¹ The sestertius is publised by Kraay, C.M., ‘Two New Sestertii of Domitian’, American Numismatic Society Museum Notes 9 (1960), 109–16.Google Scholar The practice of three pay-days before the raise is documented by P. Gen. Lat. 1 = Fink, R. O., Roman Military Records on Papyrus (1971), 68.Google Scholar The same practice was in force again at least by the late second century as shown by P. Vindob. L 72 and 82 = Fink, Records 71; see Fink's commentary, ibid. p. 253.

² The text given here is Fink's (Records 69); see also his commentary and bibliography.

³ cf. Fink, Records p. 251.

⁴ The text given here is Fink's (Records 68) except for col. ii, 31 where the figure is 343. not 344, as is clear from the additions; see Fink's commentary and extensive bibliography.

⁵ Fink, Records 243.

⁶ Mommsen, Th., ‘Ägyptische Legionare’, Hermes 35 (1900), 443–452Google Scholar = Ges. Schr. vi, 124 ff.

⁷ Mommsen suggested the 75 denarii were reckoned to equal 300 drachmae, but drachmae in billon, worth 6 obols each, i.e. 1,800 obols. But 1,800 obols, reconverted into silver drachmae, worth 7¼ obols apiece, result in 248 drachmae 3 obols, not the 247½ drachmae that P. Gen. Lat. 1 records.

⁸ This was maintained first by Th. Mommsen, l.c. (above, n. 6), and in the latest instance by Marichal, R., ‘La solde des armées romaines d'Auguste à Septime Sévère,’ Annales de l'Institut de Philologie et d'Histoire Orientales et Slaves 13 (1953), 339–421Google Scholar; also by Fink, Records p. 245.

⁹ IGRR i, 1337 gives a long list of soldiers with tria nomina in the cohors I Hispanorum in A.D. 85, cf. Lesquier, , L'armée romaine d'Egypte (1918), 88Google Scholar, n. 5. Likewise, P. Oxy. 1022 of A.D. 103 (= Fink, Records 87) names six recruits of the cohors II or III Ituraeorum, all have the tria nomina. There are other similar cases.

¹⁰ The other four documents on P. Gen. Lat. 1 (= Fink, Records 9; 10; 37; 58) record men with tria nomina exclusively. As has been said, however, this does not prove P. Gen. Lat. 1 to refer to a legion. Fink (Records 9 and 58) assumes the verso to refer to Legio III Cyrenaica, yet conclusive proof is wanting. Even if it could be produced, nothing could be inferred for the recto, on which the pay-records were written some ten years later, for by then the verso was treated as waste paper, and was glued together and recycled (see Fink, Records p. 107). Legionary and auxiliary detachments often shared a camp, and thus might easily have used each other's ten-year-old waste paper.

¹¹ Johnson, A. C., Roman Egypt to the Reign of Diocletian (1936), 670 ff.Google Scholar Ironically, Johnson's thesis may have started from a misunderstanding of Lesquier, L'armée romaine d'Egypte, whom he thought to have opted for five-sixths. Johnson's thesis was adopted by Passerini, A., Le coorti pretorie (1939), 101Google Scholar, n. 2 and Forni, G., Il reclutamento delle legioni da Augusto a Diocleziano (1952), 32 ff.Google Scholar

¹² cf., e.g., Marichal, ‘Solde’ (above, n. 8); Watson, G. R., ‘The Pay of the Roman Army, The Auxiliary Forces,’ Historia 8 (1959), 372–78Google Scholar; Webster, G., The Roman Imperial Army (1969), 260Google Scholar; Brunt, P. A., ‘Pay and Superannuation in the Roman Army,’ PBSR 28 (1950), 50–71Google Scholar; Fink, Records 68.

¹³ Watson, G. R., ‘The Pay of the Roman Army, Suetonius, Dio and the Quartum Stipendium,’ Historia 5 (1956), 332–340Google Scholar, and in The Roman Soldier (1969), 104 ff. suggested such a difference, but Fink (Records p. 245) had rightly pointed out that if this were correct one would expect different sums to be retained at different times, since the expenses they covered varied greatly; also the text says clearly accepit stip(endium). Above all, Watson's thesis would force one to assume an unlikely basic change in book keeping between the date of P. Gen. Lat. 1 and P. Gen. Lat. 4.

¹⁴ p. Hamb. 39 = Fink, Records 76.

¹⁵ Domaszewski, A.v., ‘Der Truppensold der Kaiserzeit,’ Neue Heidelberger Jahrbücher 9 (1899), 218–241.Google Scholar Domaszewski's argument that stipendium denotes 75 denarii was demolished by Brunt, I.c. (above, n. 12), 54. Brunt's own three-fifths thesis collapses with the new date (A.D. 192) for P. Berol. 6866, cf. Fink, Records 70.

^l6 AE 1969–70, 583.

¹⁷ Speidel, M., ‘The Captor of Decebalus,’ JRS 60 (1970), 142–153.Google ScholarBreeze, D., ‘Pay Grades and Ranks below the Centurionate,’ JRS 61 (1971), 130–5Google Scholar even argues that a vexillarius received double pay—but the evidence is tenuous; especially since P. Dura 100 and 101 show that vexillarius was not a permanent rank.

¹⁸ cf. Speidel, I.c. A cut in pay, as suggested by Breeze, I.c., would be totally unparalleled. A time of warfare certainly does not mean ‘abnormal conditions’ for a soldier of Domitian and Trajan, nor does this man's career (or any other) indicate that promotion prospects in the auxilia were better than in the legions. cf. the commentary to AE 1969–70, 583.

¹⁹ The demand of the Batavian cohorts in A.D. 69 for duplex stipendium (Tacitus, Hist. 4, 19) cannot be taken as a demand to double the auxiliary pay in general, nor, therefore, as evidence that it must have amounted to less than half the legionary pay, as Watson, l.c. (above, n. 12), 373, assumes; cf. also Brunt l.c. (above, n. 12), 64 f. It will rather have been a demand for some special favour similar to augeri equitum numerum: perhaps double pay for merit (cf. Dessau 9098, of the time of Septimius Severus), or wholesale promotion in the ranks. For duplicarius as a pay-grade encompassing several ranks, cf. Gilliam, J., ‘The Moesian Pridianum,’ in Hommages à A. Grenier (Collection Latomus 58, 1962), 747–756Google Scholar and Fink, Records p. 16 f.

²⁰ For the text and interpretation see Fink, Records 70.

²¹ No pay-raise seems to have occurred during the second century A.D. before Septimius Severus. cf. Passerini, A., ‘Gli aumenti del soldo militare da Commodo a Massimino,’ Athenaeum 24 (1946), 145–159.Google Scholar

²² In P. Beatty Panop. 2, 36 ff. and 291 ff. the stipendium of the cohorts amounts to twice their annona, the stipendium of the alae to three times their annona. If the annona was the same for both kinds of units, then their stipendium was in a relation of 2:3, which would fit better the two-thirds thesis. But perhaps the annona of the alae was higher, too, so that this ratio need not apply. Jones, A. H. M., The Later Roman Empire (1964) II, 623Google Scholar, commenting on this papyrus, assumes the pay of a miles legionis to equal the pay of an eques alae, and two-thirds of this to be the pay of a miles cohortis. His figures, however, are not meant to be more than hypothetical, because the number of men in each detachment or unit could vary widely. (E.g., the donatives for the alae could be the same as those for the legions; if so, the detachment of the ala II Herculiana dromedariorum in lines 168 ff. was 43 men strong rather than 211, which is not in itself impossible.)

²³ P. Brit. Mus. 2851 = Fink, Records 63; BGU 696 = Fink, Records 64; see Gilliam, I.c. (above, n. 19).

²⁴ Breeze, l.c. (above, n. 17).

²⁵ We follow Watson's convincing proposition that the figures be divisible by 25 so as to be accountable in aurei as was the viaticum (in BGU ii, 423 and P. Berol 6866) and the legionary pay throughout. This assumption is now confirmed by the new figures for auxiliary pay: 7½ aurei before A.D. 84 and 10 aurei thereafter.

²⁶ cf. Forni, I.c. (above, n. 11), 33. 25% of the recruits for the alae on Rhine and Danube were citizens by Trajan's and Hadrian's time, as is revealed with some precision by the number of equites singulares Augusti that do not change their names to that of the ruling Emperors, cf. Speidel, M., Die Equites Singulares Augusti (Bonn, 1965 = Antiquitas ii, 11) 67.Google Scholar

²⁷ Deposits: P. Fay. 105 = Fink, Records, 73. Wife and concubine: P. Wisc. 14, cf. Gilliam, J., ‘P. Wisconsin 14,’ BASP 5 (1968), 91–8.Google Scholar

²⁸ Extortion: Luke 3, 14; trade: Tacitus, Ann. 13, 35 and 51 ; other business: e.g. BGU ii, 462 = Wilcken, Chrest. 376. The argument that soldiers had a rather low standard of living (Forni, l.c. above, n. 11), does not convince, cf. Brunt, I.c. (above, n. 12), 65 f.

²⁹ For the effect of military expenses in the third century see Pekary, Th., ‘Studien zur römischen Währungs- und Finanzgeschichte,’ Historia 8 (1959), 443–489.Google Scholar A short version of this paper was presented at the Thirteenth International Congress of Papyrologists, Marburg, in August, 1971. I am grateful for the enlightening remarks offered by my colleagues there, but especially to Prof. E. Birley, Durham, and Dr. J. F. Gilliam, Princeton, who read the manuscript and made valuable suggestions.