Explanatory alphabetical list of Romano-British place-names

 

Part 8: M

Part 8b: Me to Mu

 

[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]

 

MEDIBOGDO             

(Celt)	(Rav)				(Mod)
Probably	Medibogdo (114)				Kitridding
Medibogldo					(Cumbria)
or
Mediboglodono

 

Note 1:

Rivet and Smith see Mediobogdum as the correct form, this being derived from hypothetical Celtic medio taken to mean 'mid, middle' and (following Williams and Walde-Pokorny) a hypothetical root beugh given the meaning 'bend, curve', the name as a whole thus meaning 'place in the middle of the curve' which, Richmond and Crawford thought, exactly described Hard Knott in relation to the Esk valley. Unfortunately, however, Medibogdo was not at Hard Knott.

 

Medibogdo will have been the name of an Iron Age hillfort or settlement transferred to a Roman post built somewhere in the vicinity. Ravenna lists Medibogdo between Galluvio at Casterton on the river Lune and Cantiventi at Watercrook on the river Kent. It thus seems clear that Medibogdo was between Casterton and Watercrook. And in that region of few hillforts there is in fact a hillfort lying more or less on a straight line drawn between Casterton and Watercrook. That is the hill-fort at Kitridding (NGR: SD 583 843). It stands at around 220 metres on a platform at the top of a steep, high hill, with views to the north, east and south, though the true summit, at 238 metres, is a little to the west.

 

There seems little doubt that Medibogdo is a topographical name, but it is incomplete. The initial med element is an old-style element meaning 'hill-summit'. The second element has the b meaning 'high' and the g meaning 'steep', but the hill-letter is missing. Given that the place-names in the Lune valley, just to the east, use the hill-letter l, as in Galluvio at Casterton, and that the river-letter t in Cantiventi at Watercrook, a little to the west, corresponds to the hill-letter l, it is probable that the missing hill-letter is l, thus giving a place-name of the form Medibogldo, where bogldo means ‘high steep hill summit’. But it is also possible that do is an abbreviation of dono, meaning ‘summit of hill’, so the place-name would comprise the three elements med, bogl (meaning ‘high steep hill’) and dono. In this case there may have been a linking vowel between the l and the d, thus giving a Celtic place-name of the form Mediboglodono. But, as noted above, the name will have been transferred to a Roman post built somewhere in the vicinity and it is that Roman post which will have been the Medibogdo of Ravenna.

 

Note 2:

Example of missing letter

Example of abbreviation

 

[The entry for Medibogdo was last modified on 22 March 2021]

 

 

MEDIOLANUM  see  MEDIOLANO

 

 

MEDIOLANO

(Celt)	(Rav)	(Ptol)	(AI)		(Mod)
Mediolano	Mediolano	Mediolanum	Mediolano		Whitchurch
	(84)		(Iter II, X)		(Shropshire)

Note:

This name is generally regarded as being derived from hypothetical Celtic medio, taken to mean 'mid, middle', and hypothetical British lano, taken to mean 'plain, level ground', the name as a whole perhaps meaning, so suggest Rivet and Smith, '(place) in the middle of the plain', or, according to Jackson, 'central plain'.

 

But the location of the Roman fort at Whitchurch cannot sensibly be described as being in the middle of a plain, nor as a central plain. The fort was in fact built on top of a hill, which is actually what the name indicates. The med part of the name is the old-style element med meaning 'hill summit', and the l is just the hill-letter l. The river-letters corresponding to the hill-letters m and l are r and t, both of which are present in the river-name Rutunio, which the Romans transferred to the fort  which they built close to the river (now the Roden) at a point, so Iter II indicates, 12 Roman miles south of Whitchurch on the road to Wroxeter. It is thus clear that the people who used the hill-letters m and l did settle in the region around Whitchurch. There can thus be little doubt that Mediolano is indeed a topographical name in the hill-letters m and l and has no connection with any plain, whether central or not.

 

But note that the relationship between Mediolano and the river Roden may be much closer than is suggested above. There is apparently no record of any Iron Age structure in Whitchurch itself. It would appear, then, that Mediolano was not a free-standing place-name in the Iron Age but was part of the name of the hillfort now known as Bury Walls (SJ 577 274), on the eastern side of the river Roden to the south of Whitchurch. The full name of that hillfort was probably Bicsimediolano, but there may have been a qualifying b (meaning ‘high’) or c (meaning ‘steep’) between the i and the o. The composite river-name of the river Roden (such names are explained in Home/Chapter 19, 11) will have been somewhat of the form Bortobicsimediolano, though the vowels used are not important. Note that the river-letters b, r and t of the river-prefix Borto correspond respectively to the hill-letters s, m and l  of Bicsimediolano. The Romans took the mediolano part of the name and applied it to the fort which they built at Whitchurch. The river-prefix was transferred to a Roman post built close to the river Roden and presumably on the road from Whitchurch to Wroxeter, the name of that post appearing as Veratino (with initial B changed to V) in Ravenna and as Rutunio (with initial B omitted) in Iter II of the Antonine Itinerary. The Romans used what was left of the composite river-name after removing mediolano, namely Bortobics, to coin a name for the tribe. They omitted the initial B and the hill-letter s, changed the t  to d and the second b to v to give the modified composite river-name Ordovic, and on the basis of this form they called the tribe the Ordovices. Note that Ptolemy assigns his Mediolanum to the territory of the Ordovices.

 

[The entry for Mediolano was last modified on 13 May 2021]

 

 

MEDIOMANO

(Celt)	(Rav)				(Mod)
Nediomano	Mediomano (81)				Tomen-y-Mur
or					(Gwynedd)
Mediolano

 

Note:

Rivet and Smith see this name as a duplicate of Mediolano at Whitchurch. That is, however, unlikely to be correct. Ravenna has come over from Brinavis at Bicester to Wroxeter, continues to Caer Gai at the southern end of Bala Lake and after Mediomano goes from Seguntio at Caernarfon to Sandonio at Sandon on the river Trent via Caerhun and Whitchurch. Mediomano is most likely to have been between Caer Gai and Caernarfon. Tomen-y-Mur is in the right place to have been Mediomano and the latter name is appropriate for the location. In addition it is just conceivable that the Tomen part of Tomen-y-Mur is derived from the dioman part of Mediomano.

 

The original Celtic name is, however, unlikely to have included the hill-letter m twice - the first m was probably originally some other hill-letter, n being the most likely, though s would also be possible. The original name was thus probably Nediomano, this comprising the old-style element ned meaning 'hill summit', the hill-letter m and an ano ending. The name is thus entirely appropriate for Tomen-y-Mur, which is indeed at the top of a hill. It is of course possible that the Tomen of Tomen-y-Mur has an origin quite different from that suggested above and that the Romano-British name for the fort was indeed Mediolano - the meaning of the name remains the same. But there was apparently no Iron Age settlement at Tomen-y-Mur, so the place-name must have been transferred to Tomen-y-Mur from some other location in that region. The nearest Iron Age hillfort is that at Bryn-y-Castell, Ffestiniog (SH 728 430), which is thought to date back to around 370 BC. The hillfort does indeed stand on the summit of a hill. Presumably that hillfort was called Nediomano/Mediolano and the Romans simply transferred the name (modified at some stage to Mediomano) to their fort at Tomen y Mur.

 

[The entry for Mediomano was last modified on 13 May 2021] 

 

 

MEDIONEMETON

(Celt)	(Rav)				(Mod)
Nedionemedon	Medionemeton (196)				Bar Hill
					(East Dunbartonshire)

Note 1:

The first part of this name is normally considered to be derived from hypothetical Celtic medio, taken to mean 'mid, middle', and the second part is believed to be a Celtic word nemet, having a meaning such as 'sacred place' and thought to be based on hypothetical nem-os, taken to mean 'heaven'. The name as a whole is thus thought to have a meaning such as '(place) in the middle of the sacred grove' or perhaps 'middle grove'.

 

Medionemeton is, however, a topographical name, though one which has been modified. The med part of the name appears to be an old-style element meaning 'hill summit' and the met part an inversion-type element meaning 'hill high', but the elements are then in the wrong order within the name. The met element must therefore have been the old-style element med originally, but one cannot have a med at the beginning and end of the name. The first letter of the name must therefore have been some other hill-letter originally, presumably an n or s, assuming that the hill-letters in the names in that area appear in the normal chronological order - n₁, s, m, r, l₁, n_2, l₂. The earlier name of the place will thus have been Nedion (or, possibly, Sedion), where ned means 'hill summit', and then the people who used the hill-letter m arrived on the scene and added their element med, which has the same meaning. The name is similar in structure to the conionemedo part of Elconionemedo (Launceston), where the earlier name was Conion and the med element was added by later settlers. And of course neither name - Nedionemedon or Conionemedo - has any connection with a sacred grove. But Nedionemedon will actually have been the name of the Iron Age fort on Castle Hill, some 200 metres northeast of the Roman fort on Bar Hill, and was simply transferred by the Romans to the early Antonine fortlet which stood on the same site as the later Wall fort. No doubt the name was retained by the Wall fort when this was built. For a discussion of the role of the early Antonine fortlet in the Antonine frontier see Chapter 22 (The Antonine Wall) of the Home menu. Note that the names Cerma and Cermium indicate that the hill-letter r  reached central Scotland earlier than the hill-letter m, so the first element of the Celtic name may possibly have been Redion rather than Nedion or Sedion.

 

Note 2:

Example of d → t

Example of n → m

 

[The entry for Medionemeton was last modified on 26 September 2020]

 

 

MELAMONI

(Celt)	(Rav)
Lelamon	Melamoni (15)

 

Note 1:

Rivet and Smith see this name as a duplicate of Moriduno. Richmond and Crawford offer no derivation, but suggest that Melamoni was in Devon, not far from Exeter.

 

It seems clear from the order of names in Ravenna that Melamoni was somewhere between Tavistock and Exeter. The i at the end of the name will belong to the next-following Ravenna name, Scadum Namorum. We are thus looking for a meaning and identification for Melamon. The Celtic name will have been Lelamon, since this place-name corresponds precisely with the river-name Dart, and between Tavistock and Exeter we see a West Dart and an East Dart river. The earliest hill-letter in the place-name is m, corresponding to the river-letter r in Dart. The hill-letter l in the middle of the place-name will be l₁ used in an inversion-type manner, and the corresponding river-letter t, changed to d, is also used in an inversion-type manner, i.e. it comes before the r in Dart. The initial L of Lelamon is l₂ and, as usual, the people who used the hill-letter l₂ placed their river-letter t at the end of the previous river-name, thus yielding the new river-name Dart. This river-letter t, placed at the end of the previous river-name by the people who used the hill-letter l₂, can also be seen in the river-name Derventio and in Trisantonis (the second t), as well as in the modern river-names Teviot, Tweed (t  changed to d) and Teith (t changed to th). The Celtic settlement called Lelamon will thus have been located close to the West Dart or the East Dart river. The Roman fort appearing as Melamoni in Ravenna will have been built in the vicinity of that settlement and will have taken its name from it, though with the initial L changed to M.

 

Note 2:

Example of l → m

Example of t → d

 

 

MELEZO

(Celt)	(Rav)				(Mod)
Meletio	Melezo (36)				Melbury
or					(Dorset)
Meledio

 

Note:

Williams suggested a British root cognate with hypothetical melisso-s, taken to mean 'sweet', the name thus having a meaning such as 'honey place' or 'honey-stream place'. Rivet and Smith think this semantically dubious and suggest derivation from a root such as mel or mell, apparently following Pokorny, who referred to a root mel, given the meaning 'height, swelling', or Dauzat, who proposed a hypothetical pre-Celtic mala, taken to mean 'mountain'. On that basis Rivet and Smith suggest a meaning 'hill-(place)' for Meletio.

 

The name appears to be a topographical compound in the hill-letters m and l, where the inversion-type element let means 'hill high' (z = ti). But this element may originally have been old-style led meaning 'hill summit'. From the order of names in Ravenna Melezo appears to have been at or in the vicinity of Melbury Hill in Dorset. There are ancient earthworks on that particular hill, so perhaps there had been a native settlement up on top of the hill prior to the arrival of the Romans, assuming that the earthworks date from that period.

 

[The entry for Melezo was last modified on 06 Auigust 2020]

 

 

MEMANTURUM

(Celt)	(Rav)				(Mod)
Melantvrum	Memanturum (216)				Dunfermline
					(Fife)

Note:

Rivet and Smith see this name as a corrupt form of Ptolemy's Novantarum peninsula, i.e. the Rhinns of Galloway, though it seems quite clear from the order of names in Ravenna that Memanturum was not in southwestern Scotland. Richmond and Crawford appear to go the other way and place Memanturum in northeastern Scotland. They see the name as probably being a scribal error for Nemanturum, and so think it might be a place of the Nemanturi, probably a sept or pagus of the Taexali tribe.

 

Memanturum appears to be a topographical name, but it should not include the hill-letter m twice. One of the two will thus originally have been some other hill-letter, l  being perhaps the most likely, this yielding possibly the place-name Melanturum. But it would appear that the name would actually have been Melantvrum or Melantverum. Melantverum is a topographical compound in the hill-letters m, l, n and r, where the inversion-type elements ant and ver respectively mean 'hill high' and 'slope of hill'. At some stage the ver element moved to the front of the name and in the post-Roman period a Gaelic dun was added at the front, the name then becoming Dunvermelantum, and it is this form which developed further, via Dunfermelitane in 1128, to the modern name Dunfermline. It is not clear whether the vr/ver element had this form in the original name, or whether the element was originally the old-style element br/ber meaning 'high hill'.

 

 

MESTEVIA

(Celt)	(Rav)				(Mod)
Mestevia	Mestevia (18)				Tiverton
					(Devon)

Note:

Rivet and Smith see this name as a corrupt form of Ptolemy's Antivestaeum promontory, i.e. Land's End in Cornwall. Williams considered the name to be derived from a root mess or mest, taken to mean 'acorns', the name then meaning 'place of acorns'.

 

Mestevia appears, however, to be a straightforward topographical compound in the hill-letters m and s, where the inversion-type element st means 'hill high'. The evia part of the name will just be an ending, which may or may not have had some meaning. The Roman fort at Tiverton was just north of the modern town, on the other side of the A361 and east of the river Exe. But it will have taken its name from Iron Age Mestevia on the other side of town. This is the hillfort now known as Cranmore Castle (SS 959 118) standing high above the confluence of the Exe and its tributary, the Lowman. For the reason for identifying Tiverton as Mestevia see Home/Chapter 10, 7.

 

[The entry for Mestevia was last modified on 13 May 2021]

 

 

METAMBALA

(Celt)	(Rav)				(Mod)
Nemedonbala	Metambala (50)				Lydney
					(Gloucestershire)

Note:

Rivet and Smith, following Richmond and Crawford, see Nemetobala as being perhaps the correct form of this name. They see the first part of the name as being hypothetical nemeto, taken to mean 'sacred grove', and the second part as being derived from hypothetical Celtic balma, taken to mean 'pointed rock, peak', the name as a whole perhaps meaning 'grove-hill' or 'hill-sanctuary'.

 

Ravenna gives two adjacent names as Iupania Metambala. The original names were presumably Luba Nemedonbala, but this text was at some point wrongly divided to yield Lubane Medonbala, Lubane then changing to Iupania. The earlier name of the place called Metambala in Ravenna will have been Nemedon, this comprising the hill-letter n and the old-style element med meaning 'hill summit'. And then the people who used the hill-letter l arrived on the scene and added their element bal meaning 'high hill'. There was presumably n/m confusion at some stage of medieval copying, so that Nemedonbala became Nemedambala and the initial Ne was lost to the previous name, Luba, thus leaving Medambala which, with the minor change of d → t, became Metambala. It may of course be that Nemedon was the name of the Iron Age hillfort just southwest of Lydney and that the name became associated with the hill itself, this being known as 'Nemedon high hill', i.e. as Nemedonbala. The Roman fort, assuming the Ravenna name refers to a Roman fort, will thus have taken its name from the hill, and may actually have been located at the foot of the hill, though so far as the present writer is aware no traces of such a fort have yet been found. In later times there was indeed a temple precinct inside the old hillfort, but the name Nemedonbala is topographical and has nothing to do with any sacred grove or indeed anything else sacred. Note that the river flowing down the east side of the hill on which the hillfort stands is called the Lyd, which comprises the river-letters l and t (changed to d) corresponding to the hill-letters n and l, both of which are present in Nemedonbala. The modern town Lydney will have taken its name from the river.

 

 

MILIDUNUM

(Celt)	(Rav)				(Mod)
Milidunum	Milidunum (19)				South Molton
					(Devon)

Note:

Rivet and Smith see this name as a duplicate of Moriduno and provide a derivation on the basis of the latter name. They think Moriduno might possibly have been at Sidford, Devon.

 

But Milidunum is an acceptable topographical name as it stands. It is a compound in the hill-letters m, l and n, where the inversion-type element dun means 'summit of hill'. The order of names in Ravenna appears to indicate that Milidunum was at South Molton, and there are several locations both within the modern town and in the immediate vicinity which would be suitable for the name Milidunum, the most likely being a promontory of higher land within the town, this promontory stretching eastwards towards the river Mole and being bounded by lower ground to both north and south. But note that Milidunum  is most likely to have been the name of the hillfort on Whitechapel Moors, at SS 757 264, some 4 kilometres east of South Molton. The name will have been transferred by the Romans to a fort which they built at South Molton itself.

 

 

MORBIO

(Celt)				(ND)	(Mod)
Morvio				Morbio	Castleford
(a river)					(West Yorkshire)

 

Note:

Rivet and Smith offer no derivation of their own and quote no derivations proposed by others. They think the name may be slightly corrupt.

 

Morbio is a river-name transferred by the Romans to a fort which they built close to the river concerned. The fort-name appears in the Notitia Dignitatum between Dano at Doncaster and Arbeia at Newton Kyme (not South Shields as is commonly assumed). Now, the river-letter m is normally applied to major rivers (by those Celts who used the hill-letter n) and the only major river between Doncaster and Newton Kyme is the Aire, so Morbio was presumably the then name of the river Aire. But the name was probably Morvio originally, where vio is just an ending, for if the b in Morbio were the river-letter b this would imply that in that region the hill-letters were not applied in the normal chronological order n₁, s, m, r, l₁, n₂, l₂. But we know from Loxa (at Exley Head, Keighley) and the modern river-name Worth (a tributary joining the Aire at Keighley) that the hill-letters were indeed applied in the normal chronological order in that region. The initial W of Worth is the river-letter b (corresponding to the hill-letter s) changed to v changed to w. The r in Worth is the river-letter r corresponding to the hill-letter m, so we know that in that region the hill-letter s  was indeed used earlier than the hill-letter m, and so the river-letter b was used earlier than the river-letter r. The river-letter b should not therefore follow the river-letter r in Morbio, so Morbio was probably originally Morvio, the v simply having changed to b at some stage, a fairly common change in Romano-British names. There is no problem with the m and r in Morvio - they correspond to the hill-letters n and m in Camulodono, further up the Aire at Skipton. The river-letter r corresponding to the hill-letter m in Camulodono appears in both names Eller (the tributary of the Aire on which Skipton stands) and Morvio. Those Celts who used the hill-letter n (in the dono element of Camulodono) applied the river-letter l to minor rivers, including tributaries, and the river-letter m to major rivers. We thus see the river-letter l in the tributary, the Eller, and the river-letter m in the major river, the Morvio. Note that the river-letters in each river-name are in the correct order. The element dono comes at the end of Camulodono (indeed it normally appears at the end of place-names) but it is of course an inversion-type element and the corresponding river-letters l and m are used in an inversion-type manner, that is to say the l comes before the r in Eller and the m before the r in Morvio. Turning to the Morbio of the Notitia Dignitatum we can be sure that this fort was between Doncaster and Newton Kyme and on the river Aire. Now there was of course a well-known Roman road from Doncaster to Newton Kyme and that road crossed the Aire at Castleford. It thus seems sensible to presume that Morbio was a late Roman fort at Castleford. The change of name from Flavian Lagentium to Morbio rather suggests that Lagentium (and perhaps its vicus, too) had been abandoned some time before the later fort was built.

The modern river-name Aire is presumably derived from Morvio. The m and v were dropped or lost, thus leaving orio, and this form later changed to Aire.

 

 

MORIDUNO

(Celt)	(Rav)		(AI)		(Mod)
Moriduno	Moriduno (23)		Moriduno		Hembury Hill (Somerset)
			(Iter XV)

Note:

The first element is thought to be from hypothetical British mori, taken to mean 'the sea', and the second element to be Celtic duno, from hypothetical dunos, which is thought to have meant 'hill' originally and to have taken on the meaning 'fort' later. The meaning of the name is thus taken to be 'sea-fort'.

 

The m and r  of Moriduno  are the hill-letters m and r and duno is an inversion-type topographical element meaning 'summit of hill'. By considering in conjunction the order of names in Ravenna, the mileage given in Iter XV for the stretch between Moriduno and Isca at Exeter, and the mileage and location shown on the Peutinger map, it is clear that Moriduno was the hill-fort on Hembury Hill, a location having no connection with the sea. What is not clear is whether the name in Ravenna actually refers to the hill-fort or to a Roman post established inside the hill-fort after the occupants had been forced out, or to a Roman post on lower ground nearby, to which the name Moriduno had simply been transferred by the Romans. One sees the same m, r compound in Maridunum at Carmarthen, and Carmarthen, although not far from the head of the tidal section of the river Tywi, is also some considerable distance from the sea. Note that it is possible that r is the earliest hill-letter in the name so that the place-name may have existed in the Riduno form before the Dumnoni  added their hill-letter m in the inversion-type manner. Note further that the Peutinger map does actually give the form Ridumo.

 

[The entry for Moriduno was last modified on 01 February 2024] 

 

 

MORIONIO

	(Rav)					(Mod)
	Morionio (30)					Norton Camp (Somerset)

Note:

Rivet and Smith see this name as a duplicate of Moriduno. Richmond and Crawford refer to Williams' suggested derivation from hypothetical morion, said to mean 'ants', the place-name then perhaps meaning 'a place infested by ants'.

 

Morionio appears to be a straightforward topographical compound in the hill-letters m and r, though there may originally have been a consonant between the first i and the following o. The order of names in Ravenna appears to indicate that Morionio was between Omiretedertis (Ham Hill/Martock) and Bolvelaunio (Wiveliscombe). There appears to have been only one pre-Roman fort along that alignment and that is the hill-fort known as Norton Camp, at Norton Fitzwarren to the west of Taunton. That hill-fort was presumably Ravenna's Morionio, though it is not clear whether the name as it appears in Ravenna refers to the hill-fort itself, to a Roman post built inside the hill-fort after the locals had been persuaded to leave, or to a Roman fort built nearby, to which the Romans simply transferred the name of the hill-fort.

 

 

MURIDONO  see  MARIDUNUM

 

 

MUTUANTONIS

(Celt)	(Rav)					(Mod)
Mutuandonis or Mutucandonis	Mutuantonis (69)					Hastings (East Sussex)

 

 Note:

Richmond and Crawford appear to have regarded this name as a conflation of a name beginning with Mutu and the river-name Trisantonis. Rivet and Smith thought this possible, but thought it more likely that initial Mu is simply a garbled version of the abbreviation Fl  for Latin Flumen.

Neither of the views expressed above seems likely to be correct. Ravenna has come over from Caleba (Silchester) to Mutuantonis via Anderelionuba (Pevensey), and is about to go on to Lemanis (Lympne). It seems hardly likely that there could be any confusion with the name of a river in West Sussex (the Trisantonis is now called the Arun).  Mutuantonis appears to be a straightforward topographical compound with two inversion-type elements mut and ant both meaning ‘hill high’. It is a compound name with two hill-letters and so refers to a Celtic settlement which existed prior to the Roman invasion. Now, there is high ground on or near the coast all the way from Hastings to Rye, but the only clear structural evidence of a Celtic settlement (so far as the present writer is aware) is that on the East Hill in Hastings, so in the absence of any alternative that settlement on East Hill is here identified as Mutuantonis. The structural evidence takes the form of a rampart built across the eastern end of the hill, and it is thought that the lowest level of that rampart might date back to the first millenium BC. (For detailed information on that rampart see Fradley M. and Newsome S.: “East Hill, Hastings, East Sussex - A Landscape Survey and Investigation”, English Heritage Research Department Report Series no. 35-2008). However, since the settlement was actually on top of the hill it is likely that the Celtic name was Mutuandonis, in which name the old-style qualifying element and means ‘hill summit’. But the ua letter combination is unusual, so it may be that the name was actually Mutucandonis, where Mutuc is an inversion-type element meaning ‘hill high steep’ – it is the same element as the Macat of Macatonion (at Dymock in Gloucestershire) but with the adjectival c and t in reverse order.  Ravenna’s  Mutuantonis, however, was most probably a Roman post, so either the Romans built a post inside the old settlement after evicting the inhabitants, or they built a new post somewhere in the vicinity, that new post simply taking its name from the Celtic settlement.