² Speidel, M., ‘The Captor of Decebalus, a New Inscription from Philippi’, JRS LX (1970), 142–153Google Scholar.

³ Speidel considered that, as the career is so detailed, earlier positions as miles and discens equitum had not been omitted from the stone, and Maximus accordingly had enlisted as an eques (l.c., 143). The lowest recorded stipendia of a legionary eques is four years (AE 1902, 41 = ILS 9090) and seven is the next lowest (V, 896 = ILS 2332), while only one auxiliary eques is certainly known to have enlisted as such (BGU 696, 28–32); cf. Gilliam, J. F., ‘Dura Rosters and the Constitutio Antoniniana’, Historia 14 (1965), 74–81Google Scholar, where the suggestion is made that equites had normally served as pedites for about ten years before promotion, though recruits when of sufficient merit and influence could enlist as equites. Many quite detailed careers, on the other hand, omit the lower posts, e.g. Florus must have held one or more posts between being made a principalis in A.D. 209 and his promotion, probably to tesserarius, in 213 (IX, 1609). It is therefore not certain that there are no omissions from the career of Maximus and that he had enlisted as an eques.

⁴ Speidel considered that these four posts could hardly have been held in less than five years (l.c., 143). This indeed would have been a minimum. Petronius Fortunatus at present stands alone in being promoted centurion after only four years service (VIII, 217 with P. 2353 = ILS 2658); the next lowest is thirteen years (XII, 2234 = ILS 2342), while the other two lowest stipendia known for optiones and signiferi are 6 and 8 (III, 2716; 4375; 5976 and 10525). It is therefore likely that Maximus had rather more than five years' service before being decorated as a vexillarius equitum (probably in 89), especially if he had enlisted as a pedes. If the dona awarded by Domitian are not placed in their correct position in the career but at the end of his service in the legion, which is not impossible, Maximus would have had ample time from before 89 to after 101, when he was appointed duplicarius, to hold the four posts.

⁵ When Maximus reached the rank of decurion, the rule of automatic retirement for milites after 25 (or 26) years service would cease to apply to him; centurions and decurions served for far longer than this term. For long serving centurions see Birley, E. ‘Promotions and Transfers in the Roman Army, II: the Centurionate’, Carnuntum Jahrbuch 8 (1963/1964), 33Google Scholar. Presumably they could apply for discharge at any time after the completion of 25 years service. There are centurions who receive honesta missio like Maximus. It is unlikely that missus voluntarius honesta missione means more than this. For a discussion of the problem in relation to legionary centurions see Dobson, B., ‘The Centurionate and Social Mobility during the Principate,’ Recherches sur les Structures Sociales dans l'Antiquité Classique ed. Nicolet, C. (Paris 1971), 101 fGoogle Scholar.

⁶ Although the appearance of this post seems to suggest that the equites legionis had funds of their own, the existence of these funds is not necessarily evidence that the equites had a permanent commander (Speidel l.c., 144). Since the legionary cohort, a tactical unit in battle, did not have a permanent commander, it might have been considered that the equites did not need one either.

⁷ VI, 3339 and VI, 3614 each mention a singularis legionis, and Passerini, A. (Dizionario Epigrafico 4, 605Google Scholar) rightly considered that they could equally well have been on the staff of the legionary legate or of the provincial governor. Since it is well-attested that the governors drew their singulares from the auxiliary units, it is perhaps probable that the above two soldiers, the only singulares legionis recorded apart from Maximus, were on the staff of the legate. C. Iulius Nepos, a pedes singularis buried at Lambaesis, may have been a legionary on the staff of the legate of III Augusta, as his brother, a beneficiarius consularis, undoubtedly was (VIII, 2911). AE 1957, 122, found in the legionary headquarters at Lambaesis, is a dedication by equites singulares who are possibly but not certainly legionaries.

⁸ There is only one other recorded promotion from a legion, in this case a miles, to duplicarius in an auxiliary unit (VIII, 2354 = ILS 305), though there are attested a number of promotions of legionaries to centurion (III, 8438 = ILS 2597; V, 8185 = ILS 9172; probably P. Mich. III, 164 and just possibly XII, 3177) and decurion (III, 647 = ILS 2538; VIII, 9370 = ILS 1357a(?); AE 1917/8, 74; P. Mich. III, 164 (2)). II, 2554 with AE 1910, 1; V, 522; III, 7449; VIII, 9002; 17619; X, 7580 = ILS 1358; and AE 1958, 156 (cited by Speidel, l.c., 146, n. 43 and 44) include no certain cases of direct promotion from a legion to centurion or decurion in an auxiliary unit.

⁹ VIII, 2562 with Neue Heidelberger Jahrbücher 9, 150 is possibly but not certainly in hierarchical order.

¹⁰ o.c. 145.

¹¹ VI, 37191 = ILS 9190.

¹² o.c. 147.

¹³ Domaszewski-Dobson, 47 and 71.

¹⁴ ibid., 71.

¹⁵ o.c. 1–6. This division only holds good for the first and second centuries; in the third century a different situation arises—as Domaszewski suggested, though on different grounds.

¹⁶ II, 2610 = ILS 2079; III, 2887 = ILS 9067; IX, 5809 = ILS 2078; XI, 710; AE 1933, 87. For a full discussion of the promotion patterns in the guard and other units see my forthcoming paper in Epigraphische Studien, ‘The Career Structure below the Centurionate’.

¹⁷ III, 7334 = ILS 2080; VI, 2794; XI, 5646 = ILS 2081.

¹⁸ XI, 20 = ILS 2082.

¹⁹ VI, 2454 = ILS 2060.

²⁰ II, 2610 = ILS 2079; III, 7334 = ILS 2080; VI, 3661; 37191 = ILS 9190; X, 1763; XI, 5646 = ILS 2081.

²¹ VI, 2440 = ILS 2077.

²² VI, 2794; IX, 5839 = ILS 2085; XI, 20 = ILS 2082; XI, 710; XIII, 6728; AE 1933, 87.

²³ III, 7334 = ILS 2080; VI, 2755 = ILS 2145; VI, 2794; VI, 37191 = ILS 9190; VI, 32887; IX, 5839 = ILS 2085; XI, 20 = ILS 2082; XI, 395 = ILS 2648; XI, 710; XI, 5960; XIII, 6728; XIII, 6823; XIV, 4626 = ILS 2742; AE 1933, 87. For direct promotion from cornicularius praefecti praetorio to centurion, see II, 2664; III, 3846 = ILS 2652; VI, 1645 = ILS 2773; X, 1763; XI, 3108; XI, 6055 = ILS 2743; Pais 1253. Cf. Dobson, B. and Breeze, D. J., ‘The Rome Cohorts and the Legionary Centurionate’, Epigraphische Studien 8 (1969), 100–117Google Scholar.

²⁴ V, 7160 = ILS 2086, from optio to evocatus and then centurion (some posts may have been omitted from this career); VI, 2454 = ILS 2060, from tesserarius to evocatus; V, 3371 and VI, 32520, a, 2, 21, from signifer to evocatus.

²⁵ VIII, 4874 = ILS 2116; IX, 1617 = ILS 2117.

²⁶ VI, 414b; VI, 2987 = ILS 2169; VI, 37295; X, 3880 = ILS 2171; XI, 1438; XI, 5693 = ILS 2666; VI, 1057 (1) 3 and VI, 1058 (4) 7; VI, 1057 (6) 13 and VI, 1058 (3) 6; VI, 1057 (2) 10 and VI, 1058 (3) 5; VI, 1057 (3) 3 and VI, 1058 (3) 3; VI, 1057 (6) 4 and VI, 1058 (4) 4.

²⁷ Domaszewski-Dobson, 43.

²⁸ III, 12411 = ILS 2666b; AE 1937, 101; cf. III, 11135 = ILS 4311.

²⁹ III, 12411 = ILS 2666b; V, 7872; AE 1937, 101.

³⁰ III, 1124; III, 1202; VIII, 217 = ILS 2658; XII, 2929, demonstrate that optio was held before signifer; AE 1937, 101 shows the sequence signifer, cornicularius, optio ad spem ordinis.

³¹ Domaszewski-Dobson, 42; H. M. D. Parker, The Roman Legions (1958), 207.

³² Passerini, o.c. 595, on the basis of VIII, 2554.

³³ V, 7004; AE 1951, 194. The soldier in the latter inscription, after being benefic(iarius) leg(ati) and cornicularius leg(ati), became optio coh(ortis) I, which may have been equivalent to optio spei.

³⁴ The career recorded in III, 12411, dating to the late first century, is the earliest corresponding to this pattern; that of III, 11135 = ILS 4311, probably dated to 235–238, the latest.

³⁵ III, 11135 = ILS 4311: librari[u]s numeris, cus(tos) arm(orum), signif(er), optio o[cta]v[a] (?) pr(incipis) pr(ioris), candidatus; III, 12411: b(ene)f(iciarius) lega(ti), opt(io) ad spe(m) ordin(is), (centurio) leg(ionis); V, 7004: [b(ene)f(iciarius)] legat(i), a comment(ariis) [.….]f., optio, centurio [legi]onis; VIII, 217 = ILS 2658: librar(ius), tesser(arius), optio, signif(er), (centurio) factus ex suffragio leg(ionis); AE 1937, 101: signif(ero), corni[cular(io)], optioni ad spe[m ordi]nis, (centurioni) leg(ionis); AE 1951, 194: benefic(iarius) leg(ati), cornicularius leg(ati), optio coh(ortis) I. ILS 8880—β(ενε)φ(ικιάριος) καὶ ἀκομενταρήσιος καὶ κορνικουλάριος καὶ (ἑκατοντάρχος) γενόμενος τῆς ἡγεμ(ονίας)–is in a sense the exception which proves the rule, for this soldier held three posts on the staff of the governor of Arabia and none in the century, and went on to be a centurion on the same staff. (The soldiers recorded in III, 14178; V, 742 = ILS 2670 and VIII, 12128, whose careers do not follow the above pattern, probably date to the third century.) The only group of careers which do not fit into the system in this period are those of the aquiliferi. Aquiliferi are known to have been promoted from the ranks of the signiferi (V, 3375 = ILS 2339; V, 5832 = ILS 2338; and possibly VIII, 2568, 81 with VIII, 2796) and even from miles (XII, 2234 = ILS 2342), and advanced directly to the legionary centurionate (XII, 2234 = ILS 2342; XIII, 6646 and 6952), but no aquilifer is known to have served as an optio.

³⁶ The letter of Iulius Apollinaris, P. Mich. VIII, 466, 18 ffGoogle Scholar., demonstrates the subtle differences between apparently similar posts.

³⁷ Promotion from optio to legionary centurion: III, 14178; V, 942 = ILS 2670; V, 7004; VIII, 2554 and 18048, b, 22 = ILS 2445 with VIII, 2848; optio spei: III, 12411 = ILS 2666 b; V, 7872; AE 1937, 101; signifer: VIII, 217 = ILS 2658 and possibly XII, 3177 with XII, 3178; aquilifer: XII, 2234 = ILS 2342; XIII 6646; 6952; beneficiarius consularis: VIII 17626; cornicularius legati: VIII, 12128; cornicularius consularis: III, 6542 with 6543 with 6598; XIII 6803; ILS 8880; cornicularius: XIII, 1832.

³⁸ The only known promotion to a legionary centurionate from a post below optio was from eques (AE 1902, 41 = ILS 9090), but this took place in the first half of the first century before the promotion system became more formalized.

³⁹ Note, e.g., the caveat of G. R. Watson, The Roman Soldier (1969), 92. Domaszewski had supposed that the amount of money given to soldiers on their retirement, as recorded in the records of the scholae at Lambaesis, all dated to the period 198 to 211, was equal to a year's pay. The cornicines were thus immunes (VIII, 2557 = ILS 2354); the librarii and the exacti, sesquiplicarii (AE 1898, 108–9 = ILS 9100); the armorum custodes, the cornicularii praefecti and the cornicularii and actarii legati legionis, duplicarii (AE 1902, 147a = ILS 9097; AE 1899, 60 = ILS 9099; AE 1898, 108–9 = ILS 9100). G. R. Watson has suggested to me that the large sum given to the optio spei, not on his retirement but on his promotion to the centurionate, was equal to four times the basic pay, or twice his own, while the unsuccessful optio received upon his retirement three times the basic, or one-and-a-half times his own, by way of compensation for not getting promoted (VIII, 2554 = ILS 2445). If this is the case, Watson suggests, on the basis of his calculations of pay in the period Severus to Caracalla, that the amounts will have been rounded up slightly (cf. o.c, 91).

⁴⁰ Gilliam, J. F., ‘The Moesian Pridianum’, Hommages à A. Grenier (Collection Latomus LVIII, 1962), 755–6Google Scholar. Cf. R. O. Fink in The Excavations at Dura-Europos, ed. A. Perkins, Final Report V, 1: The Parchments and Papyri (1959), 32–33.

⁴¹ P. Dura 100, xxxi, 24; 26; xxxii, 16; 21; 29; xxxiii, 20; cf. R. O. Fink, l.c.

⁴² P. Dura 100, xvii, 3 and 101, xxii, 13.

⁴³ ILS 9107, apparently recording a list of legionaries in hierarchical order, places the actarius above the optio but below the signifer. The presence of an actarius among the equites of VII Gemina can be explained if the suggestion of Speidel, that this soldier served in the tabularium equitum and not the tabularium legionis, is accepted (l.c. 144, n. 20).

⁴⁴ Cf. XIV, 2255; AE 1895, 204 and AE 1898, 108, for the seniority of the cornicularius to the actarius in the legions.

⁴⁵ XII, 2602 = ILS 2118.

⁴⁶ VI, 2776; Domaszewski-Dobson, 21.

⁴⁷ It is, of course, uncertain what happened in a cohors peditata and an ala.

⁴⁸ This period is chosen because under Hadrian the distinction between immunis and principalis was formalized, while by the time of Severus changes were taking place in the status of certain posts. Cf. Watson, G. R., ‘Immunis librarius,’ Britain and Rome (eds. Jarrett, M. G. and Dobson, B. (1966)), 51, and The Roman Soldier (1969), 75–79Google Scholar.

⁴⁹ Cf. Tarruntenus Paternus' list of immunes, dating to this period: Digest 50, 6, 7.

⁵⁰ The difference between the immunis and the principalis after the time of Hadrian was presumably based upon pay.

⁵¹ In the guard a beneficiarius tribuni is described as a principalis, though the inscription may date to before the time of Hadrian (IX, 5809 = ILS 2078). The singularis praefecti praetorio had a higher status than an eques and therefore was presumably a sesquiplicarius (XI, 5646 = ILS 2081), as was probably the beneficiarius tribuni, who ranked just below the singularis praefecti praetorio (III, 7334 = ILS 2080). VI, 221 listing junior staff officers in the vigiles as principales dates to 113, that is before the distinction between the immunes and the principales was formalized. However, in the vigiles also the optio ca(rceris) ranked a little below the optio centuriae and was probably a sesquiplicarius (VI, 1057 (2) 10 with VI, 1058 (3) 5). The beneficiarius tribuni and the optio co(nvalescentium), ranking just below tesserarius, were probably also sesquiplicarii (VI, 1057 (1) 3 with VI, 1058 (4) 7; VI, 1057 (6) 13 with VI, 1058 (3) 6). On the basis of analogy with the guard and the vigiles, the posts of optio valetudinarii, optio carceris, singularis, beneficiarius tribuni and a quaestionibus in the urban cohorts may have fallen into this category (IX, 1617 = ILS 2117). The haruspex and the beneficiarius acil. in the legions, according to ILS 9107, seem to have ranked above tesserarius but below optio and may have been sesquiplicarii. Probably the beneficiarii of all junior officers belonged to this group.

⁵² The tesserarius was clearly a sesquiplicarius in the auxilia, and a principalis in the guard (IX, 1609 and probably XIII, 6728) and in the vigiles (VI, 220). In the legions he was apparently below duplicarius, for three tesserarii appear on a list of soldiers of III Augusta made duplicarii by Elagabalus (VIII, 2564 with 18052 = ILS 470).

⁵³ Tarruntenus Paternus includes the armorum custos in his list of immunes (Digest 50, 6, 7), but the regulations of the schola at Lambaesis would suggest that there at least the armorum custodes were duplicarii about 200 (AE 1902, 147a = ILS 9097). The list of soldiers made duplicarii by Elagabalus includes one armorum custos, suggesting that he was either an immunis or a sesquiplicarius (VIII, 2564 with 18052 = ILS 470). Finally one of a series of dedications erected by members of a vexillation of v Macedonica and XIII Gemina in the reign of Gallienus couples together the tesserarii and the armorum custodes as if they were similar in rank (AE 1936, 55).

⁵⁴ The librarii were included in Paternus' list of immunes and also Vegetius' list of principales (II, 7). It is possible that their status had risen in the intervening years, and in support of this it may be noted that librarii and exacti receive 800 denarii on retirement according to the regulations of their schola at Lambaesis, suggesting they were sesquiplicarii (AE 1898, 108–9 = ILS 9100). Earlier the librarii had certainly been immunes, cf. G. R. Watson, ‘Immunis librarius’, l.c. 45–55.

⁵⁵ The aquilifer clearly belongs to this category, and to it may be added the imaginifer, who is placed on a list of legionaries apparently in hierarchical order between the optio and the signifer (ILS 9107). Vegetius (II, 7) included the post in his list of principales. If imaginifer has been restored correctly on the roster of 219 he was probably a duplicarius in XX Palmyrenorum (P. Dura 100, xxviii, 21; xxxi 19; cf. R. O. Fink, l.c. 32).

⁵⁶ In addition to the cornicularii cited above note also the career of M. Aurelius Augustianus (VI, 2977 = ILS 2173): 7 coh. v vig., vix. an. XXXIIII, provitus ann. XVII, exceptor presidi provincies M.S. ann. IIII, lectus in praetoriae eques sive tabularius ann. V, factus 7 in Syria, vixit ann. VIII. Augustianus apparently was an eques and a tabularius simultaneously. He was advanced from the latter post directly to the centurionate; hence tabularius was probably a senior staff post and might therefore merit the extra pay of an eques. Tabularius was a senior staff post in the vigiles (VI, 37295).

⁵⁷ See n. 41.

⁵⁸ VI, 37191 = ILS 9190.

⁵⁹ The career of Maximus itself offers no help. The nearly contemporary singularis praefecti praetorio was probably a sesquiplicarius (see n. 51), and the singularis legati legionis may also have been; but this in itself would not prevent the post next held, vexillarius equitum, being of the same pay grade.

⁶⁰ Cf. Polybius VI, 24, 6; Caesar, , BG II, 25Google Scholar; BG V, 37; V, 5832; XII, 2234 = ILS 2342; V, 3375 = ILS 2339; V, 8185 = ILS 9172.

⁶¹ o.c. 147.

⁶² Dr. Brian Dobson and Mr. G. R. Watson read this paper in typescript and made several useful suggestions, which I have gratefully incorporated.