^{page 86 note 1} L'Organisation de l'Empire rom. (Fr. trans, of Staatsverwaltung) ii, pp. 70-1; E.E. iv. pp. 224-5.

^{page 86 note 2} pp. 55, n. 1, 87, n. 2.

^{page 86 note 3} iii, 4, 20—he does not use the word dioecesis, which seems to be only once applied to Spain, in an uncertain sense, in a late Greek inscription (E.E. iv, p. 223=Dessau 8842). Modern writers adopt it for Spain on the analogy of Africa. It was applied to much larger units in the late empire. Cicero (Fam. iii, 8, 4Google Scholar; xiii, 67, 1) and Strabo (xiii, 4, 12) use the word as=conventus in Asia.

^{page 86 note 4} Plin. 19, 35; 31, 24. cf. Plin. Ep. iii, 5, 17Google Scholar.

^{page 87 note 1} Dessau 2454-5, 2644. Ritterling's difficulty (de Leg. X Gem. p. 26) that the last inscription may refer to the legions VI Ferrata and X Fretensis in Syria is not serious, since there is no reason to suppose that these were closely associated there.

^{page 87 note 2} Dessau 2648—an indication at least that Leon was a legionary camp in pre-Flavian times, if not under Augustus.

^{page 87 note 3} Braun, , Die Entwicklung der Spanischen Provinzialgrenzen (Berlin 1909), pp. 31–5Google Scholar, puts the original boundary rightly between Asturia and Cantabria, but he extends the second diocese along the Pyrenees as far as the Mediterranean, because of an ancient view that the Pyrenees ran north and south. Such a boundary would have been impossible in practice. The conventus Cluniensis touches the W. end of the Pyrenees and so satisfies Strabo's ‘μέχρι Πυρήνης’ Kornemann counts Asturia an original part of the second diocese.

^{page 87 note 4} Plin. Ep. ii, 14, 9Google Scholar.

^{page 88 note 1} Dessau 1011; Gsell, Domitien, p. 140 n.; Domaszewski, Rb. Mus. 46, p. 599; Suet. Vesp. 13.

^{page 88 note 2} Ritterling (J.O.A.I. 1907, pp. 299 ff) first published this explanation of these subordinates in the eastern province, and he refers to Largius as a parallel. He wrote later than the other authorities. My conclusion was formed from the Spanish side, independently of his, if that is any confirmation.

^{page 88 note 3} v, 6974 ff. with Mommsen on 6987; Dessau 1021, 1021A; P.I.R. G. 114. Domaszewski, believing that a separate iuridicus for Asturia-Callaecia was instituted in 88, thinks that Glitius' title implies that he was sole iuridicus, but not that he was competent for the whole province. The peculiar order of the inscr. of Aquae Flaviae (ii, 2477) ‘D. Cornelio Maeciano leg. Aug. L. Arruntio Maximo proc. Aug. | Leg. VII Gem. Fel.’ might suggest that Maecianus was a successor of Largius as iuridicus, only that the leg. legionis could hardly have been omitted. The order may rather be explained by ref. to Strabo's statement connecting the proc. closely with the military service; Arruntius is bracketed with the legatus as a legionary officer and may thus have been proc. of Asturia-Callaecia only, rather than for the whole province as suggested (p. 87, with n. 1 there). Truttedius Clemens (ii, 2643) was probably one of his successors before the officers referred to on p. 86.

^{page 88 note 4} xii, 3167 = Dessau 1016; xiii, 1802.

^{page 88 note 5} ‘Hispania Ulterior Lusitania’ is so styled only once, so far as as I can find, in extant records (vi, 18190).

^{page 88 note 6} viii, 14291 = Dessau 1096. Liebenam dates his Spanish office about 150.

^{page 88 note 7} Dessau 8975, viii, 2747 = Dessau 1070.

^{page 89 note 1} v, 6987; xii, 3167; ii, 2959; Plin. Ep. ii, 14, 9Google Scholar.

^{page 89 note 2} Stabo iii, 4, 20, αὐτὸς δὲ ὁ ἡγϵμὼν διαχειμάζει μὲν ὲν τοῖς ὲπιθαλαττιαίοις μέρεσι, καὶ μάλιστα τῇ Καρχηδόνι καὶ τῇ Ταρράκωνι δικαιοδοτῶν, θέρους δὲ περίεισιν, ἐφορῶν ἀεί τινα τῶν δεομένων ὲπανορθώσεως.

^{page 89 note 3} Caes. B.C. ii, 18, 21Google Scholar; B. Alex. 51, 56; Dio xliii, 14, 29-31, 40; Florus iv, 2, 75-6; Caes. B. Hisp. 37, 40.

^{page 89 note 4} Florus iv, 12, 49; Tac. Ann. iv, 5Google Scholar; Plut. Ant. 68; Fiebiger, de Classium Italicarum historia (Leipziger Studien xv), pp. 292–3Google Scholar.

^{page 89 note 5} p. 100, n. 1. Even the fleet at Ravenna seem, by A.D. 6 to have been too weak to secure the Adriatic (Dio lv, 29).

^{page 90 note 1} Vita Marci 21, 1; Dessau 1354, 1354A. Calpurnius Siculus iv, 40 ‘trucibus obnoxia Mauris pascua Geryonis’ records raids in Nero's reign—the earliest reference. See Mommsen, , Provs. ii, p. 324Google Scholar, n. 3, and n. 5 on next page.

^{page 90 note 2} So ii, 4114 = Dessau 1140, with n. in addenda, is now read. Mommsen l.c. followed by Cagnat, L'armée rom. d'Afrique, p. 275, n. 2, refers this to another Moorish invasion. It may rather have been the last struggle of the friends of Albinus in the wars of succession (Vita Severi 12)—the rebellion was against Severus, not against Rome. Cf. p. 99.

^{page 90 note 3} ii, 4063. Cf. Livy xxii, 19.

^{page 90 note 4} Tac. H. iii, 43Google Scholar; Hirschfeld in xii, pp. 38-9, and nos. 257-8. One of these records a M. Aurelius—not necessarily of the second cent., but even so hishe may belong to Misenum.

^{page 90 note 5} Tac. H. ii, 16Google Scholar; Mommsen in x, pp. 777, 787, and no. 8329; E.E. viii, 709-12, 800-1; Fiebiger, op. c. pp. 329-332.

^{page 90 note 6} Cagnat l.c. makes no ref. to the praef. orae in discussing the Mauretanian fleet. He is recorded only once in an inscription which is fragmentary and not exactly datable (xi, 5744). My suggestion (p. 61, l. 5) of a joint control of that fleet must be withdrawn, first because it was not independent, but only a detachment from Syria and Egypt, and secondly because it was probably too small (Cagnat, pp. 283-4) for anything but local duties. Also its head-quarters were at Caesarea (Cherchel), too far away for convenient co-operation. Such as it was, it is possible to attribute its creation to Vespasian (Cagnat, p. 278) though it may be earlier.

^{page 90 note 7} Pliny (p. 60, n. 2) indicates that the praef. orae Ponticae had few troops in his service, unlike hishe Spanish colleague.

^{page 90 note 8} p. 60, n. 4.

^{page 91 note 1} Ptol. ii, 6, 18; Strabo iii, 4, 7-8—Tarraco is ἀλίμενος …. Ἐρατοσθένης δὲ καὶ ναύσταθμον ἔχειν φησὶν αὐτὴν, ούδέ ἀγκυροβολίοις σφόδρα εὐτυχοῦσαν, ὡς ἀντιλέγων εἴρηκεν ᾿Αρτεμίδωρος καὶ ἡ σύμπασα δ᾿ ἀπὸ Στηλῶν σπανίζεται λιμέσι μέχρι δεῦρο, ἐντεῦθεν δ᾿ ἤδη τὰ ἑξῆς εὐλίμενα καὶ χώρα ἀγαθὴ τῶν τε Λεητανῶν.…καὶ ἄλλων τοιούτων μέχρι Ἐμπορίου τοιούτων μέχρι‘ Εμπορίου. Livy xxii, 22, on Tarraco, hardly weakens this testimony.

^{page 91 note 2} ii, p. 599, Hübner. Mela ii, 6, 5.

^{page 91 note 3} ii, 4508, 4536-48; Martial i, 49, 40, vii, 47; Victor, Ep. 13, 5; ii, 4509 = Dessau 1029. Natalis was cos. suff. in 106 with the younger Granianus (x, 5670) who had been specially honoured by the ordo of Baetulo, another coast-town of Laeetania, in Domitian's reign, ii, 4609=Dessau 1028. For Iluro, another of its towns, see p. 81, n. 5.

^{page 91 note 4} ii, 4264 = Dessau 2716.

^{page 91 note 5} Some hint of the probable system of land defence in Spain against sea-raids is given by Livy xxn, 19, 6-7 (cf. Plin. 11, 181; xxxv, 169) for earlier times at least. But Livy's words ‘Multas et locis atlis positas turres Hispania habet quibus et speculis et propugnaculis adversus latrones utuntur. Inde primo conspectis hostium navibus datum signum Hasdrubali est…’ seem to be of Permanent application also. The cohortes tironum might have been employed.

^{page 91 note 6} The harbour of Emporiae is better attested than any other except Carthago Nova on the Mediterranean coast of Hither Spain, but it was not included in Laeetania and was also too far north to have been head-quarters.

^{page 92 note 1} ii, 4225 = Dessau 2714.

^{page 92 note 2} ii, 4239, recording a praef. orae who had been previously praef. alae Thracum Herculaniae in the East, is dated first cent, by Cichorius in P.W. i, 1263. Only 4266 = Dessau 2717, presumably the latest inscr., records Coh. II (indicating some extension of the system)—unless it is implied in 4264 = Dessau 2716, which names Coh. Prima, not Nova as before. All the praefecti were men of local distinction.

^{page 92 note 3} Op. c. (n. 7 supra), p. 98.

^{page 92 note 4} Müller on Ptol. denies it: Hübner in ii, p. 912, makes Avila = an ancient Avela, ignoring any other identification. On his authority we have counted the town in Citerior, p. 75; so Sprüner-Sieglin's map. Kiepert in Formae Orbis Antiqui now marks it in Lusitania.

^{page 92 note 5} iii 4, 20; iii, 3, 3, where the clause ὑπενατίως δὲ τοῖς νῦν ἔνιοι καὶ τούτους Λυσιτάνους ὀνομάζουσιν should so be read with Kramer, not τὰ νῦν as Meineke emends, τούτους probably refers to Καλλαικοί only. But our point is not affected whatever be the reading and interpretation adopted, Sprüner-Sieglin (tab. 29) assigns to ‘Hisp. Ulterior Lusitania’ beforethe death of Agrippa (12 B.C.) all the Vettones, Vaccaei, and Astures (but not the Carpetani) perhaps on the authority of this pasasge. Yet if Agrippa's measurements of his Lusitania are at all correct (Plin. 4, 118—540 by 536 miles) it must have included the Carpetani and more. Cf. Strabo iii, 4, 12.

^{page 93 note 1} Plin. 25, 84 ‘Vettones in Hispania’ is ambiguous (p. 85, n. 6). Id. iii, 6, is vague. The reading ‘Ocelenses qui et Lancienses’ which Sillig keeps in 4, 118, would strengthen the view that the town was distinct from Ocelodurum; ‘Ocelenses Lancienses’ may be similarly explained, and it is arbitrary to omit ‘Lancienses’ with Mayhoff against the MSS.

^{page 93 note 2} Salmantica is Vaccaean in Polybius and Livy.

^{page 93 note 3} ii, 857-8-9, 5033.

^{page 93 note 4} Plin. 4, 118, counts in Lusitania ‘Mirobrigensis qui Celtici cognominantur’ (so Ptol.) and, as he does not mention Polibedenses, presumably (p. 77) they like the other two places were outside the province. We assume that the Mirobriga of the inscriptions is the same as Pliny's and Ptolemy's.

^{page 94 note 1} ii, 460, Ptol. counts Lanc. Opp. in Vettonia; therefore Pliny's Lusitanian Lancienses (4, 118) must be referred to the Lancienses Transcudani, a separate commune, as marked by Sprüner-Sieglin, though he counts it in Vettonia. Thus Igaedi, not mentioned by the geographers (cf. p. 73), must be an instance of a commune organised by Augustus in his later years. The position of Mirobriga as given in the maps (conjecturally) is somewhat too far north for our view.

^{page 94 note 2} The epithet Val which is attached to the names Mirobriga, Salmantica, and Bletisa in ii, 875-8, has not been satisfactorily explained—see Hübner and Mommsen in ii, p. 107, and Ursin p. 827. Our argument suggests as at least possible the view that Val=Vallares or the like. It would follow that some kind of vallum was at least projected as an inter-provincial boundary—perhaps like the fossa regia in Africa. But there is no other trace of it, and it would be strange that anything so elaborate should be contemplated for that purpose in Spain.

^{page 94 note 3} ii, 656. ii, 862, is a dedication at Bletisa (Ledesma) to Domitian in 82 A.D. ‘decreto decurionum’: the occasion is not stated, but may be connected with the reconstruction in progress.