¹ It is a pleasant duty to acknowledge the help that I have received from my friends and colleagues, Mr. J. A. Crook, Mr. J. E. Jones, Mr. F. H. Thompson and Dr. G. D. B. Jones, who kindly read this paper in typescript and made numerous suggestions for changes and improvements. I am also indebted to Dr. A. R. Birley for several helpful observations. None of these kind readers is in any way responsible for the shortcomings that remain.

² RE XII, s.v. legio, 1608–9. Earlier Hasebroek had briefly expressed the opinion that the division must be later than 197; see Hasebroek, J., Untersuchungen zur Geschichte des Kaisers L. Septimius Severus (Heidelberg, 1921), 100.Google Scholar

³ Arch. Ael. 4th series XI, 1934, 131–7.

⁴ CAH XII, 36 f., though he tentatively accepted Herodian's date.

⁵ E.g. Barbieri, , Epigraphica XIV, 1952, 9Google Scholar; Pflaum, , Les procurateurs équestres sous le haut empire romain (Paris, 1950Google Scholar; hereafter Pflaum, Procc.), 155, though in his Les carrières procuratoriennes équestres (Paris, 1960; hereafter Pflaum, Carr.), II, p. 640, Herodian's date is given with some qualification; Steer, , in Roman and Native in North Britain, edited by Richmond, I. A. (London, 1958), 91.Google Scholar

⁶ E.g. Wright, , JRS LI, 1961, 192Google Scholar, who writes ‘after the division into two provinces about 213’. In this he is following Birley in his list of the legates of Britain, published as Appendix I to the Coinage of Roman Britain, by G. Askew (London, 1951), which is also, regrettably if inevitably, deprived of references and arguments.

⁷ See ILS 1157: ‘Iunoni regina[e], pro salute [a]c im[perii] diuturnitate [imp.] M. Aurelli Antonin[i] Pii Fel. Aug. et Iulia[e] Piae Fel. Aug., matri[s] Antonini Aug. ca[s]trorum senatus ac patriae, C. Iul. Cerealis cos., leg. Aug. pr. pr. pr. H(ispaniae) n(ovae) c(iterioris) Anton[i]nianae, post divission. provinc. primus ab eo m[issus]’.

⁸ See note 3 above.

⁹ See note 4 above.

¹⁰ LV, 23, 2, 3 and 6.

¹¹ As Miller noted, CAH XII, 37, n. 7. F. Millar has suggested that we can calculate closely the time of composition not only of Dio's history as a whole, but also of each book: see A Study of Cassius Dio (Oxford, 1964), 28–31Google Scholar and the resulting table, 193 f. The argument seems objectionable on general grounds, firstly, because it assumes in the author a regular output of history per annum and, secondly, because it requires the belief that Dio did not, as a rule, go back over his work or re-write (cf. Appendix III, 208–10, where the passages collected seem, in fact, to show that Dio could, as one would expect, do this). But the steps by which Millar reaches his conclusion are also questionable, since they depend on a forced interpretation of what Dio writes in LXXII, 23.

¹² CIL VIII, 5180.

¹³ CIL VIII, 2080.

¹⁴ The decisive connection between the number of legions and the rank of the governor was clearly brought out by Domaszewski, , Die Rangordnung des römischen Heeres (Banner Jahrbücher Heft 117, Bonn, 1908), 173–182.Google Scholar

¹⁵ See JRS XI, 1921, 101–107 (= AÉ 1922, 116).

¹⁶ RIB 745; the appearance of a ‘beneficiarius cos. provinci[a]e superior[is]’ at Greta Bridge seems to show that it lay outside his area of command; see, e.g., Huebner, notes to CIL VII, 280.

¹⁷ RIB 1738.

¹⁸ ILS 2762.

¹⁹ l.c. (n. 3 above), 132.

²⁰ As by Birley and Miller, for instance, locc. citt., though Atkinson thought it very uncertain; see JRS XII, 1922, 71, n. I. A terminus ante quern might also be thought to be provided by the interesting lead sealing from Combe Down near Bath; see RIB, notes to 179. Its inscription, P BR S, is almost certainly to be expanded ‘P(rovinciae) Br(itanniae) S(uperioris)’; see Wright, R. P., Trans. Cumb. and West. Ant. and Arch. Soc. n.s. LIV, 1955, 102 f.Google Scholar; and it was associated with a building which was restored under Caracalla (or possibly Elagabalus); see RIB 179. But obviously one cannot press this connection to yield an exact date for the sealing. Even less can any exact date be assigned to the similar lead sealing from York, inscribed ‘P(rovinciae) B(ritanniae) I(nferioris)’; see Wright ibid.

²¹ RIB 1280.

²² CIL XIII, 3162, but see the edition by Pflaum, , Le Marbre de Thorigny (Bibliothèque des Hautes Études, Fasc. 292, Paris, 1948)Google Scholar, where the circumstances behind the text are worked out. It is clear that the Concilium Galliarum at which Sollemnis' action saved Paulinus from an accusation was the first held under his successor Aedinius Julianus (Pflaum, p. 2), and Paulinus' letter from Britain, of which a copy is given on the left face (II) of the stone, together with the presents it enumerates, will have been his immediate reaction to the news of the service rendered him by Sollemnis (cf. Pflaum, 33). On the promotion among praetorian provinces see Domaszewski, l.c.

²³ Advanced by Hasebroek, o.c. (n. 2 above), 96, and generally accepted; see, e.g., Groag in Ritterling, E., Fasti des römischen Deutschland unter dem Prinzipat (Vienna, 1932), 76 f.Google Scholar; Miller, CAH XII, 12; Barbieri, , Albo Senatorio (Rome, 1952), no. 528.Google Scholar

²⁴ LXXV, 5, 4 (Exc. U^R 18 (p. 414); where Huebner's reading Παρθικῷ seems far superior to the παροίκῳ of the MSS., in spite of its defence by Boissevain, Cassius Dio III, p. 346, and Hasebroek, o.c., 104.

²⁵ RIB 637, 730. The reading of II. 1–2 of the former inscription is difficult. In addition to RIB see the supplements proposed by Simpson, G., Britons and the Roman Army (London, 1964), 47.Google Scholar For her suggestion that Lupus was the legate responsible for reconstruction at Caernarvon, , see Archaeologia Cambrensis CXI, 1962, 113Google Scholar; but it has no epigraphical support. The possibility that he carried out building work at Corbridge rests on a very heavily restored inscription, RIB 1163.

²⁶ JRS LI, 1961, 192, no. 4 (= AÉ 1962, 260).

²⁷ See ILS 2618 = RIB 1234, RIB 1909, 722, and perhaps 1462.

²⁸ In addition to those in the preceding note, see RIB 1151 (the name is wholly restored, and we now know that ‘C. Valerio Pudente’ could also stand in the gap); RIB 740, 746.

²⁹ On the chronology of Senecio's governorship and career see below pp. 101–2.

³⁰ I discuss the governors more fully below pp. 101–3.

³¹ l.c. (n. 3 above), 135 f.

³² III, 14, 9.

³³ This is often assumed; e.g. by Birley, ibid. and in his list of legates (see n. 6 above); but I doubt if it is right. When the imperial family were in Egypt the praefectus Aegypti was not thereby deposed, even though the emperor sat in judgment at the conventus. See, most notably, P. Col. 123, especially 11. 17 f., 49–51, published in Apokrimata, by Westermann, W. L. and Schiller, A. A. (New York, 1954)Google Scholar, corrected by Youtie, and Schiller, , Chronique d'Égypte XXX, 1955, 327 ff.CrossRefGoogle Scholar Here the emperor delegates business to the prefect. Obviously an Augustus (Geta became this in 209; see, e.g., RIC IV, I, Geta no. 67) who was in a province, could take over whatever business he liked from his legate, but it seems most unlikely that he would replace the legate and himself undertake all the day-to-day business transacted by him.

³⁴ See ILS 8841; I discuss the chronology below, p. 103.

³⁵ o.c. (n. 3 above), 135.

³⁶ ILS 1151.

³⁷ See below, p. 103.

³⁸ See ILS 478; cf. Pflaum, Carr. II, no. 237, who discusses the whole career and makes a good case for dating the procuratorship of Britain immediately after the defeat of Clodius Albinus; it must at latest have been soon afterwards.

³⁹ See Pflaum, Carr. III, 990 (Addenda to p. 628, n. 7); however, if he is right in this view, can he also be right in thinking that the procuratorship of Syria lost position as a result of the division (628, n. 7)?

⁴⁰ As Birley requires, o.c. 136.

⁴¹ ILS 1396; see Pflaum, Carr. no. 295.

⁴² Procc. 148.

⁴³ CIL XIII, 3162, III, 1–3; cf. n. 22 above.

⁴⁴ As Miller, , CAH XII, 37Google Scholar.

⁴⁵ Roman Britain and the English Settlements, 2nd edition (Oxford, 1937), 173Google Scholar.

⁴⁶ Cf. the section on consular governors in Domaszewski, o.c. 175–182.

⁴⁷ RE XII, s.v. legio, 1609.

⁴⁸ His theory required that the procuratorial province included a legion, VI Victrix at York, which seems too anomalous to accept, even though he also postulated that the legionary commander would be an equestrian prefect, like those in Severus' new province of Mesopotamia.

⁴⁹ See above, nn. 27 and 28.

⁵⁰ See n. 3 above.

⁵¹ Thus he, surprisingly, did not object to Ritterling's hypothesis about the Sixth Legion (above n. 48), but this is perhaps understandable since he was rejecting the theory as a whole.

⁵² o.c. 136 f.

⁵³ Trans. Cumb. and West. Ant. and Arch. Soc. n.s. LIII, 1954, 52–62.

⁵⁴ See p. 59.

⁵⁵ Arch. Ael. 4th series XVI, 1939, 240–3; RIB 1462.

⁵⁶ RIB 2283.

⁵⁷ Although he says that the dedication by a procurator for the safety and health of Caracalla ‘found beyond the Irthing’ (ILS 9317 = RIB 2066) has a formula consistent with the procurator also being praeses, as a straightforward dedication by a procurator, called simply ‘[pr]oc. Aug. n.’, the inscription can hardly rank as a separate, positive, argument.

⁵⁸ The matter is treated fully by Stein, A., Die Reichsbeamten von Dazien (Diss. Pannonicae no. 12, Budapest, 1944), 54.Google Scholar

⁵⁹ Compare, e.g., CIL III, 1377 with CIL III, 7647 (Stein, o.c. 59 f.).

⁶⁰ Stein gives examples, o.c. 54. The titles of the very numerous inscriptions of Q. Anicius Faustus in Numidia illustrate the desire at this period to emphasize the title ‘cos.’ in addition to, and sometimes at the expense of, ‘leg. Aug. pr. pr.’; see the useful table given by Pflaum, Libyca V, 1957, 66–73.

⁶¹ JRS LI, 1961,190, no. 1 (= AÉ 1962, 258). R. P. Harper has proposed that ‘su[b A]ṛ[ad]io Rufino’ be read in the third line, in place of ‘su[b A. T]ṛ[iar]io Rufino’ suggested earlier; see Anat. Stud. XIV, 1964, 166Google Scholar. His reasons seem convincing, if not absolutely decisive, and it will not be denied that the command in Britain fits Q. Aradius Rufinus' known career much better than the little we know of that of the cos. ord. of 210. In which case, as Harper shows, the inscription from Reculver cannot be earlier than the 220's.

⁶² o.c. (n. 53 above), 59.

⁶³ Earlier observers were sure that ‘MP LIII’ was the sole inscription the stone had ever borne (till modern times); see, e.g. Watkin, , Arch. Journ. XXXI, 1874, 353Google Scholar (whom Huebner follows, EE III, 119) and Collingwood, , Trans. Cumb. and West. Ant. and Arch. Soc. n.s. XXV, 1925, 367 f.Google Scholar Birley, however, (apparently forgetting that he has used the absence of the emperor's name as an argument against a second-century date) suggests the possibility that Geta's name was originally on the stone and that ‘MP LIII’ is a secondary inscription after its predecessor had been ‘almost completely erased’ (l.c. 61). If this were so, the possibility that the present inscription implies that Carlisle was a provincial capital would depend on our belief that it was inscribed after Geta's death in 212 but before the end of the procuratorial province, dated by Birley in 213. Since this would make the hypothesis even more shaky, it is perhaps best to ignore this rather audacious suggestion about a possible earlier inscription and erasure.

⁶⁴ Though if we are to use them, some milestones from Britain suggest that the emperor's name was not omitted in the Severan period; see RIB 2266, 2298, 2313.

⁶⁵ E.g. Collingwood, ibid.

⁶⁶ As Birley noted, l.c. 55.

⁶⁷ Having stated the two possibilities, he then expands them by admitting, in addition to the provincial capital, ‘command headquarters’. Thus Chester and Wroxeter are regarded as possible capita viarum in the earlier period, as well as London. Whether or not this is valid (there is no evidence for it in Roman Britain), it makes no difference to the present argument, for, as he points out, the praefectus alae at Carlisle is unlikely to have been responsible for so great an area.

⁶⁸ One milestone at least, measured, it seems, from a cantonal capital, but without its name, is known in Britain; see below, n. 81.

^68a Though the argument has been made less decisive by the newly discovered milestone at Brocavum, which bears the name of the ‘R(es) P(ublica) C(ivitatis) Car(vetiorum)’: see JRS LV (1965), 224 (no. II). The milestone was published after my paper was complete.

⁶⁹ CIL VIII, 2, p. 859.

⁷⁰ See, e.g., Collingwood, Trans. Cumb. and West. Ant. and Arch. Soc. n.s. XXXVIII, 1938, 296 f.; Stevens, , Eng. Hist. Rev. LII, 1937, 196CrossRefGoogle Scholar, followed by Wright, , Arch. Ael. 4th series XVII, 1940, 120Google Scholar, though he has to make an exception of the Military Way.

⁷¹ With some such phrase as ‘R. P. Silensium’, for instance: CIL VIII 10295.

⁷² RIB 2219–2314; but see n. 68a above.

⁷³ l.c. (n. 53 above), 57.

⁷⁴ The assumption that London was ‘by the Flavian period already the capital of Roman Britain’ (Birley, ibid.) is presumably based on the tomb stone of the procurator Julius Classicianus discovered in London; cf. Collingwood and Myres, Roman Britain and the English Settlements, 170 f. One cannot help wondering if this is sufficient evidence to justify our placing the seat of the governor of a major military province so far from his legions. The few tombstones and dedications of individual soldiers from London, which Ritterling related to a detachment of soldiers at the seat of the governor (RE, s.v. legio, 1461, 1774) do not seem to require such an explanation. Mommsen conjectured that York was the seat of the governor; see Provinces of the Roman Empire I, 193; and, even if the palace at York in Severus' time is only known from the SHA (Vita Severi XXII, 7), and may well have been specially built to accommodate the imperial family, it is at least worth noting that York was their seat while in Britain.

⁷⁵ RIB 2248, 2281. I gladly acknowledge here that this section on the milestones has benefited greatly from Mr. R. P. Wright's expert knowledge of Romano-British inscriptions, which he most generously put at my disposal, before the appearance of RIB, in a long letter on the stones in question.

⁷⁶ RIB 2244; see also Eng. Hist. Rev. LII, 1937, 196Google Scholar.

⁷⁷ RIB 2241. Haverfield conjectured that another milestone bore a distance from Lincoln; see EE IX, 1251, 6, of which he wrote ‘fuit fortasse [a L(indo)] R(atas) [MP LII]’. But Wright has recognized that the true reading of the line is R(es)p(ublica) L(indensis); see RIB 2240.

⁷⁸ RIB 2274.

⁷⁹ From Navio, RIB 2243; from the camp at Lancaster, if the marks on the stone in fact refer to its Roman name, RIB 2272; from Canovium, RIB 2265; (also from Canovium, RIB 2266, if we follow Haverfield's restoration, EE VII, 1100; Wright, RIB, does not); from Trimontium, RIB 2313.

⁸⁰ Three milestones which have been printed as giving measurements may now be excluded on the authority of the superior readings in RIB, viz.; CIL VII, 1173, line 4 of which was read by Watkin as ‘M.P. III’ (Roman Lancashire (Liverpool, 1883) 182Google Scholar); see now RIB 2270; CIL VII, 1180, lines 8 f. of which have been thought to contain a numeral, as EE IX, p. 638; see now RIB 2276; and EE VII, 1114, where the brief and enigmatic inscription does not seem likely to be a measurement; see RIB 2305.

⁸¹ RIB 2235; cf. Stevens, , Eng. Hist. Rev. LII, 1937, 199.CrossRefGoogle Scholar

⁸² RIB 2299. That the distance must be from Corbridge is shown by Wright, , Arch. Ael. 4th series XVII, 1940, 118 f.Google Scholar

⁸³ RIB 2306. For discussion of the reading and mileage see Wright, l.c. 118, 119 f., who notes that even the solution Portgate is unsatisfactory, since the distance is 20⅕ Roman miles in contrast to the stone's ‘XVIII’.

⁸⁴ Birley allows that subsidiary roads were measured from the point where they left the main road, but this exception does not save the rule, for it only explains one of our Romano-British examples (see previous note), and that not certainly.

⁸⁵ Cf. Antonine Itinerary, ‘Iter Brittaniarum I’, ‘a limite, id est a vallo’ (Cuntz, O., Itineraria Romana (Leipzig, 1929), p. 71Google Scholar). The general concept of measurement from capital to frontier hardly requires documentation; cf. CIL v, 8002, 8003, for a road from Italy to the Danube; and, within a province overseas, the African roads from the military capital to the frontier; e.g. CIL VIII, 10023. For measurement to, instead of the commoner measurement from, see, e.g., RIB 2241.

⁸⁶ See above n. 55.

⁸⁷ IGR III, 335 = Ditt., OGIS 538. On the subject as a whole see Hirschfeld, O., Die Kaiserlichen Verwaltungsbeamten, second edition (Berlin, 1905), 402.Google Scholar

⁸⁸ IGR III, 466.

⁸⁹ See Pflaum, Carr. I, 146, and Schlumberger, Syria XX, 1939, 52–61. Note that Pflaum prints the text more optimistically than Schlumberger, the original editor; but the important words for the present purpose, ‘p[ro]c. A[ug.]’, are well justified by Schlumberger (60 f.), partly by reference to other known appearances of P. Postumius Acilianus; cf. Pflaum, o.c. no. 62. Schlumberger also relies on the inscription cited in n. 87 above when he writes (61) ‘L'activité de Postumius Acilianus comme assistant du gouverneur dans une opération d'abornement appartient aux attributions normales d'un poste tel que celui qu'il occupait’.

⁹⁰ Unfortunately not, it seems, properly published, but see A. Stein, Die Reichsbeamten von Dazien, 63, from whom Pflaum quotes the inscription, Carr. II, 691. For a discussion of its date see Pflaum, 692 f.

⁹¹ After the rearrangement of Dacia in 167 the procurator of Porolissensis must have been financial, as the praetorian legate of Leg. V Macedonica became the leading man in that part of Dacia. For the evidence and discussion see Stein, o.c. 40 ff., whom Pflaum follows, Procc. 149.

⁹² Another example of an equestrian procurator and his senatorial superior collaborating and recording their collaboration from the same period (though not in a province) is to be seen in the lead pipe inscribed ‘succur. Thrasia Prisc, cos. c.v. et Vari Marcelli proc. Augg.’ (ILS 8687).

⁹³ o.c. (n. 53 above), 60.

⁹⁴ See Pflaum's discussion, Procc. 146 ff.

⁹⁵ The size of the garrison of Judaea while it was under a procurator is not known for certain. It consisted of at least 3000 men, but probably not many more. See Abel, F. M., Histoire de la Palestine (Paris, 1952) 1, 427.Google Scholar

⁹⁶ See Daicoviciu, C. and Protase, D., JRS LI, 1961, 63–70Google Scholar, especially 68 f. The new diploma there published, which proves that Dacia Porolissensis already existed in 133, has rendered obsolete the previous chronology of the divisions of Dacia, worked out, e.g., by Stein, o.c. 18, 33 f., 38 ff.

⁹⁷ CIL XVI, 75, 110; III, 12601a + 13793; ILS 8909, 9180; and Stein 18, 30 f.

⁹⁸ As, for instance, by von Premerstein, A., Wiener Eranos (1911), 256 f.Google Scholar; Stein, o.c. 20; Pflaum, Procc. 148. Brandis, however, RE, s.v. Dacia, 1971, and Stein earlier, RE, s.v. Flavius, no. 67, impressed by the command over troops, argued for the independence of the procurator.

⁹⁹ See n. 97 above.

¹⁰⁰ First in Le marbre de Thorigny, 23; repeated, Procc. 148.

¹⁰¹ See, for example, CIL XVI, 99, where vexillations from Pannonia Inferior, actually discharged in Africa by the procurator of Mauretania Caesariensis, are described as ‘sub’ the governor of Pannonia Inferior; or 37, where veterans from the classis Moesica are described as ‘sub’ the governor of Moesia Inferior, not the praefectus of the fleet; see Stein, , Die Legaten von Moesien (Budapest, 1940), 58Google Scholar, and cf. Starr, C. G., The Roman Imperial Navy (Cornell Studies in Classical Philology XXVI, Ithaca N.Y., 1941), 107.Google Scholar The legal position in these cases is perhaps more fully expressed by 32, where the Praefectus Aegypti and the Praefectus classis Alexandrinae are both named. That it was thought important to name the senior officer is shown by 108, where a vexillation is described as under the temporary commander, without mentioning their permanent commander, no doubt because he was of inferior rank.

¹⁰² See n. 96 above.

¹⁰³ See Pflaum, Carr. I, no. 157 bis.

¹⁰⁴ ibid., and Procc. 129 f.

¹⁰⁵ This is demonstrated by Stein, Reichsbeamten von Dazien, 41 f., using the careers of the legionary legates in Dacia.

¹⁰⁶ For references see Stein 42, nn. 5 and 6.

¹⁰⁷ The largest otherwise would seem to be Raetia; Cheesman, , The Auxilia of the Roman Imperial Army (Oxford, 1914), 151 f.Google Scholar, arrives at a total of 12,500. But Schleiermacher's list of the garrison in the mid-second century gives a total of 10,000; see Schleiermacher, W., Der Limesführer, second edition (Berlin, 1964), 231 f.Google Scholar Cheesman gives Mauretania Caesariensis 10,200 (165 f.). The northern frontier of Britain, however, appears to have had a garrison of more than 20,000; see Cheesman, 109 f., and cf. Birley, o.c. (n. 53 above) 60, who notes that these troops will have been at least equal to the three legions in numbers.

¹⁰⁸ The argument (Procc. 125 ff.) rested too heavily on the over-formal argument that only senatorial governors enjoyed the imperium. As Sherwin-White rightly objected (JRS XLIII, 1953, 171Google Scholar) the Praefectus Aegypti makes nonsense of this argument. But Pflaum's numerous instances of the special appointments of duces or praefecti for important campaigns are sufficient to show that the procurator as such was not normally employed in this capacity.

¹⁰⁹ One notes, for instance, that Raetia ceases to be procuratorial and is given a legion under Marcus Aurelius.

¹¹⁰ SEG XVII, 505; Ephesus, Inventory 1273. First published by J. Keil, Sitz. Bayr. Akad. 1953, Heft 3, ‘Ein ephesischer Anwalt des 3. Jahrhunderts durchreist das Imperium Romanum’. The absence of a photograph in this editio princeps is most regrettable.

¹¹¹ The information in 11. 1–6, especially the names partly erased, places the inscription beyond any doubt in the reign of Macrinus, and Diadumenianus' titles imply a date between the end of May 217 and the end of May 218, as the editor shows, p. 5. The slight inconsistency between this view and his earlier dating indication (p. 3), where April 217, the beginning of Macrinus' reign, is given as the terminus post quem, may in fact be a justifiable recognition that Diadumenianus could have been called Caesar before he officially received the title. In any case the date 217–8 ensures that the journeys took place in the reigns of Severus and Caracalla, with whose movements they are perfectly consistent.

¹¹² Atkinson's list of the legates of Britain, JRS XII, 1922, 60–73, is the last to be published with full documentation, and is naturally long out of date. Birley's list (see above, n. 6), though far more satisfactory, has, as we have seen, no references. In the discussion that follows I refer to these two lists by the author's name alone.

¹¹³ Atkinson, no. 32, Birley, no. 35 (197–202 ?), Barbieri, Albo Senatorio, 528 and additions (p. 604).

¹¹⁴ See nn. 24 and 25 above.

¹¹⁵ Digest XXVIII, 6, 2, 4. The words ‘imperatoris nostri’ were taken by Atkinson (ibid.) to mean that Caracalla was Augustus when the rescript was sent. This does not seem certain but, in any case, as Atkinson saw, no precise conclusion as to chronology can be drawn from this reference.

¹¹⁶ See p. 93.

¹¹⁷ Barbieri, Albo Senatorio, 514.

¹¹⁸ JRS LI, 1961, 192, no. 4, 11. 6–8: ‘[dd.] nn. imp. Antonino II et Geta Caesare cos.’

¹¹⁹ See E. Ritterling, Fasti des römischen Deutschland, no. 36.

¹²⁰ Not after, as in the ill-informed article RE, s.v. Valerius no. 322, 214; cf. Thomasson, B. E., Die Statthalter der römischen Provinzen Nordafrikas von Augustus bis Diocletianus (Lund, 1960) 11, 110, n. 443.Google Scholar

¹²¹ This was the conclusion of Ritterling, , Arch. Ertesitö XLI, 1927, 293.Google Scholar The inscriptions from Pannonia are undated, unless we press conclusions from the singular Aug., as Thomasson, ibid.

¹²² See Thomasson, o.c. I, 55.

¹²³ Thomasson II, 110 f.

¹²⁴ Cf. Thomasson I, 31.

¹²⁵ Atkinson, no. 35; Birley, no. 36 (202 ?–208 ?); Barbieri, Albo Senatorio, 25.

¹²⁶ ILS 2618 = RIB 1234.

¹²⁷ But it is probably correct, since on all seven of the inscriptions, which certainly or probably record his activity in Britain, Geta is still Caesar. For the inscriptions see above, nn. 27, 28.

¹²⁸ Evidence: Jalabert, L. and Mouterde, R., Inscriptions grecques et latines de la Syrie I (Paris, 1929) 39Google Scholar (= ILS 5899), 40 and, probably, 41; discussion: Gilliam, , AJP LXXIX, 1958, 228 f.Google Scholar, with earlier bibliography.

¹²⁹ Barbieri, ibid.; while Atkinson (p. 70) puts his governorship of Syria Coele after Britain, without a single argument.

¹³⁰ E.g., Stout, S. E., The Governors of Moesia (Princeton, 1911), 37Google Scholar, n. 74(b); Harrer, G. A., Studies in the history of the Roman province of Syria (Princeton, 1915), 43 f.Google Scholar; A. Stein, Die Legaten von Moesien, 54 f.; Jalabert and Mouterde, o.c., notes to no. 39; Pflaum, Carr. pp. 441–3.

¹³¹ See Pflaum, ibid.

¹³² Inscr.…de la Syrie I, 42–4. These bear dedications to Severus, Caracalla and Julia Domna; the assumption that there was originally a fourth column, bearing an inscription to Geta, seems almost certain to be correct; see Jalabert and Mouterde, o.c. 42.

¹³³ It is, no doubt, unjustifiable to plot imperial journeys solely on the basis of dedications and municipal privileges. But the example of Lambaesis, where the numerous dedications of the year 203 support the definite evidence of an imperial visit in that year (see Hasebroek, o.c. 133 f.), shows that it is legitimate to entertain the possibility of an imperial visit whenever a number of dedications, all of the same date, are found in one place.

¹³⁴ On the chronology see Hasebroek, o.c. 117–20, whose conclusions have only been strengthened by later discoveries; see Milne, , Journ. Egypt. Arch. IV, 1917, 180 f.Google Scholar, Westermann and Schiller, o.c. (n. 33 above), 30. Hannestad, , Classica et Mediaevalia VI, 1944, 194 ff.Google Scholar, is unreliable. The knowledge that Severus was probably in Egypt from August 199 to after August 200 makes it improbable that the inscriptions at the bridge in Commagene give the exact date for his presence there. But obviously the suggested visit which led to the decision to repair the bridge could have been a little earlier or a little later than the inscriptions which celebrated the task's completion.

¹³⁵ See Gilliam, l.c. 229 f., n. 15.

¹³⁶ Though it is not certain that Pudens passed directly from Germania Inferior to Britain. However, the known dates of his governorships (between 195 and 198 Germany, Britain 205) make it possible, if we assume fairly long occupations of both commands.

¹³⁷ Groag in Ritterling, E., Fasti des römischen Deutschland unter dem Prinzipat (Vienna, 1932), 77.Google Scholar

¹³⁸ Cf. Domaszewski's table, o.c. 180–2.

¹³⁹ Dio LXXIV, 4, I; SHA, Vita Pert. II, II and III, 5; cf. Vita Commodi VI, 2.

¹⁴⁰ Atkinson, no. 33; Birley puts him between nos. 35 and 36 (202 ?); Barbieri, Albo Senatorio, 34.

¹⁴¹ Atkinson, no. 53a; Birley, no. 38 (217 ?–220 ?, governing Britannia Inferior); Barbieri, o.c. 312.

¹⁴² Atkinson, no. 34; Birley, no. 39 (220 ?–223 ?, governing Britannia Superior); Barbieri, o.c. 413.

¹⁴³ ILS 1151.

¹⁴⁴ CIL VI, 32326.50; 32327.10; 32332.3.

¹⁴⁵ Meiggs, R., Roman Ostia (Oxford, 1960), 177.Google Scholar

¹⁴⁶ CIL XIV, 325.

¹⁴⁷ CIL VIII, 597 and 11763.

¹⁴⁸ Cf. Stein's discussion, Die Legaten von Moesien, 87 f.

¹⁴⁹ o.c. 82–4.

¹⁵⁰ o.c. (n. 120 above), II, 96 ff.

¹⁵¹ See Pick, B., Die antiken Münzen Nord-griechenlands (Berlin, 1898) 1, 355 f.Google Scholar, nos. 1252–1263 and, on the date, Stein, ibid. The objections to this identification are set out by Stein. But, as Thomasson observes, the major difficulty, the 46 or more years separating Pollienus Auspex the Younger's consulate and that of his son, Tib. Pollienus Armenius Peregrinus, is removed by the old hypothesis that the son was adopted, which the names can be held to suggest.

¹⁵² Moesia, Dacia and Spain; see ILS 8841: ὑπατικοῦΒριταννίας, Μυσκίας, Δακίας, Σπανίας.

¹⁵³ This statement needs qualification if the restoration ‘su[b A. T]r[iar]io Rufino’ on the newly discovered inscription from Reculver is correct (JRS LI, 1961, 191, no. 1 = AÉ, 1962, 258); for we know that A. Triarius Rufinus was cos. ord. in 210, though nothing else of his career; see Barbieri, Albo Senatorio, 504. But there are good reasons for choosing another restoration, which would place the inscription in the reign of Severus Alexander; see above, n. 61. The statement in the text would also require modification if Gordian was a governor in Britain under Caracalla; cf. RIB, nn. to 1049, 1279.

¹⁵⁴ The analogy is regularly advanced: cf., e.g., Miller, CAH, XII, 15; Ritterling, RE, s.v. legio, 1609.

¹⁵⁵ On the date see Harrer, , AJA XXXVI, 1932, 287–9.CrossRefGoogle Scholar

¹⁵⁶ For a fully-documented account of Severus' treatment of the Eastern frontier see Magie, Roman Rule in Asia Minor I, 672–5; II, 1540–5.

¹⁵⁷ ILS 1140: ‘Tib. Cl. Candido … duci exercitus Illyrici expeditione Asiana item Parthica item Gallica’; this must refer to the First Parthian War, as it comes before ‘Gallica’.

¹⁵⁸ The aims of the war are not clearly stated in the literary sources. Dio says that it was waged for glory (LXXV, I, I), though he reports that Severus himself claimed that he had created a bulwark for Syria (LXXV, 3, 2). A punitive motive can be assumed also from the implication of the passage LXXV, I, 2–3, where the sequence of events is compressed to the point of obscurity, but the words ἡττηθέντες ὑπὸ Σεουήρου suggest that the Osroenians and Adiabenians had fought for Niger at first. This would agree with Herodian III, I, 2–3 (cf. II, 8, 8).

¹⁵⁹ We have to assume either that he annexed the whole kingdom in 195 and made it a province, but soon after reinstated Abgar VIII/IX on the throne, or that he made part of the kingdom a province in 195, but confirmed Abgar in his rule of the rest. See Magie's discussion, o.c. II, 1543 f. The second assumption avoids the necessity of postulating a major change of mind by Severus.

¹⁶⁰ For the chronology see Hasebroek, o.c. 110–3. Ctesiphon must have been captured ca. December-January 197–8; see Murphy, G. J., The Reign of the Emperor L. Septimius Severus from the evidence of the inscriptions (Philadelphia, 1945), 24–6Google Scholar. This is a mere ten or eleven months after the battle of Lugdunum.

¹⁶¹ Cf. Dio LXXV, 8, 3–5; Herodian III, 5, 2; 8, 6; SHA, Vita Severi X, 2.

¹⁶² The rearrangement of the legions in Pannonia so that Superior and Inferior each had two legions is securely dated between 212 and 217 (at the latest); ILS 2382 proves that Pannonia Superior still had three legions in 212, and Suetrius Sabinus, consular governor of Pannonia Inferior, was recalled from there in 217; see Dio LXXVIII, 13, 2 and ILS 1159. Attempts to fix the date more closely have naturally been made; for a recent one see Fitz, Accad. d'Ungheria in Roma, Quaderni di Documentazione II, no. 2, 1961, ‘Il soggiorno di Caracalla in Pannonia nel 214’, especially 12 ff., and Acta Antiqua Acad. Scient. Hung. XI, 1963, 287–9.

¹⁶³ Cf. Miller, CAH XII, 27; Pflaum, Libyca V, 1957, 75.

¹⁶⁴ On the Tripolitanian frontier see Goodchild, and Ward-Perkins, , JRS XXXIX, 1949, 81 ff.Google Scholar; on the whole African frontier, Romanelli, P., Storia delle province romane dell' Africa (Rome, 1959), 398–408.Google Scholar Cf. also Picard, G. C., Castellum Dimmidi (Paris, 1944), 49 ff.Google Scholar

¹⁶⁵ On the recognition of Numidia as a separate province by Severus see Pflaum, o.c. 61–75.

¹⁶⁶ Cf. Domaszewski's account, o.c. 175–9.

¹⁶⁷ Thus earlier divisions usually followed reinforcement of a threatened frontier, which, by stationing more than four legions in a province, made division necessary. E.g., the divisions of Moesia and Pannonia: cf. Domaszewski, 178.

¹⁶⁸ Cf. Frere's remarks in Arch. Journ. CXVIII, 1961, 256Google Scholar f.

¹⁶⁹ Dio LXXV, 5, 4 (Exc. U^R 18, (p. 414), Boissevain III, 346).

¹⁷⁰ It has been suggested that similar threats had to be met in Wales; see Simpson, G., Archaeologia Cambrensis CXI, 1962, 108, 110 f.Google Scholar; CXII, 1963, 72–5; Britons and the Roman Army, 48–52, 177 f. But the picture there drawn of widespread revolt followed by equivalent military reconstruction seems to rest on evidence too slight and uncertain as yet to bear its weight. The only explicit testimony comes from Severan building-inscriptions at Caerleon and Caernarvon. Both blame peaceful decay for the need for repairs. This may represent, as Dr. Simpson maintains, Roman unwillingness to admit enemy destruction, but we need clearer and more substantial evidence than two inscriptions hypothetically interpreted. Nor do the destructions which she finds in the archaeological record at several sites require the assumption of a general, large-scale revolt. More evidence seems to be needed to test this ambitious hypothesis. For a very different view of the military history of Wales in Severan times see Jarrett, M. G., Bull. of the Board of Celtic Studies XX, 1962–4, 215–7.Google Scholar

¹⁷¹ See Strabo II, 115 f.; IV, 200 f.; described with reason by Mommsen as ‘the governmental version’, Provinces of the Roman Empire I, 172, n. 2. For a stimulating hypothetical reconstruction of the Roman policy regarding Britain under Augustus, see Stevens, , ‘Britain between the invasions’, in Aspects of Archaeology, edited by Grimes, W. F. (London, 1957), 332–44.Google Scholar

¹⁷² The recall of Agricola by Domitian can be variously interpreted, but there is no reason to think that it was prompted by a decision to abandon the policy of conquering the whole island; cf. Tac., Hist. I, 2, Agric. 40, and Syme, CAH XI, 156. As for the withdrawal of Leg. II Adiutrix in ca. 86, this reflects the fears about the Danube frontier rather than general decisions about the long-term future of Britain; cf. Collingwood, Roman Britain and the English Settlements, second edition, 116–9, and Richmond, , Roman Britain (London, 1955), 45.Google Scholar

¹⁷³ The advantages are described by Richmond, o.c. 50.

¹⁷⁴ The explicit evidence of the inscriptions (cited above, nn. 25–8) is the basis for this reconstruction, to which the evidence from the buildings can be added; see Steer, in Roman and Native in North Britain, edited by Richmond, I. A. (London, 1958), 94 f.Google Scholar

¹⁷⁵ SHA, Vita Severi XVIII, 2; Aur. Victor, Caesares XX, 18, not to mention later writers.

¹⁷⁶ LXXVI, 13, 1.

¹⁷⁷ E.g., Steer, ibid., who is followed by F. Millar, A Study of Cassius Dio, 148 f.; Richmond, Roman Britain, 57–9; Collingwood, o.c. 158 f.

¹⁷⁸ LXXVI, 12.

¹⁷⁹ III, 14, I.

¹⁸⁰ LXXVI, 10, 6.

¹⁸¹ See Dio LXXVI, II, I; Herodian III, 14, 2.

¹⁸² See Birley, R. E., Scot. Hist. Rev. XLII, 1963, 126–34Google Scholar, and Proc. Soc. Ant. Scot. XCVI, 1962–3 (appeared 1965), 184–207; cf. also JRS LIII, 1963 127 and Plate X.

¹⁸³ Leg. II Aug. is now attested at Carpow in addition to VI Victrix; see JRS LV, 1965, 223 f.

¹⁸⁴ Its discovery also seems to strengthen the case made by E. Birley for assuming that there was a Severan plan to make a legionary headquarters at Corbridge; see Arch. Ael. 4th series XXXVII, 1959, 12–20. For that too could only be explained by a forward policy in Scotland.

¹⁸⁵ See III, 5, I.

¹⁸⁶ III, 10, 2.

¹⁸⁷ III, 10, I.

¹⁸⁸ See, e.g., E. Hohl, Kaiser Pertinax und die Thronbesteigung seines Nachfolgers im Lichte der Herodiankritik (Sitz. Berl. Akad., Phil.-Hist. Klasse, 1956, no. 2), passim and Anhang I on the fall of Plautianus.

¹⁸⁹ o.c. (n. 53 above), 61.

¹⁹⁰ ibid., 61, with Arch. Ael. 4th series XI, 1934, 127–131.

¹⁹¹ See, e.g., ILS 2382 (Pannonia), presumably after Dec. 10th 212, and CIL XIII, 7465 and 7616 (Germany) of 212 and 213.