[13] That this was the custom followed for the socii and auxilia during the period of the republic is suggested by Pliny, N. H., 25, 33, 6, and the same general proportion seems to have been observed later, as Tacitus, Ann., IV, 5, in speaking of the “sociae triremes alasque et auxilia cohortium,” adds, “neque multo secus in numero virium.” Detailed information regarding the size of these auxiliary contingents is nowhere given. See Liebenam, art. “Exercitus,” Pauly-Wiss., VI, 1601, 1607. G. L. Cheesman (The Auxilia of the Roman Imperial Army, Oxford, 1914 p. 53 ff.) finds that the auxilia under Augustus were at least as numerous as the legionaries, and later became more so. He calculates 180,000 for the year 69 A. D., and 220,000 for the middle of the second century A. D. Cf. also Delbrück, op. cit., II, p. 203 (2nd ed.).
[14] Cf. Ed. Meyer, ibid., p. 909. The evidence for the enrollment of foreigners in the legions at this time is conveniently summarized by Liebenam, art. “Dilectus,” Pauly-Wiss., V, 611 ff.
[15] Mommsen, Eph. Epigr., V (1884), p. 159 ff.; Hermes, XIX (1884), p. 1 ff., esp. p. 11.
[16] We must remember that this restriction in the recruiting sources of the legionaries was wholly an act of free choice on the part of Augustus, whatever the motive may have been. That suggested by Seeck, l. c., p. 611, does not seem very probable; it involved a change in the usage to which men had already become accustomed in the civil wars, and it was gradually but completely abandoned by his successors. There was nothing in the general conditions which required it.
[17] Rh. Mus., XLVIII (1893), p. 602 ff. His conclusions in part rest on none too certain foundations, and introduce an insufficiently motivated complexity in the system of levying troops, for Augustus at the beginning of his career used non-citizen soldiers freely, and after the defeat of Varus, of the two new legions which were raised one was a Galatian contingent, the Deiotariana, which was given citizenship and a place in the army (Mommsen, Res Gestae Divi Augusti, 2nd ed. (1885), p. 70; O. Seeck, Gesch. d. Untergangs d. ant. Welt, 3rd ed., I, p. 260), and the other was recruited from the non-citizen population of Rome (Tac., Ann., I, 31, “vernacula multitudo”; cf. Mommsen, Hermes, XIX (1884), p. 15, n. 1). Seeck’s statement of the system which he believes Augustus followed is: “Prätorianer und Stadtsoldaten rekrutirten sich aus Latium, Etrurien, Umbrien, und den frühesten Bürgercolonien; den übrigen Italikern sind die Legionen zugewiesen, den Bürgern der Provinz die Freiwilligencohorten; aus den Libertinen setzen sich die Mannschaften der Flotte und der Feuerwehr zusammen; die Nichtbürger bilden Cohorten und Alen und einen Theil der Flotte.”
[18] In B. C. 8 it was 4,233,000; in A. D. 14, 4,957,000. See the Mon. Anc., 8. Of course if we accept the view still defended by Gardthausen and Kornemann that this number represented only the male population (Augustus, II, 532; and Jahrb. für Nationalökonomie u. Statistik, III, 14 (1897), p. 291 ff.), a citizen army of more than a million men might have been raised, but the view of Beloch and Ed. Meyer that the numbers in the Mon. Anc. include women and children seems the only one possible. See Meyer’s complete refutation of Kornemann, Jahrb. für Nationalökonomie, III, 15 (1898), p. 59 ff.
[19] This figure, 800,000, is modest, amounting to roughly 1½% of the total population, about the same proportion which Germany and France have for some time past kept under arms in time of peace, while their war strength is several times as great as this. Rome did actually at one time, the crisis of the Second Punic War, have at least 7½% of her total population in the field, even according to the most conservative estimates. Cf. H. Delbrück, Gesch. d. Kriegskunst, I, 2nd ed. (1908), pp. 349, 355 ff.
[20] Mommsen, Hermes, XIX (1884), p. 3, n. 3, gives the number of Roman citizens who were engaged in the war between Octavian and Antony as 300,000, which makes a total of 600,000 troops or more.
[21] The details in Marquardt, Röm. Staatsver., V, 2 (2nd ed., 1884), p. 444 f.
[22] Or possibly 28; see von Domaszewski, “Zur Geschichte des Rheinheeres,” Röm.-Germ. Korrespondenzblatt, 1910, on the date of the establishment of the twenty-first and twenty-second legions. There is some question about the exact date at which the increase in the size of the legions was made (see the literature cited by Gardthausen, Augustus, II, p. 775), but that does not affect our argument. See above [note 7].