[NB. The reader will probably find it easier to understand this page if he or she has already acquainted himself or herself with the contents of Chapters 1 and 2. Chapter 1 concerns the basic building-blocks used in Romano-British place-names and Chapter 2 explains how those building blocks were assembled by the Celts to form compound names.]
(S)aranus to Navimago Regentium
Roman place-names in central southern England
1 Ravenna starts its list of place-names in this part of the country with Aranus (34), though it would appear that the s at the end of the previous name, Alavna Colonea(s), is really the first letter of (S)aranus, or (S)armis/(S)aramis if one prefers the reading of Rivet and Smith to that of Richmond and Crawford. Whatever the spelling of the name the place concerned was at Old Sarum, just outside Salisbury. Note that the river-letter corresponding to the hill-letter s in Saranus is b, and this appears, changed to v, in the river-name Avon – this river was Ptolemy’s Alaunus (= Alavna). It has long been assumed that Sorviodunum/Sorbiodoni was at Old Sarum, but this was never more than an assumption, an assumption which made it quite impossible to reconcile the mileages quoted in the Antonine Itinerary Iter XV with the actual distances on the ground. But once one has located (S)aranus at Old Sarum the next few Ravenna names all fall into place. Ibernio (37) was clearly the fort at Hod Hill, the name surviving in Iwerne, the modern name of the river flowing past the fort. It can thus be seen that Anicetis (35) will have been at Ansty, to the west of Old Sarum and Melezo (38) at Melbury, to the south of Shaftesbury. Bindogladia thus turns out to have been the hill-fort called Weatherby Castle. It may be that this short list of names, (S)aranus to Bindogladia, was taken from an early Claudian map and documents Vespasian’s advance when he was sent by the then governor to subdue the tribes of the southwest. Anicetis is an inversion-type name in the hill-letter n, nicet meaning ‘hill steep high’. Similar name-elements exist in other hill-letters, for example lucot in Lucotion (170) and Macat in Macatonion (61)). Note that Anicetis was most probably the then name of the Castle Ditches hill-fort about a kilometre NNE of the modern village of Ansty. That hill-fort was built around the summit of a steep, high hill.
2 By following the distances given in the Antonine Itinerary one can see that it was in fact Sorbiodoni that was at Badbury Rings. With the above identifications of Sorbiodoni and Bindogladia one can see that the Iter XV distances for the stages from Sorbiodoni to Bindogladia and Bindogladia to Durnonovaria at Dorchester are in fact correct. Note that Bindogladia comprises two old-style elements, the first, bind, meaning ‘high hill summit’ and the second, glad, meaning ‘steep hill summit’, so the name is eminently suitable for Weatherby Castle. It is of course likely that when the Romans built the road from Badbury Rings to Dorchester they built a post on that road at the foot of the hill crowned by Weatherby Castle, and that they simply transferred the name of the hill-fort to that new post. This new post would be the Vindogladia of the Antonine Itinerary, the original b of the name having changed to v. After Bindogladia Ravenna lists Noviomagno (39), Onna (40) and then Venta Velgarom (41) at Winchester. It is thus clear that Ravenna has used Old Sarum as a node, with one group of names on an alignment to the west-southwest out of Old Sarum and another on an alignment to the east-southeast. Noviomagno will have been at Maiden Castle. The hillfort itself was probably just called Magno, this being an old-style compound in the hill-letters m and n, the corresponding river-letters r and m (for major rivers) being seen in the name of the river Frome, which flows west-east just a little to the north of Maiden Castle. Noviomagno was probably a new settlement founded just north of Maiden Castle and to which many of the inhabitants of Magno will have been moved by the Romans. At some stage, perhaps even after the founding of the new Roman town at Dorchester, the name acquired the ending aria, the new town then being Duronoviomagnoaria (the prefix Duro appears to have been used here and elsewhere in the sense of ‘Roman new town at or near’), this later being shortened, understandably enough, to Duronovaria.
3 Onna was then most probably between Old Sarum and Winchester, perhaps at the crossing of the river Test, around Horsebridge. Note that there was a Celtic settlement a little to the east of Horsebridge (at SU 394 301) on the fairly steep slope of Ashley Down. A name of the form Conva would be entirely appropriate for that location (the old-style element conv means ‘steep hill slope’). But there was also an Iron-Age enclosure up on top of the ridge at Fir Hill, Bossington, on the western side of the river Test, just west of Horsebridge (for a report on excavations at Fir Hill see Brown, 2009). For that location a name of the form Conda would be appropriate (the old-style element cond means ‘steep hill summit’). The change from Conda/Conva to Onna is straightforward. Ravenna omits the initial letter or letters of a name on a number of occasions, as for example in (Lu)coganges, (L)itucodon and apparently in (C)armis, (C)ardaoneon and (C)arduaravenatone. The change from nd to nn is seen in Gabaglanda → Amboglanna, and the change from nv to nn is seen in Banva → Banna. What is not clear is whether the name Onna in Ravenna actually refers to one of those two settlements or whether it refers to a Roman fort built to control a crossing of the river Test, the name then being transferred from one of the two Celtic settlements to the Roman fort.
4 All of the names from (S)aranus to Venta refer to Celtic settlements, in most cases to a hillfort, so it is probable that this pattern is maintained in the names after Venta. Navimago Regentium will refer to the hillfort known as The Trundle or to Chichester, to which the name of the hillfort, together with its inhabitants presumably, was transferred. Navimago will earlier have been Navimagno, essentially the same name as Noviomagno discussed above. It is most likely, however, that The Trundle hillfort was just called Magno and that the Romans transferred the inhabitants of the hillfort to a new settlement on substantially the same site as the later Romano-British town at Chichester. That new settlement will have been Navimagno, presumably the same name as Noviomagno discussed above. Note that an existing native settlement (albeit one only recently founded) was later simply romanised, so it kept the same name, Navimagno/Navimago. Had it been a new Roman town on a different site it would probably have been called Duronavimago. It is highly probable, then, that Armis and Ardaoneon were hillforts between Winchester and The Trundle. We can exclude Old Winchester Hill since that hillfort is apparently associated with the name Clausentum in Iter VII of the Antonine Itinerary. The other hillforts between Winchester and The Trundle are Butser Hill, Harting Beacon, Torberry and Goosehill Camp. On purely geographical grounds it seems sensible to identify Butser Hill as Armis. As to Ardaoneon we can rule out Goosehill Camp since that stands on a steep slope and so its name would not include the old-style element ard which means ‘hill summit’. That leaves Harting Beacon and Torberry. We seem not to have any dating evidence for Harting Beacon but are told (see, for example, the Pastscape website of Historic England) that a timber entrance gate to Torberry was destroyed around 100 BC, and that this hillfort was only used sporadically after that. Of course this does not rule out the possibility that it was occupied at the time when the Romans advanced into that area, but on balance it seems safer to identify Harting Beacon as Ardaoneon. And of course we don’t know whether in giving the names Armis and Ardaoneon Ravenna was actually referring to the hillforts themselves, or to Roman posts perhaps built inside the hillforts, or even to Roman posts built in the vicinity of the hillforts and to which the names of the hillforts were transferred. Note that the hillforts on Butser Hill and Harting Beacon stand at the top of steep slopes, so it is probable that the names of the hillforts were actually (C)armis and (C)ardaoneon. For some reason Ravenna omits the initial letter(s) of a Celtic name on a number of occasions, as noted above.
5 Ptolemy has only two names in this region, Venta at Winchester and Noeomagus in the territory of the Regni, the latter clearly corresponding to Ravenna’s Navimago Regentium.
6 The Ravenna names discussed in this chapter are shown in the lower half of the map reproduced below. The numbering of the place-names is that of Richmond and Crawford. The identifications of the place-names are given below.
(S)aranus Old Sarum Onna Horsebridge
Anicetis Ansty Venta Velgarom Winchester
Melezo Melbury (C)armis Butser Hill
Ibernio Hod Hill (C)ardaoneon Harting Beacon
Bindogladia Weatherby Castle Navimago Regentium Chichester
Noviomagno Maiden Castle
[This page was last modified on 19 September 2019]