¹ I shall refer to this study and to the edition of the Feriale Duranum in Yale Classical Studies, VII (1940) by author and page. The three editors of the Feriale are R. O. Fink, A. S. Hoey, and W. S. Snyder. The Feriale Duranum is a papyrus which was discovered during the excavations of Dura-Europos. It was written under Severus Alexander between ca. 224 and 235.

I wish to thank Professor Fink, who has read my manuscript and who examined with me some parts of the papyrus itself in New Haven during the summer of 1953.

² Nock, pp. 195–197. Cf. Hoey, pp. 31, 172–174 for comments on the development of the military calendar, though he does not directly consider its origin.

³ It seems improbable that the feriae publicae would have been prescribed for the auxilia under Augustus. Auxiliary soldiers may not regularly have been given citizenship and conubium upon discharge until the time of Claudius; see H. Nesselhauf, CIL, XVI, p. 148 and Birley, E., J. Rom. St., XXVIII (1938), 226Google Scholar. Hoey, p. 29, n. 45, raises the question of the auxiliary calendar before 212.

⁴ On this see Syme, R., The Roman Revolution (Oxford, 1939), 286–287Google Scholar. Augustus' legionaries were of course by no means all Italian in origin. But enough of them were to make it possible to maintain at least “the fiction of a national army,” in Syme's phrase, and the fiction was perhaps inspired more by hope than by fraud. See Syme, op. cit., 456–458 and Rostovtzeff, Storia soc. e econ., dell' Impero Romano, 47–48, where the statements made in the English edition are considerably revised.

⁵ Cf. Aeneid VIII 688: sequiturque nefas! Aegyptia coniunx and Seneca Ep. 83.25: M. Antonium … quae alia res perdidit et in externos mores ac vitia non Romana traiecit quant ebrietas nee minor vino Cleopatrae amor? Legionaries were not regularly given conubium upon discharge, as were auxiliaries in most periods, to legitimatize irregular unions formed with peregrine women.

⁶ Col. II 15. The appearance of Vesta and the legend VESTA P. R. QVIRITIVM in the “Military” coinage of A.D. 68–69 seems to show that her name was thought to have meaning among soldiers at that time; on the coins see Mattingly, H. and Kraay, Colin, Num. Chron. 6th ser. XII (1952), 72–77Google Scholar and 78–86. I do not see why p. R. Quiritium need be specifically “civilian” in tone (see Kraay, p. 84); cf. legiones p. R. Quiritium, found in the Severan acta sacrorum saecularium (C.I.L. VI 32329) and restored with almost complete certainty in the Augustan acta (for the formula see Nock, p. 193). Apparently Vesta and I. O. M. Capitolinus, the only other deity found on these coins, were chosen as symbols intelligible to soldiers of loyalty to Rome and traditional values.

⁷ Nock, pp. 195–197.

⁸ For the predominance of the imperial cult in the Feriale Duranum, see Hoey, p. 173. In the extant portions of the text there are preserved or restored with some probability eight or nine ordinary public festivals which presumably were in the military calendar at the time of Augustus. It may be doubted whether there would ever have been as many as eight entries relating to divus Iulius and Augustus himself, though of course there was no lack of occasions to be celebrated, as e. g. the Feriale Cumanum shows. In any event, there would have been far fewer imperial anniversaries than had accumulated by the third century.

⁹ On the question whether the Feriale Duranum implies a deliberate policy of Romanization cf. Hoey, especially pp. 206–210 and Nock, pp. 203–229.

¹⁰ See Nock, p. 229. Even in official religion the cult of the signa surely had greater meaning for the soldiers than the holidays of the Feriale.

¹¹ Hoey, pp. 116, 169–73; Nock, p. 203.

¹² Suet. Aug. 71; cf. 75: festos et sollemnes dies profusissime, nonnumquam tantum ioculariter celebrat.

¹³ A letter from a governor of Syria which concerned the assignment of a mount was dispatched on IIII Idus Maias, the day of the circenses Martiales (A.D. 208). See Yale Classical Studies, XI (1950), 183Google Scholar.

¹⁴ The date VI Idus Maias in Dura Papyrus 9 recto, line 14 is almost certainly that of the probatio (A.D. 239. Published in Yale Classical Studies, XI (1950), 217Google Scholar). In P. Mich. 455a verso there are entries for. IIII Idus Aug., the natalis divi Iuli. The text is unfortunately not complete and requires further study. In line 10 vel feriatae is read before a lacuna. On the document see Am. J. Phil. LXXI (1950), 437Google Scholar.

¹⁵ See Vegetius II 23 : armaturam, quae festis diebus exhibetur in circo ….

¹⁶ See e.g. the recently published P. Mich. 465 and 466 (A.D. 107). For collections of evidence see Marquardt, J., Römische Staatsverwaltung, II² (Leipzig, 1884), 566–569Google Scholar; Cagnat, R., L'armée romaine d'Afrique, I (Paris, 1913), 356–367Google Scholar; Lesquier, J., L'armée romaine d'Égypte (Cairo, 1918), 227–248Google Scholar.

¹⁷ On leaves see von Premerstein, A., Klio III (1903), 45–46Google Scholar and F. Lammert, R. E., VII A, 2030–31, s.v. Vacatio. Premerstein discusses P. Gen. Lat. 1 verso, part 5, which is a record of assignments. For regular intervals for drill and exercises see Premerstein, loc. cit., 40, n. 3. Scipio Africanus set up a five day cycle which included a day of rest; Polybius X 20.1–4 and other references cited by Premerstein. But there is no evidence to my knowledge of regular days of rest in the imperial army.

¹⁸ For dates of rose festivals, which naturally varied according to locality, see Nilsson, R. E., IA, 1112. Those in the Feriale were perhaps better suited to Italy than some other areas in the Empire; cf. Richmond, I. A., Arch. Aeliana, XXI (1943), 162Google Scholar, n. 89. If Fink's conclusion that the stipendium was paid on May 9 is correct, this might explain the date of the first Rosaliae; see Fink, pp. 70–71. Fink now suggests to me that the date of II 8 may be ṿịị[idu]s maias, which would strengthen the possibility. But see the discussion of the entry for VII Idus lanuarias in the next section.

¹⁹ A suggestion for Col. I 1 was made by Oliver, J. H., Am. J. Arch. XLV (1941), 540–541Google Scholar, but this was shown to be impossible by Fink, R. O., Am. J. Arch. XLVIII (1944), 17–19CrossRefGoogle Scholar. The excellent plates and facsimile permit one to examine readings almost as satisfactorily on them as on the papyrus itself.

There are a number of instances in which the editors disagree among themselves. In the case of Col. I 11–12, which Fink interprets as the natalis of L. Seius and Hoey as that of L. Aelius, I agree with Hoey (p. 183, n. 870) in being skeptical about the existence and readings of the erased letters at the end of line 11. Hoey's reconstruction is accepted by Taylor, , Am. J. Arch. XLVI (1942), 310CrossRefGoogle Scholar and Oliver, Harv. Theol. Rev. XLII (1949), 35, n. 1, and it seems the more attractive, if one must choose between these two alternatives. As regards the date, [. Idus I] ạnuarias in the Feriale instead of Philocalus' Idib. Ian., the error might be that of Philocalus, the manuscripts, or his sources, and not of the papyrus. The next two dates in Philocalus contain a mistake: Kal. Ian. instead of Kal. Febr. (C.I.L. I². P. 255), and a numeral may have been omitted before Id. Ian. in L. Aelius' entry at some stage.

²⁰ Weinstock objects that the entry as reconstructed has no religious content; J. Rom. St. XXXII (1942), 128Google Scholar. The point perhaps escapes me, but I do not see why the honesta missio was not as proper an occasion to be celebrated as Alexander's first consulship (Col. II 18); cf. also dedications made by individual soldiers upon being discharged. Weinstock is also so doubtful about vel in a document of this kind as to regard the passage as corrupt. Like Weinstock, L. R. Taylor rightly, I believe, prefers legio = “legion” instead of privi]legio[rum; Am. J. Arch. XLVI (1942), 310Google Scholar. But the signa of the Rosaliae Signorum are not necessarily legionary and are no support for this view. Cf. e.g. ad signa in the Palmyrene cohort's acta; Yale Classical Studies, XI (1950), 234Google Scholar.

²¹ For dates of discharge see H. Nesselhauf, C.I.L. XVI, p. 186 and Fink, pp. 67–68.

²² Fink, p. 70. The Dura papyrus suggests the restoration Ianuari[arum, rather than -[is. If there were reasons to preserve the restoration of Col. I 7–9, one might explain stip. Kal. Ianuar. in two ways. First, possibly the payment had once been made on the Kalends and the name was retained after the date was changed. The suggestion of an Augustan origin of the date (Fink, p. 69) would then have to be modified. Secondly, perhaps accounts were calculated to the Kalends, but payments were made a week later.

²³ See e.g. Vegetius II 9: … non tantum binae legiones sed etiam plures numeri … ; III 4 : … milites, sive illi legionarii sive auxiliares sive equites fuerint, cum ad expeditionem ex diversis convenerint numeris … ; III 9 : sed cum legiones auxilia vel equites, … singulos numeros …. Numerus is not found in the sense of “unit” before the middle of the first century A.D.; Rowell, H. T., R. E. XVII, 1328Google Scholar.

²⁴ Cf. the table of privileges from Brigetio: … aput signa … quo tam legionarii milites quam etiam equites … ; Ann. épigr. 1937, no. 232 (A.D. 311). January 7, one of the two possible dates of the entry, was the anniversary of Augustus' first imperium, as Fink points out, pp. 66–67. Possibly it is the occasion of the entry. At any rate, the anniversary may well have had a place in the Augustan military calendar.

²⁵ For the position of frg. 5, see Fink, p. 14. The statements there about Col. I 7–10 assume the correctness of the printed restoration.

²⁶ Cf. Snyder, p. 92 and Hoey, p. 198.

²⁷ See Fink, pp. 15–18. The top margin of the fragments seems to show that they come from the first line of a column.

²⁸ See Snyder, p. 153, n. 696. The possibilities he suggests should not be rejected in order to retain frgs. 9 or 10, but the letters extant farther on in the line support neither of them. As Snyder points out (p. 153) iu] ḷiạ ṃ [a] ṃ [aea can be restored in line 27, and it is not impossible that the entry in line 26 was continued.

²⁹ It is noted, but not discussed, by Fink, pp. 77, 82 and Snyder, pp. 152, 156.

³⁰ For Nerva, see Snyder, p. 156. The possibility is entirely based on restorations and is ignored in summaries by Hoey, p. 185 and Nock, p. 188.

³¹ For stress on the nomen Antoninorum see e.g. S.H.A. Caracalla 9.2, Gord. 4.7, and Hoey, p. 186, n. 884. It seems doubtful whether Trajan's dies imperii was added because he was a favorite of Severus Alexander, as Hoey suggests, p. 186. Probably it had been in the calendar continuously since his reign.