¹ Jones, A. H. M., The Cities of the Eastern Roman Provinces ² (1971), App. 1, 503–8Google Scholar. Such lists, like so much of imperial documentation, were occasionally published and preserved on inscriptions: for one such, from Ephesus, listing cities under their ethnic titles (bar two) which are arranged under διοικήσεις (Latin conventus, assize districts), see Habicht, C., ‘New Evidence on the Province of Asia’, JRS LXV (1975), 64–91Google Scholar. The Late Roman dioceses, the sphere of authority of the vicarius, in whom authority over provincial governors and wide juridical powers were invested, were of course much larger than the previous conventus. Assizes under the Early Empire are fully discussed by Burton, G. P., JRS LXV (1975), 92–106Google Scholar.

² e.g. historical information in chronicles and epitomes, codification of laws, lists of bishops at councils, poetic catalogues and summaries of earlier works.

³ On Gaul, see Festus, , Breviarium 6 (ed. Eadie, J. W., 1967)Google Scholar, and Amm. Marc, xv, 11, 7–15. Mommsen, discussed the sources for Ammianus in Hermes XVI (1881), 602–36Google Scholar (= Ges. Schr. VII, 393–425). In his edition of the Not. Gall., MGH, Auct. Ant. IX (1892), 553, he assumes Ammianus refers to a Gallic Notitia of the 380's, which differs in some entries from the Not. Gall., but Ammianus's information could well derive from the period in which he was himself in Gaul (355–7): see below nn. 36 and 53. Note that, in Ammianus, Eauze, later metropolis of Novempopulana, is part of Narbonensis I, and Bourges, later metr. of Aquitanica I, is listed under Lugdunensis I. Such fluidity might tend to devalue the status of a provincial secular metropolitan.

⁴ See Bury, J. B., ‘The Provincial List of Verona’, JRS XIII (1923), 149–51Google Scholar, and Chastagnol, A., ‘Notes chronologiques sur l'Historia Augusta et le Laterculus de Polemius Silvius’, Historia IV (1955), 176–80Google Scholar.

⁵ On provincial reorganizations in Gaul, see Jones, A. H. M., ‘The Date and Value of the Verona List’, JRS XLIV (1954), 21–9Google Scholar, repr. in The Roman Economy (ed. P. A. Brunt, 1974), 263–79, and Eadie op. cit. 163–6, with Jones, LRE III (1964), App. III, 381–2.

⁶ Bury, art. cit., 127–48.

⁷ Amm. Marc. XXVIII, 3, 7: ‘et Valentia deinde vocaretur arbitrio principis’.

⁸ Ed. Mommsen, , MGH Auct. Ant. IX (1892), 552–612Google Scholar, and by Seeck, O., Notitia Dignitatum (1876), 261–74.Google Scholar The former is followed here.

⁹ Amm. Marc. XXVII, 6, 4. The name Gratianopolis is first attested at the Council of Aquileia in 381.

¹⁰ The date is controversial. See Chastagnol, A., ‘Le repli sur Arles des services administratifs gaulois en l'an 407 de notre ère’, Rev. Hist. CCXLIX (1973), 34–40Google Scholar, contra Palanque's date of c. 398. The date is associated with that of the Council of Turin in 398 or 417, on which see E. Ch. Babut, Le Concile de Turin (1904) and, most recently, Chadwick, H., Priscillian of Avila (1976), 162–3, n. 4Google Scholar.

¹¹ Rivet, A. L. F., ‘The Notitia Galliarum: some questions’, Aspects of the Notitia Dignitatum (British Archaeological Reports (Suppl. series xv), 1976), 119–41Google Scholar. The issues raised by Professor Rivet's paper formed the initial inspiration for this paper.

¹² Jones, LRE, II, 712.

¹³ Sulp. Sev., Chron. 11, 32: ‘serius trans Alpes Dei religione suscepta’.

¹⁴ cf. the compromise view of Duchesne, L., who does not commit himself to an ecclesiastical Notitia, but accepts the consequences of so doing, Fastes Episcopaux de l'ancienne Gaule 1 (1894), 76Google Scholar: ‘il est vraisemblable que, dès les dernières années du IV^e siècle, chacunes des cités existantes avait son église et son évêque.’

¹⁵ Rivet, op. cit. (n. 11), 122–4. He questions the omission of Carpentras, colonia with Latin rights under the early Empire, with a bishop by 442, the emergence of the minor city of Albi, and the surprising absence of civitates between Narbonne and the Pyrenees. On the secular view, Carpentras had lost civitas status by the late fourth century (or was omitted by accident): her possession of a bishop by 442 is not inconsistent with lack of civitas status, witness the presence of a bishop from the castrum Uzès. Other doubts are raised by the absence of the civitas Caletum, perhaps absorbed by Rouen, and the survival of the unimportant civitas Diablintum. These points deserve attention, but the inconsistencies with the fourth-century ecclesiastical structure appear to the present writer even harder to explain away (see below, pp. 29 f.).

¹⁶ Later revisions or, conversely, failure of a compiler to use up-to-date information are an occupational hazard in the dating and use of Late Roman official lists in general. On anachronism in the Verona List see Barnes, T. D., ‘The Unity of the Verona List’, ZPE XV (1975), 275–8Google Scholar.

¹⁷ Innoc, Ep. 2, of 15 Feb. 404 (PLS xx, col. 469), Seeck, Regesten, 306, that no ordination to take place ‘extra conscientiam metropolitani episcopi’, with Ep. 24 (to Alexander of Antioch) opposing the confusion of ecclesiastical with secular status, ‘non vere visum est ad mobilitatem necessitatum mundanarum dei ecclesiam commutari’. The De Septem Ordinibus Ecclesiae (early fifth century) refers to the recent allocation of episcopalis electio to the metropolitan. For documentation on the hegemony of Arles, see the Epistulae Arelatenses, MGH Epist. III (1892), 1–83Google Scholar.

¹⁸ cf. a letter of Ruricius of Limoges to Caesarius of Arles in 506 (Rur., Ep. II, 33) protesting against this tendency: ‘Quia, si aliis nomen urbium praestat auctoritas, nobis auctoritatem demere non debet urbis humilitas’.

¹⁹ Aug., Civ. Dei XVI, 4. On status, Jer., Ep. 46, 3, 4, ‘totius provinciae gloria metropoli vindicatur’. For corrupt indices, see CTh XV, 1, 14 (of 365) and for further laws CTh XIII, 3, 11 (to the PPO Galliarum, 376), and Novell. Theod. 22, 1, 8 (of 442). For Arles see MGH, Epist. III (1892), Epist. Arel. 8, p. 14, metropolytana, id est, in Arelatensi urbe, incipiant Septem Provinciae habere concilium’.

²⁰ Author of a Commonitorium, CSEL XVI (1888), 205–43Google Scholar.

²¹ Sid. Ap., Ep. VII, 6, 7.

²² Greg. Tur., Glor. mart. 12. Conc. Arel. 1, Concilia Galliae 1 (A.D. 314–506), ed. Munier, CC CXLVIII (1963), 3–25Google Scholar.

²³ Dax, Lectoure, St. Lizier, Pau, Aire, Bigorre, Oloron.

²⁴ On Italian bishoprics of the sixth century and their survival in Byzantine and Lombard Italy, see Duchesne, L., ‘Les évêchés d'Italie et l'invasion lombard’, MEFR XXIII (1903), 83–116Google Scholar, continued in ibid, xxv (1905), 365–99. He suggests that the Lombard invasions led indirectly to a multiplication of bishops owing to the ordination of some ‘bishops in exile’. Many of his conclusions were, and are, controversial.

²⁵ Jer., Ep. 123, 15, 3.

²⁶ Duchesne, , Fastes III, 161, 163, 114, 76, 122, 106 and 130Google Scholar.

²⁷ For details see Courcelle, P., Histoire littéraire des grands invasions germaniques ³ (1964)Google Scholar, and, for an analysis of the distortions of literary evidence as applied to the devastations of the third century, see Whittaker, C. R., ‘Agri Deserti’, in M. I., Finley (ed.), Studies in Roman Property (1976), 137–65.CrossRefGoogle Scholar

²⁸ A bishop of Reims attended the Council of Arles in 314. See above, n. 22.

²⁹ Paul. Nol., Ep. 18, 4, designates the area as ‘terra Morinorum situ orbis extrema’ (Thérouanne, the civitas Morinorum) and ‘in remotissimo Nervici litoris tractu’(Boulogne, possibly Bavai). On the accuracy of Paulinus, see de Moreau, E., ‘St. Victrice de Rouen, apôtre de la Belgica Secunda,’ Rev. belge de phil. et d'hist. v (1926), 71–9CrossRefGoogle Scholar, and compare the activities of Martin of Tours outside his province, Sulp. Sev., Vit. Mart. 16–17 (Trier), 19, 3 (Vienne), 8, 1 (possibly Agen, see ILS 6117 and 6117a and below n. 56).

³⁰ Duchesne, , Fastes III, 130–7Google Scholar.

³¹ Duchesne, , Fastes III, 106–14Google Scholar. On the settlement of the Cambrai region before 406, see Chauvin, L. and Tuffreau, A. in Rev. du Nord LI (1969), 373–91CrossRefGoogle Scholar.

³² See Vetters, H., ‘Zum episcopus in castellis’, Anz. Öst. Akad. Wiss. CVI (1969), 75–93Google Scholar, on alternative seats and castra. I do not believe these can be shown to exist in the fifth century, contra Rivet, op. cit., (n. 11), 122. The presence of bishops from the castrum of Uzès and from the locus of Toulon can be explained on political grounds: the sees were invented to pack the councils of the 440's by Hilary of Arles. For the imperial view of Hilary's activities see Nov. Vol. 17 (of 445).

³³ Duchesne, , Fastes II, 244–88Google Scholar. Greg. Tur., Hist. Franc., x, 31 on bishops of Tours. Conc. Gall., 135–9 (Angers), 150–7 (Vannes).

³⁴ e.g. Pontius Asclepiodotus, CIL XII, 138, with PLRE I, 116. Pontius was praeses in 377. Seeck prints ‘Metropolis ‘before civitas Ceutronum (268).

³⁵ Rivet, op. cit. (n. 11), 123.

³⁶ Leo, Ep. 66, Lectio dilectionis vestrae (PL LIV, col. 884); Seeck. Regesten, 384.

³⁷ For text of dedication, CSEL XXXI (ed. Wotke, , 1894), 173Google Scholar. On the see, Duchesne, , Fastes I, 238Google Scholar. On Tarantasia, ibid. 236–7.

³⁸ Amm. Marc. XV, 11, 12. Ammianus accompanied Ursicinus to Gaul to deal with Silvanus in late 355 (xv, 5, 22) and left with him in autumn, 357.

³⁹ Conc. Gall. I, 61–75 (Riez), 76–93 (Orange), 94–104 (Vaison).

⁴⁰ Conc. Tour. canon 2.

⁴¹ There is difficulty with the identification of the civ. Rigomagensium. Rivet accepts Barcelonnette, but see Duchesne, , Fastes I, 285Google Scholar, and Griffe, E., La Gaule chrétienne II (1966), 120–1Google Scholar.

⁴² Duchesne, , Fastes I, 277–8Google Scholar (bishop at Gap first attested at the Council of Epône, 517); 278–9 (Sisteron: first bishop, John, died in 509, see Vit. Marti, I, 2 (PL 80, col. 27)); 282–3 (Pentadius of Digne at Conc. Agath. 506, but foundation earlier); 283 (Marcellus of Senez at Agde, 506); 283–4 (Glandève, bishop at Council of Orleans of 541).

⁴³ Duchesne, , Fastes I, 285Google Scholar.

⁴⁴ Duchesne, , Fastes I, 286 f.Google Scholar

⁴⁵ Ep. Hilarii 8, 3. Seeck, Regesten 410.

⁴⁶ At the Councils of Arles, 314, and Aquileia, 381. The quarrel of Marseille with Arles over the metropolitan rights of the latter lasted for much of the fifth century and underlies the attempt by Arles to replace Nice, the suffragant of Marseille, with Cimiez, friendly to Arles.

⁴⁷ The diocese of the Five Provinces could also, alas, be known as the Seven Provinces, as on the heading of the Council of Nîme s of 394 or 396 addressed to bishops ‘of Gaul and the Seven Provinces’. The Five Provinces still exist on 29 January 399 (CTh XVI, 10, 15, to the Vicar of the Five Provinces), but may have been united with Galliae by the time of CTh I, 15, 15, to the praetorian prefect, Vincentius, instructing the Vicar of the Seven Provinces to collect tax arrears (18 June 400).

⁴⁸ The definitions of castrum and castellum are discussed, with references, by Rivet, op. cit. (n. 11), App. II, 134–5. For castrum not as a proper name, see Servius on Aen. VI, 775 (‘castrum autem civitas est’); Nepos, Alc. 9, 3; Cic., Tusc. II, 10, 23; Livy XXXIV, 21,2; Dig. XXVII, 1, 17. For the later recollection (with the proviso that memory is not the same as truth) see Isid., Orig. XV, 2, 13, with ibid, IX, 3, ‘castra sunt ubi milites steterunt’, and, for the sixth century castrum Divionense, Greg. Tur., Glor. Conf. 42; cf. the letter of Maurice to Gregory the Great on a letter received from ‘episcopos civitatum et castrorum, quos Langobardi tenere dinoscuntur’. For the elevation of a vicus to a castrum see Greg. Tur., Hist. Franc., v, 5 on the creation of a bishop ‘apud Arisitensem vicum’ in 576 (not on early MS of Not. Gall.) with an addition to an eighth-century MS from Albi of the castrum Arisidensium. For the military stations of the Rhine, see Schönberger, H., ‘The Roman Frontier in Germany: An Archaeological Survey’, JRS LIX (1969), 146–97Google Scholar.

⁴⁹ Sid. Ap., Ep. IV, 25. Duchesne, , Fastes II, 190–5.Google Scholar

⁵⁰ Episcopus civitatis Vindonensium’ (541), ‘episcopus ecclesiae Vindoninsis’ (549)- See Duchesne, , Fastes III, 210–23Google Scholar. The use of civitas here for the castrum of the Notitia illustrates the blurring of the distinction between the two by this date. Note that Ammianus xv, 11, 3, calls the civitas Constantia of the Notitia ‘castra’.

⁵¹ Amm. Marc. XV, 11, 12. The ‘aedificia semiruta’, which Ammianus may well have seen, count for more in his eyes than the other ‘more obscure’ cities of Alpes Graiae et Poeninae.

⁵² Rivet, op. cit. (n. 11), 122.

⁵³ Duchesne, , Fastes I, 303–5.Google Scholar

⁵⁴ Duchesne, , Fastes II, 101–2.Google Scholar

⁵⁵ Amm. Marc. XV, 11, 6–16; Hilary, De Synodis, proem. (PL X, col. 479); Festus, Brev. 6. See n. 5 above. Changes in Viennensis/Quinque Provinciae are confined to the splitting of Aquitanica (one province in Hilary and Ammianus, redivided by the time of Festus) and Narbonensis (still a single province in Festus, redivided by the time of the Council of Aquileia in 381). The term ‘Five Provinces’ refers to a time when both were undivided, but remained in use, even when the number was restored to the Seven of the Verona List.

⁵⁶ CIL XII, 921 a and b = ILS 6117 and 6117a: ‘Cl. Lupicino v.c. consulari Maxime Senonie ob inlustria merita civitas Senonum patrono suo dedicavit’. The second is to the same effect, a dedication from Auxerre, also a city of Lugdunensis Senonia. See PLRE 1, Cl. Lupicinus 5, p. 520. Perhaps to be identified with a client of Martin of Tours, Sulp. Sev., Vit. Mart. 8, 1; see Matthews, John, Western Aristocracies and Imperial Court, A.D. 364–425 (1975), 155Google Scholar. n. 8.

⁵⁷ A praeses of Lugdunensis III, Valerius Dalmatius, received a verse dedication, Mommsen, Ges. Schr. 11, 150–4 = ILS 8987, set up near Mursa in Pannonia. See PLRE 1, Valerius Dalmatius p. 241. For the consularis, see previous note.

⁵⁸ Nesselhauf, H., ‘Die spätrömische Verwaltung der gallisch-germanischen Länder’, Abhl. der Preuss. Akad. der Wiss. (1938), 5–23Google Scholar, esp. 22 on the Notitia and provincial organization.

⁵⁹ See above n. 7.

⁶⁰ AE 1957, 311 = Emerita XXVII (1959), 372–374 (with photograph) = AE 1960, 158, suggesting that ‘Ma …’ should be restored to Ma[ur(etania)]. But see PLRE 1, Antonius Maximinus 9, p. 578–9 supporting Ma[xima) …(?)]. The inscription was found near Jaca in NE Spain, a remote location for an inscription about Mauretania Tingitana.

⁶¹ Sulp. Sev., Chron. II, 49, 7 (PLRE 1 Anonymous 7, p. 1005), and ibid. 50, 7, ‘viro acri et severo’ with Vit. Mart. 20, 4, ‘praefectus idemque consul Evodius, vir quo nihil umquam iustius fuit’. (PLRE 1, Fl. Evodius 2, p. 297).

⁶² CIL XIV, 231, (at Ostia) Fasti, Rossi I, 359–64. On Maximus’ diplomacy, see Piganiol, A., L'Empire chrétien ² (1972), 266–9Google Scholar and Matthews, op. cit. (n. 56), Chs. VI and VII.

⁶³ CTh IX, 36, 1.

⁶⁴ For his earlier service, Amm. Marc., XXVIII, 5, 1 (of 370), and XXXI, 10, 6–10 (of 378). For his incompetence under Maximus see Greg. Tur., Hist. Franc. 11, 9 (quoting Sulpicius Alexander).

⁶⁵ On Maximus and Priscillian, see Chadwick, op. cit. (n. 10), 111 f. On religious policy, his letters are preserved in Coll. Avellana, nos. 39 (to Valentinian 11 ) and 40 (to Pope Siricius). See also Sulp. Sev., Vit. Mart. 20, and Chron. 11, 49–50. For his judgement of Maximus, a man, but for the fact of usurpation, ‘omni vita merito praedicandus’ see Dial. 11, 6,2, a judgement echoed by Oros. VII, 34, ‘vir quidem strenuus et probus atque Augusto dignus, nisi contra sacramenti fidem per tyrannidem emersisset’.