¹ The name Corstopitum has been hallowed by the early excavation reports of 1909–14, Arch. Ael. vi (1910); xiiGoogle Scholar (1915) and by long general usage as seen in the guidebooks and on the site itself. See for example, Corbridge Roman Station (Corstopitum), Official Guide (1935) (1954); O.S. Map of Roman Britain (3rd. ed. 1956); E. Birley, Research on Hadrian's Wall (1961), 149; P. Salway, The Frontier People of Roman Britain (1965), 44; and most recently Roger Wilson, A Guide to the Roman Remains in Britain (1975), 250; O.S. Map of Roman Britain (4th ed. 1978), North Sheet.

² E. Ekwall, Concise Oxford Dictionary of English Place-Names (4th ed. 1960), 122. Early variants are Corebricg (c. A.D. 1050); Colebruge (A.D. 1100–7); Corebrigge (A.D. 1158). An early form of Corchester is Colchestre (A.D. 1394), but this is unlikely to represent a true etymology of the name. It perhaps represents dissimilation of the ‘r’ to ‘1’.

³ Corbridge Roman Station…, 18; Woolley, C. Leonard, Arch. Ael. iii (1907) (for the bridge)Google Scholar; Arch. Ael. iv (1908), 205–14Google Scholar, pi. Ill; site plan on 217; JBAA xii (1906), 202.Google Scholar The general lay-out of the site is perhaps most conveniently to be seen in R. E. Collingwood, Roman Britain (1934), 74.

⁴ Richmond, I. A., Arch. Ael. xxi (1943), 4Google Scholar; Salway, P., Arch. Ael. xxxvi (1958), 227 ff.Google Scholar; E. Birley, Research on Hadrian's Wall (1961), 132–50; P. Salway, The Frontier People of Roman Britain (1965), 45 ff.; D. J. Breeze and B. Dobson, Hadrian's Wall (1976), 20–6.

⁵ G. Parthey and M. Pinder, Itinerarium Antonini Augusti (1848), 222; 464. 3; Cuntz, O., ltineraria Romana i (Leipzig, 1929).Google Scholar The MS variants are CORSTOPILUM, COSTOPITUM, COSTOPITU, COR STOPITU.

⁶ ‘The form Corstopitum is impossible in Celtic and must be corrupt’, Williams, Ifor, Archaeologia xciii (1949), 30.Google Scholar This view is followed by Collingwood, R. G. and Wright, R. P., Roman Inscriptions of Britain i (1965), 368Google Scholar, where it is called ‘this corrupt form’. See also Jackson, K., Britannia i (1970), 71.Google Scholar A. R. Burn was not so sure that the letter combinations were impossible for Celtic, but he could only point to Textoverdi, more properly Tectoverdi, A. R. Burn, Romans in Britain (1969), 128.

⁷ Schnetz, J., Itineraria Romana ii (1940)Google Scholar; Crawford, O. G. S. and Richmond, I. A., Archaeologia xciii (1949), 19, 30.Google Scholar

⁸ For Coriosopites see Holder, A., Alt-celtischer Sprachschatz i (1928), 1127Google Scholar, and for the relatively numerous references to the better attested tribe of Coriosolites see Pauly-Wissowa, Realencyclopa'die iv. 1235 (Ihm). The reference in Notitia Galliarum iii. 7 is also probably to Coriosolites and not to Coriosopites (so Mommsen argued), O. Seeck, Notitia Dignitatum (1876), 264 n. 7. The Vita Ronani 4 gives in ecclesiam Coriosopitensem, Holder, A., Alt-Celtischer Sprachschatz i, 1127.Google Scholar The most recent etymological dictionary of French names derives Corsept tentatively from cors | etum, Dauzat, A. and Rostaing, Ch., Dictionnaire des noms de lieu de France (Paris, 1963), 214.Google Scholar

⁹ JRS vi (1966), 223, pl. X. 7.Google Scholar

¹⁰ J. Moreau in Dictionnaire de geographie historique de la Gaule et de la France (1972), 95. I am grateful to Professor Smith for drawing my attention to this item.

¹¹ According to Moreau, Quimper was called Coriosopitum only from the late Middle Ages. The view expressed in the text is a tentative one put forward by Professor Smith in a private communication, but he adds the rider that, ‘even if the form with p resulted from a copying error, it still had sufficient real existence at one time to produce Corsept’.

¹² Corionotatae (RIB 1142). These were, according to Frere, Britannia (1967), 55, probably a sept of the Selgovae, but they could have been an early band of raiders from Ireland, where Koriondoi were near neighbours of the Hibernian Brigantes in E. Ireland. The Corionotatae are linked with Corbridge by T. Charles-Edwards in Antiquitates Indogermanicae, a study of ‘Native Political Organization in Roman Britain’ (Innsbruck, 1974), 38. I owe this reference to Professor Smith.

¹³ The inscription from Beltingham, West of Hexham in the valley of the South Tyne was first studied by Stevens, C. E., Arch. Ael. xi. 4 (1934), 142–5Google Scholar, who collected a number of references from Britain and the Continent to Coria and Curia. For the location of the sept of Textoverdi, see Frere, S. S., Britannia (1967), 55Google Scholar and map opp. p. 1. Rivet and Smith in their forthcoming book of the place-names of Roman Britain treat this name as Tectoverdi.

¹⁴ For Coritiotar, see Archaeologia xciii (1949), 30Google Scholar, and for Ptolemy's Kuria of the Otadinoi see Geographia ii 3. 10, in Muller, C., Geographia Claudii Ptolemaei (Paris, 1883), 96.Google Scholar A. H. A. Hogg in Aspects of Archaeology (1951), 200–20; I. A. Richmond (ed.), Roman and Native in North Britain (1961), 56, 77 ff.

¹⁵ The continental *corio names are to be found in A. Holder, op. cit. i, 1127 (Coriosedenses, modern Collias, dep. Gard); i, 1127 (Coriovallum, modern Heerlen); i, 1200, (Curia Raetorum, Chur in Switzerland); ii, 978–80 (Pétrucorii= Perigueux); ii, 1950 (Tricorii).

¹⁶ Williams, Ifor, Archaeologia xciii (1949), 29–30.Google Scholar

¹⁷ Archaeologia xciii (1949), 28Google Scholar, but there it is an unknown and unexplained place-name on the line of the Antonine Wall. Rivet and Smith in their forthcoming book identify Cibra in Ravennas as the Koria Damnoniorum or Dumnoniorum of Ptolemy.

¹⁸ The corruption may have occurred through the use of abbreviations in the MS, such as Cor. Votad, as Professor Smith suggests.

¹⁹ Archaeologia xciii (1949), 19Google Scholar, 55. 140–2. Jackson, in JRS XXXVIII (1948)Google Scholar, 56, agreed with Stevens that Coria would mean ‘hosting-place’ hence ‘tribal centre’ but he objected to deriving the two name forms Kuria and Koria from the same root-word. Now C. Smith argues that there is some close connection between the two forms, and that Kuria may have undergone some Latinization to approximate to the Latin word Curia, s.v. Curia Dumnoniorum in the forthcoming book on Romano-British place-names.

²⁰ Northumberland County History x (1914), 8–9Google Scholar; A. R. Burn, Agricola and Roman Britain (1953), map on front end-paper; Stevens, C. E., Eng. Hist. Review lii (1937), 199–200.CrossRefGoogle Scholar J. S. Stevenson argued against the derivation from *corio on the grounds that the modern name should have been developed into Cher not Cor, North. County History x, 9, 774, but in Brittany *Corio leads to Corseul and Corsept.

²¹ Crawford and Richmond, op. cit. 12.

²² S. S. Frere, Britannia (1967), 55, n. 1 and map opp. p. 1.

²³ Crawford and Richmond, op. cit. 11, 32, 43.

²⁴ Camden and his followers sited Epiacum at Ebchester, Horsley put it at Hexham, Ptolemy (ed. C. Muller, 1883), 96 n.; A. L. F. Rivet, Town and Country in Roman Britain (1958), 142.

²⁵ Crawford and Richmond, op. cit. pi. viii.

²⁶ Of Ptolemy's nine Poleis in Brigantia (ii. 3. 10) three are listed in Ravennas at §§ 110–12 as Alicuna, Camulodono, Caluvio (almost certainly Ptolemy's Olikana, Kamunlodunon, Kalagon), three are to be found at 134, 136–7 as Vinovia, Catabactonion, Eburacum (Ptolemy's Vinnovion, Katurhaktonion, Eburakon). Two are more doubtful Dixiolugunduno (140) probably incorporates Rigodounon, and Corielopocarium is the name or names under discussion. This leaves Isurion (= Ant. Itin. Isurium or Isubrigantum, Itinera i, ii, v) as the only Brigantian polls not appearing even in a corrupt name-form in the Ravenna Cosmography. For Isurium, which is known to have been at Aldbrough near Boroughbridge, see Britannia i (1970), 40, 42, 47, 75.Google Scholar

²⁷ I. A. Richmond, Roman and Native in North Britain (1961), 140 n. 1., Corsley in Wiltshire and two Corstons in Somersetshire and Wiltshire are said to be derived from the river-names Corsan and Corsaburna. There is also Corse in Gloucestershire, derived directly by Ekwall from corsen (pi. cors)— ‘reed’, Ekwall, Dictionary…, 123. Fo r the plural, meaning ‘swamp’, ‘marsh’ in Welsh (y Gorse), Williams, Ifor, Archaeologia xciii (1949), 30Google Scholar s.v. Corsula.

²⁸ For the Pictish place-name element pett, see F. T. Wainwright, Archaeology and Place-Names and History (1962), 70; C. Thomas, Britain and Ireland in Early Christian Times A.D. 400–800 (1971), 44–7.

²⁹ A. Longnon, Les noms de lieu de la France (1920–9), 49, §§ 102–4. Several of these ritu names are to be found in Ptolemy Augustoriton (ii. 7. 9), Anderedon (ii. 7. 11), Danoriton or Darioiton (ii. 8. 6), Lokoriton (ii. 11. 14), Vagoriton (ii. 8. 7), Bandriton {Tabula Peutingeriana). Longnon, op. cit, 49 derives Chambor d from Camboritum and Niort from the Carolingian Noiortum, earlier Novioritum. For the numerous names in *ritu on the continent see now D. Ellis Evans, Gaulish Personal Names (1967), 249–51. See also Holder, op. cit., ii, 1195.

³⁰ For Maporiton and Tadoriton see Schnetz, J., ltineraria Romana ii. 231Google Scholar; Crawford, and Richmond, , Archaeologia xciii (1949), 40, 46.Google Scholar

³¹ For Carbantium, Carbantoriton see Crawford and Richmond, op. cit., 27. For Camboricum and Durolitum as *ritu names, see Jackson, K., Britannia i (1970), 20, 73.Google Scholar The word in O.W. was rit, moder n Welsh rhyd. Nennius gives us a site of a battle between Britons and English in the fifth century A.D. as taking place at Episford, or ‘in our language’ Rithergabail, Lot, F., Nennius et UHistoria Brittonum (Paris, 1934), 184.Google Scholar The Roman name for Pevensey — Anderidos = Anderitori) contains the same element, Jackson, K., Language and History in Early Britain (Edinburgh, 1953), 257 n. 1Google Scholar; 272. It is also found in the modern Penrith, ibid. 555.

³² Woolley, L., Arch. Ael. iii (1907)Google Scholar, esp. plan on 168. E. Birley in the guidebook (see n. 1) stresses that Corbridge controlled one of the best crossings of the Tyne, and that there were excellent foundations in the bed of the river for a bridge. This would apply equally t o its virtues as a ford.

³³ For ‘Heerford’ see Ekwall, Dictionary…, 236.

³⁴ For the heroic defence of lines of communication at fords in early Celtic literature see Ifor Williams, Canu Aneirin (1938), A. 17; B. 17. K. Jackson, The Gododdin The Earliest Scottish Poem (1969), 105, 122. For the defence of Rhyd Forlas on the Welsh marches by Gwen, youngest son of Llywarch Hen, see Ifor Williams, Canu Llywarch Hen (1935), 1; The Beginnings of Welsh Poetry (ed. R. Bromwich, 1972), 138–9, 150. Other famous fords in Scotland are that across the Clyde, the site of a famous battle cat yn rit Alclut- ‘battle of Dumbarton Rock Ford ‘and the ‘fords of Frew’ Rit Gwerid- over the Forth, W. J. Watson, History of the Celtic Place-Names of Scotland (1923), 52, 349.

³⁵ Coriovalium, Coriovallum (Heerlen in Holland ) is discussed by C. J. Guyonvarc'h in Ogam xvii (1965), 351–4Google Scholar, and in Apulum vi (1967), 119–22.Google Scholar

³⁶ A. Mawer, The Place-Names of Northumberland and Durham (1920), 52–3. Mawer mentione d the suggestion made by Chadwick, but himself preferred to link the name with Coriosopitum (Corsept-on-Loire) and Coriosolites (Corseul). It is perhaps worth stating here that Professors Rivet and Smith will treat of this name under the heading Coriosopitum, a suggestion made in the nineteenth century and adopted by Loth, J., Vemigration bretonne en Armorique de ve au vii siecle de notre ere (Paris, 1883), 60Google Scholar, but regarded as doubtful by Jackson, Language and History in Early Britain (1953), 646 n.i.

³⁷ The running together of two unfamiliar names is common in the Ravenna Cosmography. In the British section alone one can find Anderelionuba (68) where Anderida has been conflated with another name following it, and Ventacenomum, where Venta Icenorum (Caistor-by-Norwich) has been conflated and corrupted (103), Archaeologia xciii (1949), 18, 23, 48.Google Scholar

³⁸ The Times October 27, 1974; Current Archaeology 46, September 1974 (published January 1975), 325–9. The editors suggest that the site of the first Roman base was moved from Red House Farm to avoid the flood plains (haughs) of the R. Tyne, which are quite wide at that point, to the site at Corchester which had low hills hemming in the river. Across these from south to north ran the line of Dere Street. See also CBA Region 3, 9th January 1975, 2–3 and Britannia vi (1975), 230–1.Google Scholar