct as it stands, a compound in the old-style cas meaning 'steep hill' and the hill-letter l, but may alternatively be a latinisation of the Celtic ucselum, which is an old-style compound in the hill-letters s and l, ucs meaning 'steep hill'.


But the above explanation is not entirely satisfactory because the order of the hill-letters in the name would be first r and then s, which is contrary to the normal chronological order of these hill-letters. This is of course possible if, for example, the s-people had been expelled from some other region of Britain and moved into the Broomholm area, taking it over from r-people. But in the absence of any supporting evidence that the hill-letter r reached this area before the hill-letter s it seems safer to assume that Eburocaslum is in fact a river name similar to Ptolemy’s Abravannus (the Water of Luce in Dumfries and Galloway).  It would be a river-name of the kind having a river-element, here a compound of the river-letters b and r, used as a prefix to a land-name, in this case caslum. The place-name caslum, as indicated above, may be correct as it stands, or it may be a modification of a Celtic name somewhat like ucselum or perhaps even iscalum (though the ending of the Celtic name may have been different). The Celtic settlement called Caslum/Ucselum/Iscalum may have been the pre-Roman settlement at Broomholm itself (see, for example, the entry for Broomholm on the Canmore website of Historic Environment Scotland), but may alternatively have been one of the hill-forts along the course of the Esk, one of which is the fort known as Castle O’er, this standing on a steep hill between the White Esk and Black Esk rivers. But wherever Caslum/Ucselum/Iscalum was the river became known as the Eburocaslum/Eburoucselum/Eburoiscalum, and this name was transferred by the Romans  in the form Eburocaslum to the fort which they built at Broomholm.


Note 2:

Possible example of latinisation of Celtic name


[The entry for Eburocaslum was last modified on 20 January 2021]




(Celt) (Rav)       (Mod)
Elconionemedo Elconio Nemeto(tatio Tamaris)       Launceston
  (3, 4, 5)       (Cornwall) 

Note 1:

Ravenna's names have long caused difficulty, partly because it is not clear whether the subdivision of the names in the surviving manuscripts is correct. Richmond and Crawford suggested as one possibility that Nemetotatio might be restored to Nemetostatio. Rivet and Smith see Elconio as a name quite separate from Nemeto Statio. As for Elconio they take the view that what the compiler of Ravenna actually saw on his map was the river-name Fl Conio (Fl short for Flumen) and that he misread this as a place-name Elconio. Nemeto they see as a Celtic word meaning 'sacred grove' and derived from hypothetical nem-os assumed to mean 'heaven'. Statio they take to be a Latin word meaning 'posting station and/or tax-collecting centre'.


It seems clear, however, that the confusion is overcome if we take the names in Ravenna to be Elconionemeto Statio Tamaris. Statio Tamaris must surely mean something more than a mere posting station. It may have referred to the tax-collecting centre for the Tamar basin, though one asks why this particular tax-collecting centre, out of all the tax-collecting centres that must have existed, should actually be identified as such in Ravenna. Elconionemeto appears to be a Celtic topographical name, but one which has been slightly modified. The old-style element con means 'steep hill' and the inversion-type element met means 'hill high'. But the elements are in the wrong order within the name, so met must originally have been the old-style element med meaning 'hill summit'. The Celtic name at the date of the Roman invasion must therefore have been Elconionemedo. The earliest name will have been Conion and then the people who used the hill-letter m arrived on the scene and added their element med meaning 'hill summit', so that Conionemedo simply means, in modern English, 'summit of a hill called Conion'. The Celtic hill-fort or settlement called Conionemedo will have been up on high ground, perhaps where Launceston Castle now stands. The initial El  of Elconionemedo appears to be the hill-letter l  used in an inversion-type manner, so that Elconionemedo means 'hill called Conionemedo'. This could mean that the hill-top settlement called Conionemedo was replaced by a new Celtic settlement called Elconionemedo, which may have been on the low ground at the foot of the hill. The later Roman fort then took its name from that new settlement. Note that the river-letters corresponding to the hill-letters n and m of Conionemedo are m and r, and these appear in exactly that order in the river-name Tamaris, now the Tamar. The initial T of Tamaris is the river-letter corresponding to the inversion-type hill-letter l  of Elconionemedo. Note, too, that it is possible that the initial Launc of Launceston is derived ultimately from the Elcon of Elconionemedo, though this is at odds with the conventional etymology for Launceston.


Note 2:

Example of d → t 




(Celt) (Rav)       (Mod)
Eltavori Eltavori (93)       Lutterworth


Rivet and Smith appear to take the view that the name on the map used by the compiler of Ravenna was Fl Tavori, where Fl is of course an abbreviation for Latin Flumen or Fluvius, and that the Fl on the map was capable of being misread as El. Eltavori then really means 'river Tavori', though Rivet and Smith take the view that the correct form of the river-name is Tamus, this being a reference to the river Tame, a tributary of the Trent.


But the name in the original Cosmography was probably Corieltavori, this name appearing on a tegula which was found at Caves Inn, just south of Lutterworth (the inscription actually reads '..vitatis Corieltauvorum'). The Eltavori part of the name is thus probably a tribal name rather than a place-name. But it came to be applied to the place itself, to Lutterworth, since Ekwall indicates that Lutterworth was Lutresurde in 1086, and lutre is most likely to be simply a rearrangement of Eltavori (perhaps EltavoriEltauriLutriLutre). But at some stage of medieval copying the Cori of Corieltavori was attached by mistake to the previous Ravenna name, Rate, so that the names actually appearing in the surviving copies of Ravenna are Ratecorion and Eltavori. The above rather suggests that Lutterworth was the early, pre-Roman tribal centre and that it was the Romans who established Rate at Leicester as the new administrative centre of the tribe. Note that Eltavori also exists in the form Eltanori, so that Ptolemy's tribal name Coritani is probably just a Roman simplification of Corieltanori, i.e. Cori [el] tan [or] iCoritani. Note further that Eltan  may be a topographical name including the inversion-type element elt meaning 'hill high'. The people of Eltan may then have been known as the Eltanori  and Corieltanori  may mean 'meeting place of the Eltanori', where the Cori element was most probably applied by the Romans. 


[The entry for Eltavori was last modified on 11 February 2021]









(Celt) (Rav)     (ND) (Mod)
Esica Esica (150)     Aesica first, Haltwhistle Burn
          later, Great Chesters

Note 1:

This name is believed by scholars to be connected with that of the Celtic god Esus or Hesus, whose name, or a personal name derived from it, is said by Rivet and Smith to be present on the coins of three British tribes, the Dobunni, the Iceni and the Coritani. However, Rivet and Smith do not appear to suggest any meaning for the name.


There seems little doubt, however, that this is a name of the isca-family, isca being an inversion-type name meaning 'hill steep'. Ravenna's Esica was the Trajanic fortlet at Haltwhistle Burn (this is explained in Chapter 20: Rome's frontiers in northern England), the name referring to the almost vertical drop down to the stream at the western end of the fortlet. The name was later transferred to the Hadrianic Wall fort at Great Chesters and the fortlet was dismantled. At some point Esica changed to the Aesica of the ND, though the change of initial e to ae is of no significance so far as the meaning of the name is concerned.


Note 2:

Example of transfer of name to a different location

Example of initial e → ae







(Celt) (Rav)       (Mod)
Lecidanum Evidensca (189)       Inveresk
          (East Lothian)


Richmond and Crawford gave Eiudensca as their main reading, though pointed out that the Basle manuscript gives the form Evidensca. Rivet and Smith see Evidensca as the correct form but see this as a corrupt form of Habitancum, at Risingham in Northumberland, and provide an etymology on the basis of the latter name.


Evidensca is the next but one name after Alavna (187) in Ravenna, and by the time Ravenna reaches Alavna at Alnwick it has listed names from Chesters on the North Tyne up to Berwick-on-Tweed, names from High Rochester over to Alnwick, and also Newstead and Broomholm. It is therefore unlikely that Ravenna returns south to Risingham after listing Alavna, so that Evidensca is most likely to have been north of the places just mentioned. In addition, river-names of the Esk family took their names from place-names including an element of the isca-family, so we need such a name to account for the river Esk in Midlothian. Evidensca is the only such name available and it is in the right position in the Ravenna list, so it seems quite clear that Evidensca was one of the two forts on the river Esk. Inveresk was one of those forts, Elginhaugh the other. Coming up from the south one reaches Elginhaugh first, so this fort is likely to have been Olcaclavis, the name after Alavna in Ravenna. That leaves Inveresk to have been Evidensca, and the latter name is entirely appropriate for the location. Evidensca is, however, a very unusual name in having a dunum-type element - den - in the middle of the name and an isca-type element at the end. There were a number of place-names with an isca-type ending in Gaul, either on or east of the Rhône-Saône axis, for example Matisco, Tarasco and Vindisca, but Inveresk is surely too geographically remote to have belonged to that group, and by the time the isca-ending reached Britain it had changed to essa (see Updates19 October 2015). It therefore seems likely that it was the Romans themselves who added the isca-ending to the name of the Inveresk fort, more specifically the Legio II Augusta which had been based at Isca at Exeter and later simply transferred that name to their new fortress at Caerleon. It would appear that they acted in a similar manner at Inveresk. The reason is most probably that the Celtic name for Inveresk was Lecidanum, very similar to Leciodanum at Doune, so the Romans added isca to the Lecidanum at Inveresk just to distinguish the two forts more clearly. And of course they then transferred the place-name, or at least the isca-ending of it, to the river flowing past the fort, that river now being called the Esk. The Roman name of the fort was thus Lecidanisca, where the inversion-type elements lec and dan respectively mean 'hill steep' and 'summit of hill'. The name changed to Levidanisca (cf. Leciodanum → Leviodanum at Doune and Racatonium → Ravatonium at Stracathro), the initial L was lost or dropped and Evidanisca changed to Evidensca.




(Celt) (Rav)     (ND) (Mod)
Cambaglanda Gabaglanda (131)     Amboglanna Birdoswald

Note 1:

Rivet and Smith see the correct form of the name as being Camboglanna, the first element being derived from hypothetical cambo taken to mean 'curved, crooked' and the second from hypothetical glanno given the meaning 'bank, shore', the name as a whole thus meaning 'curved bank' or 'bank at the bend'.


Ravenna's Gabaglanda is, however, early - it is Flavian. It will originally have been Cambaglanda, Gabaglanda having changed the intial c to g and having lost the m. The ND form drops the initial C and changes anda to anna. Cambaglanda is a straightforward topographical compound in which nd meaning 'hill summit' is qualified by the previous name of the high ground, Cambagl, this itself comprising the element bagl meaning 'high steep hill' qualified by the element cam meaning 'steep hill'. Note that the river-letters corresponding to the hill-letters m and l in Cambaglanda are r and t, both of which are present in the name Irthing of the river flowing past the fort, the t being changed to th in the modern name.   


Note 2:

Example of cg

Example of omission of m

Example of andaanna

Example of omission of initial letter - C  




(Celt) (Rav)     (ND) (Mod)
Gabrocentio Gabrocentio (117)     Gabrosenti Hard Knott

Note 1:

Rivet and Smith see Gabrosentum as being the correct form, the first element being derived from hypothetical British gabro-s taken to mean 'horse' or 'goat' and the second from hypothetical Celtic sento given the meaning 'path', the name as a whole thus having a meaning such as 'goat path'.


Ravenna's form, however, appears to be Flavian and there is no good reason to doubt that the c is correct. Gabrocentio is a straightforward topographical compound in which old-style gabr meaning 'steep high hill' is followed by transitional element cent meaning 'steep hill high'. The name appears entirely appropriate for the Hard Knott fort. The ND form changes the c to s.


Note that in their discussion of the hypothetical root gabro-s taken to mean 'horse' or 'goat', Rivet and Smith refer to Gabromagus, now Windischgarsten in Austria, and Gabrodunum, now Jabrun in the Cantal area of France. Both of these places are in fact associated with a high, steep hill, as is the fort at Hard Knott.


Note 2:

Example of cs   









(Celt)     (AI)   (Mod)
Galava     Galava   Lancaster
      (Iter X)    


Jackson suggested a hypothetical root gala with an ava suffix, the name as a whole meaning 'vigorous stream'. The name was simply transferred from the river to a fort built on the banks of the river concerned. Holder suggested hypothetical gala, taken to mean 'bravery'. Rivet and Smith identified Ambleside as Galava.


However, the mileages quoted in Iter X point to Galava having been the fort at Lancaster. This was built on a hill on the south bank of the river Lune. The name Galava may comprise the old-style element gal, meaning 'steep hill', with an ava ending or, if the first fort was built on the hillside, the old-style element galav meaning 'steep hill slope'. It may be that the Lan of Lancaster is derived from the lav of Galava with the v changed to n.







(Celt) (Rav)     (Mod)
Gabluvion Galluvio (113)     Casterton

 Note 1:

Rivet and Smith regard Ravenna’s Galluvio as referring to the same place as the Galava of Iter X of the Antonine Itinerary, but this can surely be rejected. It seems sufficiently clear that Galava was at Lancaster and Galluvio at or near Casterton, somewhere close to the river Lune. The Romans will have transferred the place-name to the river, the modern river-name being derived from the luvio part of Galluvio with the common change of vn. The Lan of Lancaster will be derived from the lav of Galava, again with the common change of vn.

The Celtic name will have been Gabluvion, where Gabluv is an old-style topographical element meaning ‘steep high hill slope’. The Romans will have transferred the name to Casterton, with the change of bl→ll, from the hillslope fort known as Castle Hill (NGR: SD 650 779) to the southeast of Casterton. The slope of the hill is steep below the hillfort/defended settlement, more gentle at and immediately above the hillfort, and then becomes steeper again as the hill rises further.


Note 2:

Example of blll


(The entry for Galluvio was inserted on 21 March 2021]




(Celt)       (ND) (Mod)
Gariserna       Garriano, Gariannonor Burgh Castle
(a river)         (Suffolk)



Rivet and Smith see this name as being in essence that of Ptolemy's river Gariennus. They refer to a hypothetical root gar or ger proposed by Holder and Ekwall and taken to mean 'to shout, talk', but surely rightly reject the idea that the Yare might have been regarded as a babbling river. They see an association with the river-name Garumna (now the Garonne) and refer to Dauzat's proposed derivation from hypothetical car(r), taken to mean 'stone', the river-name then perhaps meaning 'stony river', though they seem unconvinced by this suggestion, too.


This place-name appears to be simply a further development of Ptolemy's river-name Gariennus, probably originally Gariserna, this being a river-name of the kind comprising a river-suffix, here serna, comprising the river-letters s and with an ena ending, attached to a place-name, here Gar. The place-name Gar is an old-style topographical element meaning ‘steep hill’ and the place concerned was probably Norwich, where there are indeed steep slopes adjacent the river. The Romans simply transferred the river-name Gariserna to the fort which they built downstream, near Great Yarmouth, the fort now called Burgh Castle. In some of the known forms of the river-name the vowel after the r is an i, in others a u, and the succeeding vowel is sometimes an a, sometimes an e, but these vowel differences are of no consequence. At some stage the s  of the river-suffix was dropped and rn changed to nn. In some variant forms of the name n → n n  or r → r r.


[The entry for Garriano  was last modified on 09 April 2020] 




(Celt) (Rav)       (Mod)
Glavo Eltabo Giano Eltabo (1, 2)       Nanstallon


Rivet and Smith refer to two possible derivations of Glano (which they regard as the original form of Giano). The name may be derived from hypothetical Celtic glano-s, taken to mean 'pure, shining', or possibly from hypothetical glanno, taken to mean 'bank, shore'. They appear to see Eltabo as a name quite separate from Giano, indeed as a river-name, where initial El is a misreading of the abbreviation Fl  for Latin Flumen, meaning ‘river’.

Giano and Eltabo may indeed refer to two different places but it seems more likely that they belong together as the name of the Roman fort at Nanstallon, just west of Bodmin. The l of Eltabo is not a misreading of the Latin abbreviation Fl, as argued by Rivet and Smith – it is simply the river-letter l applied to minor rivers by those who coined place-names in the hill-letter nEltabo is a compound river-name in the river-letters l and t (and b if the b is indeed the river-letter b and not just a name-ending) corresponding to the hill-letters n and l (and s). The hill-letters n, l and s are all present in the modern name Nanstallon, though this might just be a coincidence. The name Giano Eltabo will have been transferred to the Roman fort from the Celtic hill-fort just upstream from Nanstallon in Dunmere Wood. This hill-fort stands on a hillside, the drop down to the river being very steep, so the name Giano was probably originally Glavo, where Glav is an old-style element meaning ‘steep hill slope’. Eltabo will have been the then name of the river Camel and the Celts will have added Eltabo to Glavo just to distinguish this Glavo from another Glavo somewhere else, Glavo Eltabo then being a Celtic name rather like modern names of the form Newcastle-upon-Tyne and Newcastle-under-Lyme. Landini Tamese and Litinomago Devoni are other examples of Celtic/Romano-British names of this kind.







(Celt) (Rav)   (AI)   (Mod)
Glevon Glebon Colonia (62)   Clevo   Gloucester
      (Iter XIII)    


Jackson considered the Celtic name might have been Glevon, earlier Glaivon, to which Welsh gloyw or gloew, meaning 'bright', is related. The Romano-British name would then perhaps mean 'bright place, bright town'.


Glebon or Glevon was presumably the name of the early fort at Kingsholm, later simply transferred to the new fortress built a little further south. The Gl element is simply an old-style topographical element meaning 'steep hill'. If the original name was Glevon then evon was presumably just a name-ending, which may or may not have had meaning, but it is equally possible that the name included the old-style element glev meaning 'steep hill slope'. If ebon were the original ending then the name might in fact be a river-name of the kind having a river-suffix, here the river-letter b, attached to a place-name, here the old-style element Gl. There are no steep slopes at Kingsholm, or indeed at the site of the later fortress in Gloucester. It therefore seems most likely that Glevon was originally the name of a Celtic settlement adjacent the steep escarpment a little to the east of Gloucester, perhaps the settlement at Crickley Hill, to the right of the A417 as it makes its steep descent down to Little Witcombe. Note that the settlement was not built at the summit of the high ground, but actually on the slope, so the name probably did include the old-style element glev meaning 'steep hill slope'. Note, too, that the Roman road, once it has got down the escarpment at Birdlip, just south of Crickley Hill, heads for the fort at Kingsholm, not for the later fortress. It thus seems most likely that Glevon was the name of that settlement at Crickley Hill and was simply transferred by the Romans to their new fort at Kingsholm. The name was later transferred to the new legionary fortress at Gloucester and sometime between AD96 and AD98, after the departure of the legion, the civilian settlement at Gloucester acquired the status of colony. This colony is of course the Glebon Colonia listed by Ravenna.