^{page 68 note 1} Asbach, Gsell, Dessau (Hermes 46, 1911, p. 159, p. 5, which presumably corrects P.I.R. I, 84) and lastly Gaheis (P.W. x, 130) support 77; Mommsen Furneaux, Haverfield, Weyand (P.W. vi, 2671–2), Sleeman (on Tac. Agr. 9) and especially Mr. J.G.C. Anderson (cl. Rev. 34, 1920, pp. 158–161) prefer 78. But Prof. Haverfield in P.W. viii, 1389—perhaps his last reference—seems to suspend judgment.

^{page 68 note 2} So Borghesi, followed by Hübner and recently by Vaglieri (s.v. Consules in Ruggiero's Dizionario Epigrafico, ii, 1027), interpercts C.I.L. vi, 2016, and Tac. Agr. 17, 3, “et Cerialis quidem alterius sucessoris curam famamque obruisset ….’; and Dessau, , P.I.R. I, 216 (1897)Google Scholar seems to agree; but cf. n. I.

^{page 68 note 3} So Asbach, and Waddington, Fastes, 103, followed by Furneaux, and Kappelmacher in P.W. x, 591–2.

^{page 68 note 4} Dio 52, 23, 2, και άρχέτωσαν μήτε ἕλαττον ὲτῶν τριῶν, εἰ μή τις ἀδικήσειέ τι, μήτε πλεῖον πέντε: in practice the lower limit was usual.

^{page 69 note 1} Mr. Anderson, p. 159 b.

^{page 69 note 2} I have consulted Liebenam, Fasti consulares Imp. Romani, and Vaglieri, s.v. Consules cited above; also Mommsen, Staatsr. Fr. trans, iii, pp. 96–9; esp. p. 97, n. 2, ‘La division de l'année (A.D. 69) en une fraction de quatre mois et quatre fractions de deux mois etablie soit par Néron, soit par Galba, semble avoir servi de modéle à la pratique postérieure.’

^{page 69 note 3} Even if Domitian held office till April 30, Vespasian may have resigned on Feb. 28; and so Agricola might well have been Domitian's colleague. We should rather expect Tacitus to be silent in that case. Suet. Dem. 13 states Domitian's own practice, which favoured short terms.

^{page 69 note 4} Agr. 9, passim.

^{page 69 note 5} The date of the consular comitia under the Flavii is somewhat uncertain, and no strong argument can be based thereon. But since Nero on Jan. 3, 59, was cos. desig. for 60 (Acta Arval.) and under Trajan the suffecti also were elected at the beginning of their year (Plin. Pan. 65 ff, 92) it is very probable that the coss. ordinarii for the next year and the suffecti for the current year were normally chosen together in the first days of Jan. Cf. Mommsen, Staatsr. Fr. trans, ii, pp. 252, n. 5; 253, nn. 4, 5; 255, n. 3. Agricola may have returned to Rome before the end of 76, in good time for that date, though the presence of candidates was not necessary. If he had an interval of private life (Dio 52, 23, 3) it was before the consulship rather than after it.

^{page 70 note 1} Mr. Anderson, l.c.

^{page 70 note 2} Agr. 18. With transgressus so interpreted may we compare the use of proficiscens to cover the whole journey of a governor from Rome to his province ?

^{page 70 note 3} Agr. 28, as interpreted by Mommsen and Furneaux, followed by Mr. Anderson, p. 160 b. So too L. Schmidt, Gesch. der deutschen Stämme ii, 4, p. 412 (in Sieglin's Quellen und Forschungen 29, 30, Berlin 1915–8).

^{page 71 note 1} Tac, . Ann. xiii, 56Google Scholar, 5, ‘Ampsivariorum gens retro ad Usipos et Tubantes concessit’—i.e. from the lands formerly held by the Usipi on the Rhine. Mr. Anderson, p. 160 b, thinks that all the Usipi had been displaced probably by Gemanicus and had gone southward. Schmidt, p. 411, placing all the tribe in the north up to 58, says ‘Bald darauf werden aber auch die Usipier und Tubanten, ohne Zweifel auf Veranlassung der Römer, weiter nach Süden gezogen sein.’

^{page 71 note 2} So Jullian, , Hist. de la Gaule, iii ², p. 46Google Scholar, n. 5, referring to Caes. B.G. iv 4, I; p. 329, n. 7.

^{page 71 note 3} Ib. iv², pp. 104, n. 4; 204, n. 2; 462, n. 3.

^{page 71 note 4} It may perhaps be suggested that the older of the two earthen forts at Marienfels, in the Mühl valley N. of the Wisper, being some distance behind the limes, is a trace of earlier control (C.I.L. xiii 2, 1, p. 478). Cf. von Sarwey in Westd. Zeitschrift, 18, p. 12: ‘Es liegt die Vermutung nahe, dass der Landstreifen entlang des Rheines von der augusteischen Zeit bis zur Einbeziehung in das Reichsgebiet, als Ödland behandelt worden ist, um die Chatten von dem Strom zu trennen.’ Tac. Germ. 29, 4 supports this view; also Front, ., Strat. i, 3, 10Google Scholar.

^{page 71 note 5} Coh. II Mattiacorum appears in a diploma of Moesia Inf. in 99 (Dessau 2000). Veterans of that year must have enlisted not later than 74. We may infer a Coh. I, not yet traced. Cichorius, , P.W. iv, 314Google Scholar; Tac, . Hist. iv, 37Google Scholar.

^{page 71 note 6} Germ. 29, 3; 32, I, ‘Proximi Chattis certum iam alveo Rhenum quique terminus esse sufficiat Usipi et Tencteri colunt.’ But Schmidt, p. 412, boldly assumes that all the Usipi were under Rome from 58 and subject to conscription.

^{page 71 note 7} Schmidt, p. 412, n. 2, says, rather inconsistently, ‘Ob sie im J. 77 mit den Brukterern, die … Gallicus besiegte, verbündet waren, lässt sich nicht mit Sicherheit sagen.’ The omission in the Germania need not be significant, since other tribes also are omitted.

^{page 72 note 1} Plin, . Ep. ii, 7Google Scholar. They were neighbours of the Usipi on the east. Cf. P.I.R. V 308 for date.

^{page 72 note 2} Lehner, Novaesium (reprint from Bonn. Jahrbücher), p. 291; C.I.L. xiii, 4, p. 143.

^{page 72 note 3} Cichorius in P.W. iv, 286; C.I.L. vi, 3230, 3260, 3321 a, 32850, 32866; vii, 68, 427; xiii. 8040; D. 37 = 52 in iii, p. 879; Dessau 4760–1 with reff., esp. Scherer in S.B. Berl. 1884, p. 573. Three singulares of the Frisiavones bear the name T. Flavius. But C.I.L. iii, 4228 records a Cattus serving in an Ala Pannoniorum probably in pre-Flavian times. There are no records of individual Usipi.

^{page 72 note 4} In the former category I reckon 5,000 Batavi, 2,500 Tungri, 2,000 Nervii, 1,000 Cugerni, and 500 each of the Baetasii, Frisiavones, Menapii, Morini, and Sunuci; in the latter, 1,500 Lingones and 1,000 Vangiones.

^{page 72 note 5} Cf. Germ. 34 ‘Romanis classibus navigatos.’

^{page 72 note 6} In Seeck's Notitia Dign. p. 253, ‘Nomina civitatum trans Renum fluvium quae sunt: Usiphorum, Tuvanium, Nictrensium, Novarii, Casuariorum. Istae omnes civitates trans Renum in formulam Belgicae primae redactae. Trans castellum Montiacese LXXX leugas trans Renum Romani possederunt. Istae civitates sub Gallieno imperatore a barbaris occupatae sunt.’ Müller (on Ptol. ii, II, II) and Gsell (Domitien, p. 191) refer this whole passage to the Agri Decumates. So too on better grounds Jullian (op. cit. iv², p. 565, n. 4) who places the Usipi on the Wisper, the Tuvanes at civitas Ulpia Taunensium, and the Nictrenses at col. Ulpia Sueborum Nicretum. The last identification is attractive; but Tuvanes are rather Tubantes, and the Casuarii are also northern—they probably = the Attuari of Velleius, ii, 105. Further, Belgica Prima cannot have been contiguous with the former Germ. Sup. (Notitia Galliarum, ap. Seeck, op. cit. p. 265); it might more probably have been connected with the lower Rhine below Trèves. And surely our passage describes two districts ? The Romans were driven out of the Agri Decumates before Belgica Prima was constituted. Hence I agree with Lehner l.c. in referring the first part of the extract to Germ. Inf. Schmidt, p. 412, agrees with the majority, but ignores the difficulties. The rarity of later mention of individual tribes is due to the substitution of the inclusive name of Franks.

^{page 73 note 1} Martial, vi, 60 (61), with Mr. Anderson's comment p. 161 a. He concludes too readily that 83 is the only possible date.

^{page 73 note 2} If Ptolemy's οὐισπο'. (ii 11, 6) = Usipi, whom he does not otherwise mention, and if he is not simply in error concerning their position (as seems very probable, cf. Müller' s n.), we should have to assume a later migration of the southern section still farther south, or else a third section of the tribe. With the latter alternative the levy might be connected with the annexations of 74. Schmidt, p. 413, says Ptol's mention is ‘nur nach älteren Quellen, an ganz falscher Stelle.’

^{page 73 note 3} Tac. Agr. 39: Mr. Anderson p. 159 b.

^{page 73 note 4} So Gsell, Domitien, p. 184; Weynand in P.W. vi, 2556; Schmidt, pp. 355–7.

^{page 73 note 5} The title appears on coins of Alexandria struck before Aug. 28, 83, with mention of νίκη Γερμανῶν (Stein, in P.W. vii, 1252Google Scholar, quoting Dattan, Numi Augg. Alexandrini. p. 27); and on one of Rome struck before Sept. 13 (Cohen,² i, 520, no. 602); also on coins of 84.

^{page 73 note 6} Cohen,² i, 482, no. 135, and 509, nos. 469 ff. So the authorities cited above.

^{page 74 note 1} C.I.L. xiv, 3612 = Dessau 1025.

^{page 74 note 2} D. 14 in C.I.L. iii = Dessau 1995; Mr. Anderson, p. 161 b; Schmidt, p. 355.

^{page 74 note 3} In Mauretania A.D. 150, Dessau 9056. See infra, p. 76–7 and notes.

^{page 74 note 4} Domaszewski, Philologus 66, 1907, p. 168; Gesch. d. röm. Kaiser, ii, p. 153.

^{page 74 note 5} Dessau, 9200; Ritterling, J.O.A.I. vii, 1904, Beiblatt, 23 ff, followed by Mr. Anderson, p. 161 b, and by Schmidt, without discussion, p. 355.

^{page 75 note 1} Philo, , Leg. ad Gaium, ii, p. 576 MGoogle Scholar; Josephus, , B.J. ii, 500Google Scholar; Dessau, 2726.

^{page 75 note 2} Mommsen, Hermes 19, pp. 437 ff; Ritterling, l.c.; Dessau, 2285.

^{page 75 note 3} Hirschfeld, in C.I.L. xiii, 2, I, pp. 83–4Google Scholar; Schulten, Bonn. Jahrb. 103, pp. 34 ff.

^{page 75 note 4} So Ritterling, l.c.

^{page 75 note 5} Tac, . Hist. iv, 55, 67, 77Google Scholar; Front, ., Strat. iv, 3, 14Google Scholar.

^{page 76 note 1} Von Domaszewski, Phil. 66, p. 166, n. 26, ‘Die Papierabdrücke dieser Ziegelstempel die mir Mowat übersandt hat, bestätigen, wie nicht anders zu erwarten war, die Lesungen und Ergänzungen Mowats. Mit der Ergänzungen Ritterlings fallen auch die weitgehenden daraufgebauten Schlüsse.’

^{page 76 note 2} Tac, . Hist. i, 61, ii, 57, 100, iii, 22Google Scholar.

^{page 76 note 3} Von Domaszewski, Phil 66, 164 ff.; Tschauchner, Legionare Kriegsvexillatonen, 29 ff.

^{page 76 note 4} Tac. Germ. 8, ‘vidimus sub divo Vespasiano Veledam ….’ Statius, , Silv. i, 4, 89Google Scholar, ‘non vacat Arctoas acies Rhenumque rebellem captivaeque preces Veledae …. pandere.’ Dessau, 9052.

^{page 76 note 5} So Gsell, Domitien, p. 180, n. 6, of the Upper German diploma of May 21, 74, when also there was no missio. Weynand, (P.W. vi, 2661–2Google Scholar) partly agrees, though he shows that Clemens' war may then have been over. Groag, (P.W. 2nd series, i, 1260Google Scholar) thinks that the decisive battle was before April 15, 78, but that operations were not yet complete.

^{page 76 note 6} Of twenty-six such documents only two, or perhaps three, record complete missio; in fifteen it is partially, and in eight completely withheld. Contemporary diplomata issued to Italian garrison and to legionaries regularly indicate missio; but they are not here relevant. Cf. Mommsen, in C.I.L. iii, p. 2014Google Scholar. Dessau (on 9052) assumes, probably by oversight, that missio had actually been granted in 78 before the issue of the diploma. Dessau, 2512, is probably a concrete instance in proof of the contrary.

^{page 76 note 7} See P. 74 n. 3 ultra.

^{page 76 note 8} Tac, . Ann. i, 17Google Scholar, ‘ne dimissis quidem finem esse militiae, sed apud vexillum tendentes alio vocabulo eosdem labores perferre’—which Mommsen thinks true of later times also.

^{page 77 note 1} As Ritterling does, Korrespondenzb. d. Westd. Zeitschrift, 1906, 26.

^{page 77 note 2} No special reward is even mentioned by any diploma, Le once (D 10=8 in C.I.L. iii, p. 851), when missio is recorded to have been granted before the time to men who had served loyally in the civil war of 69; the document itself conferring only civitas and conubium as usual.

^{page 77 note 3} Von Domaszewski, , Alterthümer unserer heidnischen Vorzeit, v, 1905, pp. 184–5Google Scholar; Weynand, , P.W. vi, 2671Google Scholar; Groag, , P.W. 2nd series, i, 1260Google Scholar.

^{page 77 note 4} Dessau, 253: if Vespasian was cos. desig. ix at the beginning of Jan. 78 (p. 69, n. 5, supra). Weynand, dating the comitia early in March, makes the time limits imporobably and needlessly narrow.

^{page 77 note 5} C.I.L. iii, 470, 7203, compared with vi, 31548a. In Dessau, 252, ‘imp.xviiii’ should be read likewise; in any case ‘xvii’ is imposibble. The original is lost.

^{page 77 note 6} Tac. Agr. 18.

^{page 78 note 1} Weynand makes his decision against 77 turn upon this question; Gaheis refers to Agricola's modesty, but Mr. Anderson, p. 160, rejects his argument, mainly on the ground that thus a modest or malicious legatus might always cheat the emperor of his due honours. But surely disloyalty in such high command would be as abnormal as Agricola's modesty; and if it did occur, it would hardly thus manifest itself. On the general question of these acclamationes, cf. Mommsen, , Staatsr. Fr. trans, i, pp. 141–4Google Scholar, with notes; v, pp. 42–3.

^{page 78 note 2} Tac. Agr. 24, 25. Sleeman's notes refer to recent discussions.

^{page 78 note 3} Tac. l.c. ‘Saepe ex eo audivi legione una et modicis auxiliis debellari obtinerique Hiberniam posse; idque etiam adversus Britanniam profuturum……’