¹ Sv. de Vita Agricolae (ed. Furneaux and Anderson), p. 143; H. Mattingly takes the passage in a very similar way in his Pelican translation, Tacitus on Britain and Germany, p. 90. Furneaux and Anderson take omni in the ablative case and translate ‘along all the adjoining side.’

² Hübner, , Hermes xvi (1881), 545Google Scholar (Ugrulentum, Anon, Ravennas, 435, 21); Hübner is followed by Reed most recently, but there seems little to recommend an identification of Ugrulentum with Trucculensis. He suggests that both were a deep inlet north of Loch Broom; Britannia ii, 148. Lipsius, see Ogilvie and Richmond (eds.) Tacitus, Agricola 118, 282 (Rutupiae); Anderson and Furneaux, op. cit. 141 (Carpow or Cramond).

³ Annals xv, 44, 1; Hist, iv, 29, 2.

⁴ A. R. Burn, ‘Tacitus on Britain’, Tacitus, Studies in Latin Literature (ed. T. A. Dorey, 1969), 59. The idea is less fully developed in Agricola and Roman Britain 156-7. See also CR xviii (1968), 316Google Scholar. Nicholas Reed follows Burn in his translation, ‘The Fifth Year of Agricola's Campaigns’, Britannia ii, 144, 147–8.Google Scholar

⁵ Tacitus, Agricola 10, 6: et hiems adpetebat.

⁶ O. Seeck, Not. Dig. xl, 51.

⁷ Richmond, I. A. and Crawford, O. G. S., Archaeologia xciii, 10, 36Google Scholar; E. Birley, Research on Hadrian's Wall 224-6; Bellhouse, R. L. and Gillam, J., CW lxvi (1966), 44–5.Google Scholar

⁸ Birley, E., CW lv (1956), 30–45Google Scholar. Professor Birley still favours Moresby for Tunocelum in a letter dated 5 January 1973.

⁹ By Horsley, cf. Richmond and Crawford, loc. cit. (note 7), 36.

¹⁰ Ptolemy, , Geogr. ii, 3, 2.Google Scholar

¹¹ By Professor Frere in a letter of 21 October 1972. Birley notes that the pottery from Kirkbride is of two periods, c. 90-120 and 160-200, with no fourth-century occupation.

¹² Jarrett, M. G., CW lviii (1958), 63–7Google Scholar; lxv (1965), 118. See also xxxvi (1936), 85-99. Dr. Jarrett now considers that the fort was established under Hadrian.

¹³ CW xlviii (1948), 42–7Google Scholar (Moresby); lv (1955), 30-45; lvi (1956), 85-99 (Burrow Walls).

¹⁴ The cohortes classicae are normally treated as units raised from the fleet, which henceforth became normal land army units: Cheesman, The Roman Auxiliaries, Appendix ii, 189; C. Starr, Roman Imperial Navy (1941), 154-5, 188, 204. A. R. Burn, Romans in Britain (1932), 111, suggested that I Aelia Classica may have been raised by Hadrian and planted by him on the line of the Roman Wall c. 122. I am by no means convinced that the cohortes classicae did not still have some naval function. The known cohortes classicae are rather few; might they not have retained some of their former specialized nature? Cohors II Classica was in Syria. I Classica was in Aquitania, then in Lower Germany: K. Kraft, Zur Rekrutierung der Alen und Kohorten an Rhein und Donau (1951), 95-9; G. Alföldy, Die Hilfstruppen der römischen Provinz Germania Inferior (1968), 55 f. Another cohors classica, perhaps a section of the true fleet or one of the above, had been stationed at the naval base of Forum Julii in Provence: PWRE ‘Cohors’ 272-3 (Cichorius). To return to I Aelia Classica in Britain, it was balanced at the opposite end of Hadrian's Wall by the numerus barcariorum Tigrisiensium at South Shields. This unit is normally treated as concerned with ferrying duties across the Tyne estuary and probably up the coast. Units of barcarii certainly had such duties on rivers elsewhere in the Empire: PWRE iii², 2,648-9 ‘Classis’ (Fiebiger). There was a numerus barcariorum at Confluentes and a classis barcariorum at Eburodunum. Finally, the Wroxeter diploma of 135 lists I Aelia Classica and this is immediately followed by … n. Nautarum, this last almost certainly a unit of marines. There are grounds for thinking that I Aelia Classica had a naval function, especially as it was based at TUNOCELUM, which was certainly a coastal fort, and probably had a harbour somewhere near the Eden estuary.

¹⁵ For a parallel to the ethnic form sv. CAPITULUM-CAPITULENSIS in Lewis and Short. A further objection to RUTUPIAE is the fact that the normal adjective was RUTUPINUS and was well known in Latin literature. Lucan vi, 67; Juvenal iv, 141; Ausonius, Parentalia 18, 8.

¹⁶ De Vita Agricolae (eds. Furneaux and Anderson) sv. 38, 4; app. crit. p. 29 (conjectures by Madvig).

¹⁷ For the pluperfect in the sense of an action completed in the past, Francis and Tatum, Latin Syntax 54; Roby, , Latin Grammar ii, 200Google Scholar (an action subsequent to another past action).