¹ All dates are AD unless otherwise specified. References in brackets without further qualification are to the Roman History of Cassius Dio. The standard edition of his work is that of U.P. Boissevain, Cassii Dionis Cocceiani historiarum Romanarum quae supersunt (Berlin 1895-1931). The Loeb Classical Library edition prints Boissevain's text, with some slight differences in book numbering. The translations are my own.

² Hadrian: Birley, A.R., Hadrian: The Restless Emperor (London 1997) 87-8Google Scholar. Caracalla: Sillar, S., ‘Caracalla and the Senate: The Aftermath of Geta's Assassination’, Athenaeum 89 (2001) 407–23Google Scholar.

³ Birley, , Septimius Severus: The African Emperor (London 1999) 218Google Scholar.

⁴ Crook, J., Consilium Principis (Cambridge 1955) 36Google Scholar. For example, much attention has been paid to the limited impact that regime change had on such appointments in the Flavian and Antonine periods: Waters, K.H., ‘Traianus Domitiani Continuator’, AJP 90 (1969) 385–405Google Scholar; Eck, W., ‘The Emperor and His Advisors’, CAH² 11.195-213 at 197Google Scholar; Grainger, J.D., Nerva and the Roman Succession Crisis of A.D. 96-99 (London 2004) 70–1Google Scholar. These conclusions are supported by the detailed studies of Eck, Senatoren von Vespasian bis Hadrian (Munich 1970)Google Scholar and Alföldy, G., Konsulat md Senatorenstand unter den Antoninen (Bonn 1977)Google Scholar. For the Severan period, see Leunissen, P.M.M., Konsuln und Konsulare in der Zeit von Commodus bis Severus Alexander (180-235 n. Chr.) (Amsterdam 1989)Google Scholar, who likewise emphasises continuity, despite changes in the structure of senatorial careers.

⁵ The extent of Macrinus' innovations has not hitherto been recognised by scholars. Millar, F., A Study of Cassius Dio (Oxford 1964) 161-3Google Scholar, discusses Dio's account of Macrinus' reign, but he does not analyse the emperor's provincial appointments as part of a larger political strategy. Salway, R.W.B., ‘A Fragment of Severan History: The Unusual Career of… atus, Praetorian Prefect of Elagabalus’, Chiron 27 (1997) 127-53 at 147Google Scholar, argues that there was significant continuity between the regimes of Caracalla and Elagabalus.

⁶ The majority of Dio's account of the Severan period has to be reconstructed from excerpts preserved by the Byzantine writers Xiphilinus and Zonaras. However, we are fortunate that a portion of the original text of the Roman History, covering the reign of Macrinus, has survived. The manuscript tradition of Dio's history is discussed by Millar, Cassius Dio (n. 5) 1-4.

⁷ PIR ² 25; CIL 10.5178Google Scholar; CIL 10.5398Google Scholar = ILS 1159.

⁸ PIR ² 224. According to the Historia Augusta (Cara. 6.7), Agrippa was serving as praefectus classis at the time of Caracalla's death, but this would have been an aberration for a senator. See Starr, C.G., The Roman Imperial Navy 31 B.C. - A.D. 324 (Westport 1975) 192Google Scholar.

⁹ PIR² I 566Google Scholar; CIL 3.10473Google Scholar = ILS 1153.

¹⁰ PIR ² A 271. Dio 78.13.3 gives Ms name as Deccius Triccianus, but this is likely to be a textual corruption.

¹¹ For a review of his career, see Fitz, J., ‘Die Laufbahn des Aelius Triccianus’, ActaArchHung 26 (1978) 21-7Google Scholar.

¹² Boteva, D., ‘Legati Augusti pro praetore Moesiae biferioris, A.D. 193-217/218’, ZPE 110 (1996) 239-47 at 246Google Scholar.

¹³ Ibid. 247. For a list of such simultaneous commands, see Potter, D.S., ‘Palmyra and Rome: Odaenathus’ Titolature and the Use of the Imperium Maius’ ZPE 113 (1996) 271-85 at 275–6Google Scholar. Potter cites twenty-two examples of senators who administered more than one province, including some cases from the Danubian region, notably Q. Marcius Turbo Fronto Publicus Severus (PIR ² M 249), legate of Pannonia Inferior and Dacia in 118, and M. Claudius Fronto (PIR ² 874) who controlled Dacia and Moesia Superior in 170.

¹⁴ Boteva, , ‘On the Cursus Honorum of P. Fuscus Pontianus (PIR ² F 496), Provincial Governor of Lower Moesia’, ZPE 110 (1996) 248-52 at 248–9Google Scholar.

¹⁵ A senator named Pontius Pontianus is also attested as governor of Pannonia Inferior under Elagabalus. Boteva, ‘Legati Augusti’ (n. 14) has proposed that the two Pontiani were one and the same man, who was moved by Macrinus from Moesia Inferior to Pannonia Inferior in order to replace Aelius Triccianus. However, on current evidence, Pontianus is not attested as governor of Pannonia Inferior until the reign of Elagabalus: AE (1996) 1247Google Scholar. Since Triccianus was executed by Elagabalus (79.4.3), and suffered damnatio memoriae in the province, it seems best to regard this Pontius Pontianus as a new appointee installed by Elagabalus to replace Macrinus' former associate. Boteva's interpretation is further complicated by the fact that we know of three other men named Pontianus in this period, any of whom could be identical with one of these governors: Pontius Fuscus Pontianus, attested in 227 (AE [1998] 282Google Scholar); the ordinary consul of 238, Pontius Proculus Pontianus (PIR ² P 818); and C. Pontius Pontianus Fuficius Maximus, governor of Bithynia and Pontos in 224/225 (SEG 55 [2005] 1384Google Scholar). See Haensch, R. and Weiss, P., ‘Statthaltergewichte aus Pontos et Bithynia. Neue Exemplare und neue Erkenntnisse’, Chiron 37 (2007) 183-218 at 198-9Google Scholar, for a discussion of these men and their careers.

¹⁶ Sabinus' career is reviewed by Alföldy, , ‘Septimius Severus und der Senat’, BJ 168 (1968) 112-60 at 121Google Scholar, and Peachin, M., Iudex vice Caesaris: Deputy Emperors and the Administration of Justice during the Principate (Stuttgart 1996) 101-6Google Scholar.

¹⁷ For Castinus' relationship with Severus, see Birley, , Septimius Severus (n. 3) 215Google Scholar.

¹⁸ Ibid. 89-107.

¹⁹ Birley, , Marcus Aurelius: A Biography (London 2000) 183–9Google Scholar.

²⁰ Vespasian's revolt was aided by the presence of his ally, T. Iulius Alexander, as prefect of Egypt (Suet. Vesp. 6). See Levick, B., Vespasian (London 1999) 46Google Scholar.

²¹ Most modern accounts emphasise only Macrinus' equestrian supporters: Southern, P., The Roman Empire from Severus to Constantine (London 2001) 56Google Scholar; Potter, , The Roman Empire at Bay, A.D. 180-395 (London 2004) 147Google Scholar; Campbell, B., ‘The Severan Dynasty’, CAH² 12.1–27 at 20Google Scholar.

²² CIL 3.6924Google Scholar, revised by French, D., ‘Notes on Cappadocian Milestones: The Caesarea-Melitene Road’, EA 41 (2008) 125-34 at 127–9Google Scholar. The milestone was originally dated to the year 238 by Townsend, P.W., ‘Sex. Catius Clementinus Priscillianus, Governor of Cappadocia in A.D. 238’, CP 50 (1955) 41-2Google Scholar. However, French has suggested that the emperor mentioned on the milestone was not Ant. [Gordi]ono, but Ant[on]ino, which he restores as a reference to Macrinus' son and co-emperor Diadumenianus. On this interpretation, Sex. Catius Clementinus Priscillianus (PIR ² C 564), cos. ord. 230, would be the son of Macrinus' governor.

²³ PJR ² M 735; AE (1960) 36Google Scholar.

²⁴ Macrinus is styled providentissimus on the following milestones from Noricum: CIL 3.5708, 3.5736, 3.13534, 17.4.1.77, 17.4.1.112, 17.4.1.117, 17.4.1.141, 17.4.1.182Google Scholar; AE(1980) 664Google Scholar; AE (2004) 1086Google Scholar. Unfortunately, the identity of the governor in 217/218 remains a mystery, but he was obviously a dedicated supporter of Macrinus.

²⁵ The standard tenure of senatorial legates was approximately three years: Birley, , The Roman Government of Britain (Oxford 2005) 8CrossRefGoogle Scholar.

²⁶ PIR¹ V 560Google Scholar; CIL 3.6903 = 12163,3.6912Google Scholar.

²⁷ There was sufficient time to make this journey, since Cerialis was probably discharged from Cappadocia soon after Macrinus' death in mid-218. Elagabalus spent the winter of 218/219 in Bithynia, and did not arrive at Rome until September of 219. See Kienast, D., Römische Kaisembelle (Darmstadt 1996) 172Google Scholar.

²⁸ French, , ‘Cappadocian Milestones’ (n. 22) 130Google Scholar.

²⁹ Although the status of suffect consul had declined by the early third century, the post of consul Ordinarius had become even more prestigious: W. Eck and H. Wolff, ‘Ein Auxiliardiplom aus dem Jahre 203 n. Chr.’, in eidem (eds), Heer md Integrationspolitik: Die römischen Militärdiplome als historische Quelle (Cologne 1986) 556-75 at 561-6; Bagnali, R.et al., Consuls of the Later Roman Empire (Atlanta 1987) 3Google Scholar.

³⁰ Kienast, , Römische Kaisertabelle (n. 27) 172Google Scholar.

³¹ Peachin, M., ‘Prosopographic Notes from the Law Codes’, ZPE 84 (1990) 105-12 at 105-6Google Scholar, has suggested that Aelius Ulpianus may have been governor of Syria Coele in 216 (CJ 9.51.1), but the exact dates of his tenure are unknown. It thus remains uncertain whether Agrippinus was a recent appointee.

³² PJR ² F 20. For Agrippinus' ancestry, see Bocherens, C. and Zevi, F., ‘La schola du Trajan et la domus du consul Caius Fabius Agrippinus a Ostie’, Arch. Class. 58 (2007) 257-71 at 268–9Google Scholar.

³³ PIR ² F 19; CIL 13.8050Google Scholar. See Eck, , ‘Ein Ehrenmonument der Centurionen der legio I Minervia für Caracalla, Geta and Iulia Domna’, Bonner Jahrbücher 185 (1985) 41-5Google Scholar. He may also be identical with the Arval Brother [---A]grippinus, attested under Caracalla, though he is not the only possible contender: Scheid, J., Le collège des Frères Arvales: étude prosopo-graphique du recrutement (69-304) (Rome 1990) 448Google Scholar.

³⁴ The name of Caracalla's brother Geta was also expunged from the stone, but this did not occur at the same time as the erasure of Triccianus' name: Eck, , Die Statthalter der germanischen Provinzen vom 1.-3. Jahrhundert (Cologne 1985) 202 n. 5Google Scholar.

³⁵ Millar, , The Roman Near East, 31 B. C - A.D. 337 (Cambridge MA 1993) 145Google Scholar.

³⁶ PIR ² p 403 pica Caerianus is given in the text of Dio, but his name was undoubtedly Caesianus, given the nomenclature of the first-century senator P. Numicius Pica Caesianus (PIR ² N 203).

³⁷ Sartre, M., Trois études sur l'Arabie romaine et byzantine (Brussels 1982) 89Google Scholar, suggested that Caesianus could have been appointed by Macrinus to replace a disloyal legatus Arabiae who appears in the Historia Augusta (Diad. 8.4). Given that the passage also refers to the exile of a dux Armeniae, a position not attested until the fourth century, it should be regarded with suspicion.

³⁸ Leunissen, , Konsuln (n. 4) 268Google Scholar.

³⁹ PIR ² M 318. Secundus' support for Macrinus is noted by Dabrowa, E., ‘The Commanders of Syrian Legions, 1st-3rd C AD’, in Kennedy, D.L. (ed.), The Roman Army in the East (Ann Arbor 1996) 277-96 at 283Google Scholar.

⁴⁰ Datus: PIR¹ V 46Google Scholar. Basilianus: PIR ² I 201Google Scholar. For the failure to recognise Macrinus, see Sijpesteijn, P.J., ‘The Prefect L. Valerius Datus in P. Princ. Inv. GD 7644’, ZPE 65 (1986) 169-70Google Scholar. The dates of office are collected by Bastianini, G., ‘Lista dei prefetti d'Egitto dal 30 al 299’, ZPE 17 (1975) 263-328 at 307Google Scholar; idem, ‘Usta dei prefetti d'Egitto dal 30 al 299: Aggiunte e correzioni’, ZPE 38 (1980) 75-89 at 86.

⁴¹ Millar, , Cassius Dio (n. 5) 163Google Scholar. Senatore had been forbidden to travel to Egypt by Augustus (Tac, . Ann. 2.59)Google Scholar, but this may have changed since Egyptians were themselves admitted to the senate in the Severan period, on which see Bowman, A.K., ‘Egypt from Septimius Severus to the Death of Constantine’, CAH² 12.313-26 at 315-8Google Scholar.

⁴² P. Oxy. 3117; Bastianini, ‘Aggiunte e correzioni’ (n. 40) 86. Salway, ‘Fragment of Severan History’ (n. 6) 140-6 would add the enigmatic … atus, recorded on two fragmentary inscriptions at Rome (CIL 6.41190, 41191Google Scholar) to the list of Macrinus' supporters. However, the course of this individual's career remains uncertain, and Salway's interpretation has not won universal acceptance (see the comments of Alföldy, , CIL 6.41190, 41191 ad loc.Google Scholar). The traditional reading of … atus’ career identifies him with T. Messius Extricatus (PIR ² M 518), as argued by Cébefflac-Gervasoni, M., ‘Apostilles à une inscription de Portas: T. Messius Extricatus et les Saborrarii’, PP 34 (1970) 267-77Google Scholar. On this interpretation, Extricatus was an adherent of Caracalla who fell into political disgrace under Macrinus.

⁴³ Burreil, B., Neokoroi: Greek Cities and Roman Emperors (Leiden 2004) 292-3Google Scholar.

⁴⁴ Price, S., ‘Local Mythologies in the Greek East’, in Howgego, C., Heuchelt, V. and Burnett, A. (eds), Coinage and Identity in the Roman Provinces (Oxford 2005) 115-24 at 123Google Scholar.

⁴⁵ Burrell, , Neokoroi (n. 43) 289Google Scholar.

⁴⁶ Ibid. 292-3.

⁴⁷ PIR² I 182Google Scholar. Asper's power and influence is underlined by the fact that he was patron of no less than six provinces – Britain, the two Mauretanias and the three Spanish provinces – a fact noted on several honorific statue bases from his villa at Tusculum: AE (1997) 261Google Scholar; CIL 14.2508, 2509Google Scholar = ILS 1156. Dietz, K., ‘Iulius Asper, Verteidiger der Provinzen unter Septimius Severus’, Chiron 27 (1997) 483-523 at 514-22Google Scholar, argues that Asper defended clients from these provinces who had supported Clodius Albinus in his war against Septimius Severus.

⁴⁸ The process of sortitio for praetorian and consular proconsulships remained the accepted method of appointment into the third century: a career inscription from Moesia Inferior, dated to the reign of Gordian III, refers to a proco(n)[s(uli) prjov(inciae) Sficiliae] / sortit[o] (AE [1990] 863Google Scholar). For the process of balloting for proconsulships, see Alföldy, , Konsulat (n. 4) 122Google Scholar; Eck, , ‘Beförderungskriterien innerhalb der senatorischen Laufbahn, dargestellt an der Zeit von 69 bis 138 n. Chr.’, ANRW 2.1 (1974) 158-228 at 204Google Scholar; Bowman, A.K., ‘Provincial Administration and Taxation’, ANRW 2 10.344–70 at 347-8Google Scholar.

⁴⁹ Asper's appointment was probably for the proconsular year 217/218: Leunissen, , Konsuln (n. 4) 225Google Scholar.

⁵⁰ PIR ² A 595 Faustus had originally been favoured by Severas, serving as the first official governor of Numidia, a post that he held for five years. See Birley, , Septimius Severus (n. 3) 143-4, 151-2, 195Google Scholar.

⁵¹ PIR ² A 1385. Fronto had been consul Ordinarius in 199, but is not known to have held other provincial commands. This is unsurprising, as he was a member of a particularly eminent aristocratic dynasty, being the son of C. Aufidius Victorinus (PIR ² A 1393), cos, II ord. 183, and the grandson of Marcus Aurelius' tutor, M. Cornelius Fronto (PIR ² C 1364).

⁵² PIR ² A 1016; AE (1971) 490Google Scholar. Rufinus can be dated to the late Severan period on prosopographical grounds. His marriage to Calpumia Ceia Aemiliana is recorded on an inscription from the second quarter of the third century (AE [1995] 1653Google Scholar, with the stemmata in PIR¹ VII.1, 95Google Scholar). His son, Q. Aradius Rufinus (PIR ² A 1013a = R 143), also a governor of Syria, would marry Iunia Aiacia Modesta (PIR ² A 471), the descendant of Q. Aiacius Modestus Crescentianus (PIR ² A 470), cos. II ord. in 228, placing the second generation in the mid-third century.

⁵³ In favour of 217/218: Leunissen, , Konsuln (n. 4) 218-9Google Scholar; Beschaouch, A., ‘Une hypothèse sur la date du vice-proconsulat en Afrique de Q. Aradius Rufinus Optatus Aelianus’, Tituli 4 (1982) 471-4Google Scholar. The case for 238 was made by Birley, , ‘The Governors of Roman Britain’, Epigraphische Studien 4 (1967) 63–102 at 83Google Scholar; idem, The Fasti of Roman Britain (Oxford 1981) 175-6; Rémy, B., ‘La carrière de Q. Aradius Rufinus Optatus Aelianus’, Historia 25 (1976) 458-77Google Scholar.

⁵⁴ As noted by Rémy, , ‘La carrière’ (n. 53) 465-72Google Scholar. See the list of acting governors in Peachin, , Iudex (n. 16) 229-36Google Scholar.

⁵⁵ Millar, , Cassius Dio (n. 5) 160-1Google Scholar. Dio was resident in Rome until May/June 218, when he became curator of Smyrna and Pergamum.

⁵⁶ Sidebottom, H., ‘Herodian's Historical Methods and Understanding of History’, ANRW 2.34.4 (Berlin and New York 1998) 2775-836Google Scholar, provides the best recent assessment of Herodian's work.

⁵⁷ See Millar, , Cassius Dio (n. 5) 150-60Google Scholar, and C. Davenport, ‘Cassius Dio and Caracalla’, CQ (forthcoming).

⁵⁸ Herodian (5.1.4-8) includes the supposed text of a letter sent by Macrinus to the senate in which he justifies the promotion of an equestrian to the purple, but it is clearly the author's invention. See Alföldy, , ‘Zeitgeschichte und Krisenempfindung bei Herodian’, Hermes 99 (1971) 429-449 at 443Google Scholar.

⁵⁹ Adlection inter consulares, although rare, was not unprecedented: in the reign of Augustus, Gaius Cluvius and Gaius Furnius were given the rank of ex-consul when they were unable to assume their consulships (52.42.4), as noted by Birley, , ‘Review of P.M.M. Leunissen, Konsuln und Konsulare in der Zeit von Commodus bis Severus Alexander (180-235n. Chr.), Gnomon 62 (1990) 612-15 at 613Google Scholar. Under Commodus, P. Tarruttienus Paternus received the honour as a dubious ‘promotion’ after being removed from the praetorian prefecture (HA Comm. 4.7), and Septimius Severus adlected Aelius Antipater inter consulares before appointing him governor of Pontus and Bithynia (Philostr. VS 2.24.2).

⁶⁰ The fact that Macrinus was criticised for appointing Ulpius Iulianus and Iulianus Nestor, two former principes peregrinorum, to the praetorian prefecture (78.15.1) indicates that Dio did not have a problem with equestrians, but with the promotion of soldiers.

⁶¹ PIR ² O 9. For his career, see Rankov, B., ‘M. Oclatinius' Adventus in Britain’, Britannia 18 (1987) 243-9CrossRefGoogle Scholar.

⁶² There are several career inscriptions which recorded proconsular appointments extra sortem: CIL 9.2485Google Scholar = ILS 915 (P. Paquius Scaeva); CIL 11.1835Google Scholar = ILS 969 (L. Martius Macer); AE (1925) 85Google Scholar (M. Iulius Romulus). L. Egnatius Victor Lollianus is attested as proconsul of Asia for an extended term of three years (242/3-244/5) under Gordian III and Philip, an appointment that was clearly the result of an imperial decision. See Christel, M., Drew-Bear, T. and Tashalan, M., ‘Lucius Egnatius Victor Lollianus, Proconsul d'Asie’, Anatolia Antiqua 11 (2003) 343-59CrossRefGoogle Scholar; Haensch, R., ‘L. Egnatius Victor Lollianus: la rhétorique, la religion et le pouvoir’, in Vigourt, A.et al. (eds), Pouvoir et religion dans le monde romain (Paris 2006) 289–302Google Scholar.

⁶³ Talbert, R.J.A., The Senate of Imperial Rome (Princeton 1984) 348–9Google Scholar.

⁶⁴ Parkin, T.G., Old Age in the Roman World: A Cultural and Social History (Baltimore 2003) 116-7Google Scholar, who specifically notes the exceptional nature of the interval between consulship and proconsulship in the case of Aufidius Fronto.

⁶⁵ E.g. a monument erected for C. Salvius Nonius Liberalis Bassus recorded that he excused himself from the proconsulship of Asia (CIL 9.33 = ILS 1011). For other declined posts Usted on cursus inscriptions, see CIL 6.41124, 6.41226,14.3610 = ILS 1071.

⁶⁶ Potter, ‘Introduction: The Shape of Roman History: The Fate of the Governing Class’, in idem, A Companion to the Roman Empire (Oxford 2006) 1-19 at 16.

⁶⁷ Lendon, J.E., Empire of Honour: The Art of Government in the Roman World (Oxford 1997) 181–5Google Scholar.

⁶⁸ Cf. Millar, , Cassius Dio (n. 5) 163Google Scholar, who writes of Macrinus' need to act cautiously to avoid provoking the senate.

⁶⁹ This had happened on several occasions under both Severus and Caracalla: Salway, , ‘Equestrian Prefects and the Award of Senatorial Honours from the Severans to Constatine’, in Kolb, A. (ed.), Herrschaftsstrukturen und Herrschafìspraxis: Konzepte, Prinzipien und Strategien der Administration im römischen Kaiserreich (Berlin 2006) 115-35 at 121-4Google Scholar.

⁷⁰ We have already noted in Section 3 the adherence of M. Munatius Sulla Cerealis, cos. ord. 215, and the governors of Syria Coele and Arabia, Fabius Agrippinus and Pica Caesianus.

⁷¹ PIR ² M 308; CIL 6.1450Google Scholar = ILS 2935.

⁷² Rome: CIL 6.1452Google Scholar = ILS 2936; CIL.6.1453Google Scholar = 41186. Ardea: CIL 10.6764Google Scholar.

⁷³ Barnes, T.D., ‘The Composition of Cassius Dio's “Roman History”’, Phoenix 38 (1984) 240-55 at 244CrossRefGoogle Scholar.

⁷⁴ For the post of curator rei publicae, see Burton, G.P., “The curator rei pubhcae: Towards a Reappraisal’, Chiron 9 (1979) 465–87Google Scholar.

⁷⁵ Campbell, J.B., The Emperor and the Roman Army, 31 B.C. -A.D. 235 (Oxford 1984) 52-4Google Scholar; Phang, S.E., Roman Mihtary Service: Ideologies of DiscipUne in the Late Republic and Early Principate (Cambridge 2008) 187-8CrossRefGoogle Scholar.

⁷⁶ This is the best interpretation of several passages of Dio (78.12.7, 78.28.2-3, 78.36.1-3). See Millar, , Cassias Dio (n. 5) 166Google Scholar; Speidel, M.A., ‘Roman Army Pay Scales’, JRS 82 (1992) 87–106 at 88Google Scholar.

⁷⁷ Diadumenianus had previously been made Caesar, but was now raised to the rank of Augustus: Syme, R., ‘The Son of the Emperor Macrinus’, Phoenix 26 (1972) 275-91 at 277CrossRefGoogle Scholar. He was now formally known as M. Opellius Antoninus Diadumenianus Augustus: AE (1968) 591Google Scholar; CIL VIII 4598Google Scholar = ILS 463.

⁷⁸ Marius Secundus seems to have been in Egypt in May 218. This at least explains his failure to check the revolt of the legio III Gallica, which was under his direct command in Syria Phoenice.

⁷⁹ Millar, , ‘Emperors at Work’, JRS 51 (1967) 9–19Google Scholar; idem, The Emperor in Che Roman World (London 1977). Millar's thesis ‘is now the dominant scholarly model of how the Roman empire worked in practice’, according to H.M. Cotton and G. Rodgers, ‘Preface’, in eidem (eds), Rome, the Greek World and the East, Volume 2: Government, Society and Culture in the Roman Empire (Chapel Hill 2004) vii-x at vii.

⁸⁰ Hopkins, , ‘Rules of Evidence’, JRS 68 (1978) 178-186 at 181, 186Google Scholar.

⁸¹ Campbell, , Emperor and the Roman Army (n. 75) 337Google Scholar. There were exceptions, of course: according to Suetonius (Dom. 8.2), Domitian's governors were very trustworthy.

⁸² I would like to thank Dr Brian Jones for supervising the original thesis from which this article emerged and for his many years of support and encouragement. The present paper has benefited from the advice and criticism of Dr Jennifer Manley, Dr Meaghan McEvoy, and an anonymous referee for Antichthon whose suggestions significantly improved the final version.