¹ Britannia ii (1971), xvi–xvii (jointly by Professor S. S. Frere and the present writer).

² Antonine Itinerary 475.1; mortarium mark DVROBRIVIS, JRS xxx (1940), 190.Google Scholar

³ Britannia i (1970), 72–3Google Scholar; Rivet and Smith: The Place-Names of Roman Britain (1979), 346–8 (hereafter referred to as Rivet and Smith).

⁴ VCH Hunts i (1926), 228–48Google Scholar; for recent work, Durobrivae: a Review of Nene Valley Archaeology 1 (1973) – 7 (1979).Google Scholar

⁶ The fort was first observed in 1930 (Antiquity iv (1930), 274–5)Google Scholar and first photographed in 1939 (Antiquity xiii (1939), 178–90Google Scholar; see also JRS xliii (1953), pl. IX. For the bridge, J. Ward: The Roman Era in Britain (1911), 35Google Scholar; Dymond, D. P., Arch. Journ. cxviii (1963), 157.Google Scholar

⁶ e.g. Pliny, , HN III, 28Google Scholar on the conventus of Lucus, composed of ‘15 peoples who, apart from the Celtici and Lemavi, are unimportant and have barbarous names (ignobilium et barbarae appellationis)’ and on the conventus of Braga, where only six of the 24 peoples can be named ‘without being boring (citra fastidiuni)’. The Greeks and Romans found no romantic appeal of the Chimborazo-Cotopaxi kind in barbarous names. Their Utoplas (Atlantis, Panchaea, etc.) always had classical names and these prevailed even when one was supposedly identified, as with the Hesperides (Plutarch, Sertorius ix); one possible exception is ultima Thule.

⁷ M. Gelling, Signposts to History (1978), 236.

⁸ For example, some Normans (including Duke William himself) already had connections in the island, but the same might also have been true of Roman administrators, at any rate those of Gaulish origin, like C. Iulius Classicianus (from the Treveri). At a lower level it is impossible to tell how many Roman citizens acquired estates in the new province.

⁹ The point was well made by Claudius in A.D. 48 (Tacitus, , Annals XI, 24).Google Scholar In southern Gaul it was natural for Pompeius Trogus, as a Vocontian, to inflate the concept of Gallia in Graeciam translata (Justin XLIII, iv, 1), but modern research has made it virtually certain that it was not until the mid-second century B.C., and then with Roman assistance, that Massalia acquired a consolidated territory of any size fas opposed to isolated strongpoints and trading posts); for an excellent summary of the question see Barruol, P.: Les peuples préromains du sud-est de la Gaule (Paris, 1963), 221–30.Google Scholar On the other hand it was precisely the undiluted purity of its Hellenism that made the city a satisfactory alternative to Athens as a finishing school (Strabo IV, 1,5 (C. 181); Tacitus, Agricola 4).

¹⁰ Traditionally the site was acquired as a dowry from king Nannus of the Segobrigii (Justin XLIII, iii, 8–12; Athenaeus XIII, 5 (576)).

¹¹ Dauzat, A. and Rostaing, C.: Dictionnaire des noms de lieux en France (Paris, 1963)Google Scholar, s.v. Céreste; Barruol, op. cit., 211.

¹² Benoit, F., Recherches sur l'hellenisation du midi de la Gaule (Paris, 1965), 99 ff.Google Scholar; Barruol, op. cit., passim.

¹³ Stephanus, Ethnica, s.v. Tauroeis.

¹⁴ e.g. around the Black Sea (information from Dr Aleksandra Wasowicz).

¹⁵ The earliest example of such a hybrid (and that a Latin-Celtic one) seems to be Brutobriga (βρουτοβρια) in Stephanus, Ethnica, but unlocated), apparently founded in Spain by D. Junius Brutus c. 133 B.C.; see Knapp, R. C.: Aspects of the Roman Experience in Iberia 206–100 B.C. (Valladolid, 1977), 19.Google Scholar

¹⁶ Kikuyu (a tribal name) was first a trading station, then (1899) a railway station, later extended to include the area around it, including the site of Fort Smith (which still existed when the railway first came): Hardy, R., The Iron Snake (London, 1965)Google Scholar, R. O. Preston, ‘The Genesis of a Colony’, Kenya Graphic, New Year Issue 1922, 22. During the Second World War the name Fort Wilkinson was revived for an airfield near Todenyang.

¹⁷ The most curious survival (even in world atlases) is the name Port Victoria. The bay here was originally intended to be the western terminal of the railway, but Port Florence (now Kisumu) was eventually chosen instead. No town has ever existed at Port Victoria and it thus offers an analogy for Ptolemy's surveyed, but apparently never developed, ports in Britain, Portus Setantiorum and Sinus Gabrantuicorum, on which see Rivet and Smith, op. cit., 116.

¹⁸ The road was actually made by George Wilson at Sir William's expense; Gregory, J. W.: The Great Rift Valley (London, 1896), 206–7.Google Scholar

¹⁹ Preston, op. cit., 12–13. This somewhat obscure article, by an engineer who played an important part in the construction of the line, offers the fullest account of such aspects of it.

²⁰ It is worth noting, for comparative purposes, that this is far from being the most important bridge in this area, which is that over the Tsavo river at Tsavo. The original bridge at Darajani, from which the station presumably took its name, was a temporary one (Harding, op. cit., 154).

²¹ Alike in energy, effectiveness and popularity: S. M. Kirov, assassinated by L. V. Nikolayev near Leningrad, 1934, at the age of 30, was buried in Red Square; the ashes of Germanicus, supposedly assassinated by Piso and Plancina near Antioch, A.D. 19, at the age of 33 or 34, were deposited in the mausoleum of Augustus.

²² Barruol, op. cit., does not discuss Martius and for Apollinares (218) contents himself with drawing attention to the important temple at Riez.

²³ Strabo iv, 2, 3 (C. 191).

²⁴ Ausonius, Commemoratio Professorum Burdigalensium iv (Attius Patera; the temple of Belenus).

²⁵ ‘Augusta Ambianorum’ (Eu) is in a different category. Though there are substantial remains here, the name is not attested before the seventh century (without the suffix) and the capltal of the civitas was Samarobriva (Amiens).

²⁶ Rivet, A. L. F. in Duval, P.-M. and Frézouls, E. (edd.), Thèmes de recherche sur les villes antiques d'occident (Paris, 1977), 179.Google Scholar

²⁷ Seneca, Apocolocyntosis 8, 3, etc.

²⁸ On this aspect see Mann, J. C., ‘City Names in the Western Empire’, Latomus xxii (1963), 777–82.Google Scholar

²⁹ Rix, H., ‘Zur Verbreitung und Chronologie einiger Keltischer Ortsnamentypen’ in Tübingen Festschrift für Peter Goessler (Stuttgart, 1954), 99–107Google Scholar; Piggott, S., Ancient Europe (Edinburgh, 1965), 172–4.Google Scholar

³⁰ Durotincum is especially curious: Barruol, op. cit, 138, sees a Ligurian element in -tincum, and the name, which appears only in Peutinger (Durotinco) and the Ravenna Cosmography (Durotingo), may be a rationalization. Durostorum might be a ‘secondary’ name (see below) conferred by a Gaulish auxiliary unit before the establishment there of Legio XI Claudia in A.D. 106.

³¹ This peculiarity was noticed some years ago, e.g. by Zupitza, E., Zeitschrift für celtische Philologie iv (1903), 17 f.Google Scholar, and Weisberger, L., ‘Die Sprache der Festlandkelten’, in Bericht der Römisch-Germanischen Kommission xx (1930), 185 ff.Google Scholar; I am grateful to Dr J. P. Wild for these references.

³² e.g. Rodwell, W., ‘Coinage, Oppida and the Rise of Belgic Power in South-Eastem Britain’, in Cunliffe, B. and Rowley, T. (edd.), Oppida in Barbarian Europe, B.A.R.S. 11 (Oxford, 1976), 181–359.Google Scholar

³³ B.G. vi, 44. The suggestion that ‘the anomalous placing of duro- in most British names in which it occurs may reflect a Latin manner of forming place-names’ (Gelling op. cit., 52) therefore seems improbable.

³⁴ C. Cornelius Gallus, Cassius Dio liii, 23, Suetonius, Divus Augustus lxvi; Sallustius Lucullus, Suetonius, Domitian x, 3.

³⁵ Léon is always simply Legio or Legio Septima, in all sources, and no other name for it is known.

³⁶ Haberl, J. and Hawkes, C. F. C., ‘The Last of Roman Noricum’, chapter viii in C. F. C., and Hawkes, S. (edd.), Creeks, Celts and Romans (London, 1973)Google Scholar; I am grateful to Professor Hawkes for drawing my attention to this. See also Alföldy, G., Noricum (London, 1974), 259.Google Scholar

³⁷ For discussions see Hassall, M. W. C. in Goodburn, R. and Bartholomew, P. (edd.), Aspects of the Notitia Dignitatum, B.A.R.S. 15 (Oxford, 1976), 113Google Scholar; Breeze, D. and Dobson, B., Hadrian's Wall (London, 1976), 274–5Google Scholar: Rivet and Smith, op. cit., 223.

³⁸ N.D.Occ. xl, 34.

³⁹ The fort has not yet been identified, but the site has yielded material of a sufficiently early date (information from Mr A. Scott Anderson). Professor Hawkes has pointed out that the reference is most likely to be to a period after the conquest of the Cornovii, but auxiliary units were sometimes raised outside the Empire, as in Thrace before its incorporation as a province in A.D. 46 (Webster, G., The Roman Imperial Army (London, 1969), 55, with references).Google Scholar

⁴⁰ Ravenna Cosmography 105. 48 (No. 6 in Richmond and Crawford). Duplications in the Cosmography are much more common than Richmond and Crawford allowed (see Rivet and Smith, op. cit., chapter v), but this interpretation opens the possibility of the name's referring to a Roman fort (as opposed to an Iron Age hill-fort) in the south west; but it would also destroy the evidence for Cornovii in Cornwall at such an early date.

⁴¹ For further particulars see Jackson, K., ‘Romano-British Names in the Antonine Itinerary’, Britannia i (1970), 68–82Google Scholar and Rivet and Smith, op. cit., Alphabetic List of Names, 237–509.

⁴² Professor S. S. Frere informs me that there are now some indications of a Roman fort; see Britannia viii (1977), 424.Google Scholar

⁴³ TQ 994610. Hawkes, C. F. C., ‘Britain and Julius Caesar’, Procs. British Academy lxiii (1967), 162, n. 2.Google Scholar

⁴⁴ Hasted, E., History and Topographical Survey of Kent ii (1782), 800Google Scholar; V.C.H. Kent i (1908), 400–401Google Scholar; size, Archaeologia Cantiana ix (1874), lxxivGoogle Scholar; deplcted on O.S. 6-inch 1906. I am grateful to Mr A. Clarke of the Ordnance Survey Archaeology Branch for detailed information on this site. It is not accepted as a hill-fort by A. H. A. Hogg in his definitive list, British Hill-Forts, B.A.R. 62 (Oxford, 1979).Google Scholar

⁴⁵ Crawford, O. G. S. and Whiting, W., JRS xii (1922), 275Google Scholar; Whiting, however, appears to have changed his mind (Arch. Journ. lxxxvi (1929), 300)Google Scholar and reverted to a Roman interpretation.

⁴⁶ On Caesar, Hawkes, op. cit. (note 43), 164–5, with references; on the Claudian invasion, Cassius Dio ix, 20, where the river can only be the Medway.

⁴⁷ For the probability that the Durcinate of Ravenna Cosmography 106. 52 (Richmond and Crawford No. 100) is a corrupted form of this name see Rivet and Smith, op. cit., 351.

⁴⁸ This suggestion was first made by Mr G. C. Boon.

⁴⁹ For example, as noted above, the element -dunum always appears last in Romano-Celtic times, but later, apparently from the 6th century onwards, Dun- becomes a first element the noun followed by the defining adjective instead of vice versa; on this see Jackson, K. H., Language and History in Early Britain (Edinburgh, 1953), 225–7Google Scholar, and Gelling, op. cit., 52, and on Celtic survivals Gelling, chapter 4.

⁵⁰ cf. the passages from Plinycited in note 6, also Cicero, Pro Fonteio 33 (Gauls strolling in the forum cum quibusdam minis et barbaro atque immani terrore verborum). But see also Sherwin-White, A. N., Racial Prejudice in Imperial Rome (Cambridge, 1967)Google Scholar, especially 17–18: ‘Caesar certainly knew how to exploit prejudice in this field.… He is thoroughly enjoying himself in this exercise of misrepresentation, and does not believe a word of it. But for all that, he knows that there are plenty in his audience who will believe.’