^{page 17 note 1} Tac. Amm. i. II, ‘addideratque consilium coercendi intra terminos imperii.’

^{page 17 note 2} Suet. Galba, I.

^{page 17 note 3} ‘Legatus exercitus qui est in Germania superiore’ was the official title.

^{page 17 note 4} Ubii and Sugambri, in Lower Germany; Nemetes, vangiones, and Triboci in Upper Germany.

^{page 17 note 5} Tac. Ann. i. 56, ‘tumultuarias catervas Germanorum cis Rhenum colentium.’ Comp. id. Ann. xii. 27.

^{page 18 note 1} The foundation of Cologne necessitated the removal of the two legions hitherto stationed ‘in finibus Ubiorum,’ and the formation of two camps, one above and one below the new colony.

^{page 18 note 2} Tac. Germ. 41.

^{page 18 note 3} Under Nero the Frisii attempted to settle on the forbidden territory, but were evicted. Ann. xiii. 54.

^{page 18 note 4} See below, p. 27.

^{page 18 note 5} In 47 A.D. Claudius ordered the withdrawal of all garrisons beyond the Lower Rhine (Tac. Ann. xi. 19). There seems to be evidence that the early Roman fort recently excavated at Haltern, on the Lippe, was abandoned at this time.

^{page 18 note 6} Germ. 29. The ‘province’ must be Upper Germany; see below, p. 31.

^{page 18 note 7} The ‘laterculus Veronensis,’ Seeck. Notit. Dign. p. 253. It gives the names of 5 ‘civitates’ beyond the Rhine and adds, ‘trans castellum Mogontiacense lxxx leugas trans Renum Romani possederunt. Istae omnes civitates sub Gallieno imperatoie (253–268 A.D.) a barbaris occupatae.’

^{page 19 note 1} Strategemata, 1, 3, 10; 2, 3, 23; II,

^{page 19 note 2} Der Römische Grenzwall in Deutschland (Wiesbaden, 1884)Google Scholar. In the late summer of 1884 two Oxford men, Mr. Mowat, Fellow of Pembroke, and Mr. Crowder, Fellow of Corpus, walked the whole length of the barrier. A brief account of the walk was printed by Mr. Mowat (Oxford, 1885).

^{page 20 note 1} Major-General von Sarwey and the late Professor Hettner.

^{page 20 note 2} In addition to the report published annually in the Archäologischer Anzeiger details of the work of the commission are given in the official Limes-Blatt (Trier, 1892–1903)Google Scholar and in Ober-Germanisches-Rätisches Limes, of which 25 parts have now been published. The pottery has been treated by Koenen, , Gefässkunde ind. Rheinlanden (1895)Google Scholar and by Dragendorff, in the Bonner Jahrbuch (1895Google Scholar and following vols.) The inscriptions are given in Corp. Inscr. Lat. xiii. part 2, fasc. I (1905). Excellent summaries will be found in Koepp, , Die Römer in Deutschland (Leipzig, 1905), pp. 56–86Google Scholar, and in Fabricius, , Die Besitznahme Badens durch die Römer (Heidelberg, 1905)Google Scholar.

^{page 21 note 1} Germ. 29. ‘Non numeraverim inter Germaniae populos, quamquam trans Rhenum Danubiumque consederint, eos qui decumates agros exercent. levissimus quisque Gallorum … dubiae possessionis solum occupavere, mox limite acto promotisque praesidiis sinus imperii et pars provinciae habentur.’

^{page 21 note 2} Fabricius, , Baden, pp. 55, 58Google Scholar, takes Tacitus to refer to the whole area, ‘which projected like a peninsula into the sea of barbarism,’ a very doubtful interpretation of ‘sinus imperii.’ He also takes ‘dubiae possessionis’ as referring to the tenure on which the ‘coloni’ held their lands from the Roman government after the annexation.

^{page 21 note 3} Tac. Germ. 28. A dedication ‘deo Santio’ was found at Miltenberg. C. I. L. 13, 6607. Domaszewski, , W. D. Z. 1902Google Scholar, connects the ‘Cubii’ of Frontinus, , Strateg. 2, 11, 7Google Scholar. with the Bituriges Cubi in Gaul.

^{page 21 note 4} Ladenburg, Rottenburg, and Zarten.

^{page 21 note 5} As at Zarten and probably at Rottweil. Fabr, . Baden, pp. 16, fgGoogle Scholar.

^{page 22 note 1} E.g. the ‘deus Santius’ already referred to.

^{page 22 note 2} C.I.L. 13, 6448 (at Benningen), ‘deanæ explorator(es) Triboci et Boii.’

^{page 22 note 3} At Ladenburg, the ‘civitas Ulpia Sueborum Nicretium.’ Ibid. 13, 6404, 6420, 6421.

^{page 22 note 4} Near Miltenberg. Ibid. 13, 6610, ‘inter Teutones.’

^{page 22 note 5} Ibid. 13, 6604, also at Miltenberg.

^{page 22 note 6} Legio II. Augusta was sent to Britain. See Fabricius, , d. Besitznahme Badens durch d. Römer, pp. 30, 34Google Scholar.

^{page 22 note 7} Germ. 29, ‘limite acto promotisque praesidiis.’

^{page 23 note 1} Zangemeister, , N. Heideldb, Jahrb. 1893Google Scholar.

^{page 23 note 2} In 72 Vespasian was Imperator X. (Ephem. epig. 4, no. 807); in a diploma of May 21, 74, he is Imp. XIII., and before the end of that year Imp. XIV. See Zangemeister, I.e.

^{page 23 note 3} C.I.L. xi. 5271.

^{page 23 note 4} Ibid. xi. 5201, 5211. Cn. Domitius Tullus and Cn. Domitius Lucanus, the adopted sons of the orator Domitius Afer. Each is described as ‘præfectus auxiliorum omnium adversus Germanos.’

^{page 23 note 5} Fabricius, , Baden, p. 34Google Scholar.

^{page 24 note 1} Zangemeister, l.c., ‘iter derectum ab Argentorate in R(ætiam),’ or, as Domaszewski suggests, ‘in r(ipam Danuvii).’

^{page 24 note 2} The stations are as follows, Juliomagus = Schleitheim. Brigobanna = Hüfingen. Aræ Flaviæ. At the two former places remains of Roman buildings, tiles of leg. XI., and early Flavian pottery have been found.

^{page 24 note 3} Waldmössingen, on the watershed between Kinzig and Neckar, was originally an earth fort, but was rebuilt in stone. Fabr, . Baden, p. 40Google Scholar.

^{page 24 note 4} The large irregularly shaped enclosure found at Rottweil was at first taken to be the Roman fort (W.D.Z. 20, 178). It is now regarded as belonging to a pre-Roman Celtic settlement. Traces of the Roman fort have been found within it. Fabr, . Baden, pp. 16, 37Google Scholar.

^{page 24 note 5} We may possibly connnect Aræ Flaviæ with the formation of a province ot G ermania Superior by Domitian, (Beitr. z. Alt. Gesch. i. p. 133)Google Scholar.

^{page 25 note 1} For a general survey of our knowledge respecting the extent and organisation of the ‘saltus Cæsaris’ see Hirschfeld, , d. kaiserlichen Verwaltungs-Beamten (Berlin, 1905), p. 125Google Scholar, and Beitr. z. Alt. Gesck. 2, p. 284.

^{page 26 note 1} Inscrr. Grac. ad R. hist. pert. iii. I, 70. έπίτροπον Σεβαστοὖ⋯ χώρας Σ⋯μελοκεννησ⋯ας κα⋯ ύπερλιμιταν⋯ς (cf. Mommsen, in W.D.Z. 1886, p. 260)Google Scholar.

^{page 26 note 2} χώρα = ‘tractus’ or ‘regio’ rather than ‘saltus.’ Each ‘saltus’ had also its own procurator.

^{page 26 note 3} C.I.L. 13, 6365, ‘ex decreto ordinis saltus Sumelocennensis.’

^{page 26 note 4} It has been suggested that ‘decumates’ is a Celtic form of the orthodox Latin ‘decumani.’ Fabr, . Baden, p. 58Google Scholar.

^{page 26 note 5} Plin. N. H. xxxi. 30, mentions the Wiesbaden springs (‘aquæ Mattiacæ’) ‘fontes calidi trans Rhenum’; for the silver in the Taunus see Tac. Ann. xi. 20 The pottery, &c., found at Wiesbaden point to its having been in Roman occupation during the principate of Augustus, . Mittheil. d. Nassauer Alter' thumsvereins, 1904–1905, pp. 4, 36, 43Google Scholar.

^{page 26 note 6} As in the time of the elder Drusus. Dio. 54, 33.

^{page 27 note 1} Tac. Hist. iv. 37.

^{page 27 note 2} Tac. Germ. 30, ‘alios ad prœlium ire videas, Chattos ad bellum.’

^{page 27 note 3} A permanent bridge was erected at least as early as 56 A.D. C.I.L. 13, 6820.

^{page 27 note 4} ‘Castellum Mattiacorum,’ orig. built by Drusus, , έν χ⋯ττοις παῤ αύτῷτῷ ρήνῳ. Dio. 54, 33Google Scholar. Koepp (p. 38) wrongly indentifies it with the ‘presidium in monte Tauno.’

^{page 27 note 5} The remains at Höchst indicate its occupation in the Augustan age, while those at Hoiheim belong to the time of Gaius or Claudius. Bericht über d. Fort-schritte d. röm.-german. Forschung (Frankfort, 1905) p. 29Google Scholar.

^{page 27 note 6} Tac. Ann. i. 56.

^{page 27 note 7} Limes-Blatt, p. 525. It was at a later time replaced by-a stone fort of the normal size (157 m. x 144 m.)

^{page 27 note 8} He had at least four legions—8th Augusta, 14th Gemina, 21st Rapax, and the 11th Claudia.

^{page 27 note 9} Frontinus, , Strat. i. 1, 8Google Scholar, ‘bellum inopi???atum.’

^{page 27 note 10} Ibid. i. 3, 10.

^{page 28 note 1} Frontinus, , Strat. i. 3, 10Google Scholar, ‘limitibus per cxx m.p. actis, subjecit ditioni suaæ hostes quorum refugia nudaverat.’

^{page 28 note 2} Excellent summaries of the results obtained by the Reichslimes-Kommission on the Taunus are given by ProfKoepp, , D. Römer in Deutschland, p. 60 foil.Google Scholar; Fabricius, , Baden, pp. 47–55Google Scholar; and W.D.Z. 1901, p. 177 foil. For the formation of this earliest ‘limes’ see also Arch. Anz. 1900, 2, pp. 94, 95.

^{page 28 note 3} E.g. at Saalburg, Capersburg, Feldburg, Zugmantel. The earth fort at the Saalburg measures 83 m. × 78 m., the later stone fort 147 × 221; at Capersburg the earth fort is 89 × 89, the stone fort 122 × 134.

^{page 28 note 4} For a description of these towers and of the frontier line see Limes-Blatt, pp. 617–647. The wooden towers rested on stone foundations and were surrounded by palisades. They must have closely resembled those represented on the Trajan column as guarding the Danube frontier. For the evidence that the so-called ‘Begleithiigeln’ marked the site of wooden towers see Arch. Anz. 1899, 2, 84. In some cases in the Taunus and on the Odenwald line the towers were built in the Gallic fashion(Cæs. B. G. 7, 23) ‘alternis trabibus et saxis.’ Between the towers in some cases there was a light fence. Arch. Anz. 1900, 2, 95, 96.

^{page 28 note 5} Arch. Anz. 1904, p. 152. The stone fort measured 282 m. × 185 m. Okarben measured 295 × 192, Friedberg 245 × 154, Kesselstadt 375 × 375. The legionary camp recently excavated at Neuss is 571 × 432.

^{page 29 note 1} Fabricius, , Baden, p. 51Google Scholar, notes that the frontier line from the Rhine at Rheinbrohl to the junction of the Main and Kinzig measures roughly 120 Roman miles. The ‘80 leugse’ of the Verona list must be derived from the 120 Roman miles of Frontinus.

^{page 29 note 2} The date is fixed by the Ada Fr. Arval. (ed. Henzen, ), p. cxxiGoogle Scholar. Suet. Dom. 7, ‘aput duarum legionum hiberna.’ The place was certainly Mainz, not, as Mommsen thought, Windisch.

^{page 29 note 3} See the ingenious arguments of Ritterling, , W.D.Z. 1893Google Scholar; cf. Koepp, , d. R. in Deutschl. p. 63Google Scholar.

^{page 29 note 4} O.R.L. 16. The apparent destruction of Okarben, in the Wetterau, and also of the small frontier forts and towers between the Aar and the Lahn, suggests the inference that the Chatti moved down both the Main and Lahn valleys to the Rhine. Arch. Anz. 1900, 2; 81. ‘Die Holzthiirme auf der Rheinischen Strecke siidlich der Lahn bald nach ihrer Erbauung niedergebrannt und an den gleichen Platzen erneuert worden sind.’

^{page 29 note 5} Suet. Dom. 6: ‘ipsa dimicationis hora resolutus repente Rhenus.’

^{page 30 note 1} Both were sent to the Danube; the fate of the 21st legion is uncertain, but possibly it was the legion which was cut to pieces in the Suebo-Sarmatian war, 89–93 A.D. Suet. Dom. 6.

^{page 30 note 2} For this legion see Weichert, , W.D.Z. 1902, p. 124Google Scholar. 3 Germ. 29.

^{page 30 note 3} A ‘cohors II. Mattiacorum’ was in Lower Moesia in 99 A.D. C.I.L. 3, Suppl. diploma xxxi.

^{page 30 note 4} Tac. Agric. 28. Cohors Usipiorum per Germanias conscripta.

^{page 30 note 5} It is possibly to the forts built on this section of the frontier that Frontinus (2, 11, 7) refers when he says that Domitian paid a price for the ground he took ‘quum in finibus Cubiorum castella poneret,’ and thus won the confidence of the natives. See Domaszewski, , W.D.Z. 1902Google Scholar; Fabr, . Baden, pp. 19, 52Google Scholar.

^{page 30 note 6} The fort at Kongen is not later than the end of the first century, but very probably belongs to the time of Domitian rather than of Vespasian. Arch. Anz. 1896, p. 190. Fabricius thinks that the earth fort at Siilz marked the northern limit of Vespasian's annexation.

^{page 31 note 1} Fabricius, , W.D.Z. 1901, p. 183Google Scholar; Limes-Blatt, pp. 444, 464, 497, 527. The fort at Hesselbach measures 82 × 72 m., that at Wiirzberg 79 × 71.

^{page 31 note 2} Oberscheidenthal is the first of these larger ‘cohort forts.’

^{page 31 note 3} For Gross Gerau see Livm-Bl. p. 848 and C.I.L. 13, 2, 1, p. 234, and for Neuenheim Ibid. p. 224. Both were cohort forts of the normal size.

^{page 31 note 4} C.I.L., 13, 6298.

^{page 31 note 5} Fabricius, , Baden, p. 70Google Scholar.

^{page 31 note 6} C.I.L. 11, 5271. Cornelius Clemens is ‘legatus pro. pr. exercitus qu(i est in Germania Sup.)’

^{page 32 note 1} C.I.L. 3, 2864. Possibly the selection of a distinguished jurist to be the first governor of the province is significant.

^{page 32 note 2} Ibid. 13, 2, 1, p. 426; Fabr. Baden, p. 63.

^{page 32 note 3} Ibid. 13, 7566. ‘Vicani Aquenses’ Kastel (= castellum Mattiacorum), or rather the ‘vici’ outside the fort, would seem also to have belonged to the ‘civitas Mattiacorum.’

^{page 32 note 4} Ibid. 13, 6404, 6421. The statement of Eutropius, 8, 2, ‘urbes trans Rhenum in Germania restituit’ (Trajan), may refer to his organisation of such ‘civitates.’

^{page 33 note 1} C.I.L. 13, 7265, 7352, 7370.

^{page 33 note 2} Heddernheim after the removal of the troops developed into a large ‘vicus,’ with walls 2,700 m. in circuit. Ibid. 13, 2, 1, p. 426. Much the same happened at Wimpfen. Arch. Am. 1899, 2, p. 90.

^{page 33 note 3} C.I.L. 13, 6482, ‘genio civitatis Alisin(ensium).’

^{page 33 note 4} Ibid. 13, 6339.

^{page 33 note 5} In Ibid. 13, 6365, we have still the ‘saltus Sumelocennensis,’ but with an ‘ordo,’ a local council of elders. A later stage is marked by 6384, ‘decurio ciyitatis Sumel.;’ cf. 6358. That Kongen (Grinario) was a ‘vicus’ within the domain of Sumelocenna is proved by an inscription found there a few years ago. W.D.Z. 1900, Korr.-Bl. p. 16.

^{page 33 note 6} Hirschfeld, , d. kaiserl. Verwalt-Beamten, p. 143Google Scholar. Premerstein, in Beitr. z. Alt. Gesch. 3, p. 28Google Scholar; cf. an inscription found near the site of the legionary fort at Regensburg, Archaol. Anz. 1900, p. 25, ‘Aurelius Artissius sedilis territorii contrarii (i.e across the Danube) et K(astrorum) R(eginorum).’

^{page 34 note 1} C.I.L. 13, 6541; O.R.L. pt. 5.

^{page 34 note 2} Tac. Agric. 41: ‘tot exercitus in Mœsia Daciaque…et Pannonia… amissi.’ Cf. Hist. i. 2, ‘coortae in nos Sarmatarum ac Sueborum gentes,’

^{page 34 note 3} Vespasian enlarged the camp at Carnuntum. The camps at Vindobona, Ratiaria, and Œscus were apparently established under the Flavian emperors.

^{page 34 note 4} C.I.L. 13, 2, 1, p. 37.

^{page 34 note 5} Fabricius, Baden, p. 41 sqq Cannstatt was clearly an important station. The fort is one of the larger ones (217 × 178 m.). That it was a road centre is implied by the dedication ‘biviis triviis’ (C.I.L. 13, 6437) and the presence of transport officers, ‘beneficiarii’ (Ibid. 6442).

^{page 35 note 1} Apparently the fort here dates from Domitian, and it seems to have been abandoned about 150 A.D. See O.R.L. pt. 24.

^{page 35 note 2} The fort at Aalen, a large one (288 m. × 214 m.), was probably built early in the second century. It commands not only the route westward down the Rems valley, but routes northward down the Kocher to the Neckar, and southward up the Kocher and then down the Brenz to the Danube. See O.R.L. pt. 23.

^{page 35 note 3} Aur. Victor, , De Cæs. 13Google Scholar: ‘iter conditum per feras gentes quo facile abusque Pontico mari in Galliam permeatur.’

^{page 35 note 4} C.I.L. 3, Suppl. 8267: ‘Montibus excisis anco(ni)bus sublatis via(m) f(ecit).’

^{page 35 note 5} O.R.L. pts. 6 and 9.

^{page 35 note 6} Hadrian visitfd Germany in 121 A.D.

^{page 36 note 1} Arch, Anz. 1902. A good instance is the Saalburg.

^{page 36 note 2} The stone towers are clearly older than the earthen barrier, behind which they stand at varying distances (7 yds. to 33 yds.) Livies-Bl. p. 721. They are on an average about 5 yards square.

^{page 36 note 3} The evidence is summed up in Arch. Anz. 1896, p. 177. The stone towers seem clearly to belong to the new line, the wooden ones to the old. The interval between the old and new line is in some places as much as 700 m. (Ibid. 1900, 2, p. 82), and the former is often traceable only by the positions of the wooden towers. No certain traces of such towers have been found either on the new line in the Taunus and Wetterau, or on Pius' new frontier from Walldiirn to Haghof (Arch. Anz. 1900, 2, p. 93).

^{page 36 note 4} Fabricius, , W.D.Z. 1901, p.185Google Scholar.

^{page 37 note 1} Vit. Hadriani 12, ‘in plurimis locis in quibus barbari non fluminibus sed limidbusdividuntur, stipitibus magnis in modum muralis xepisfunditus jactis atque conexis barbaros separavit.’

^{page 37 note 2} Limes-Blatt, p. 483 sgq. The palisade was a very substantial obstacle. It stood nine feet high. For the upright stakes, as Spartianus calls them, the trunks of oak trees have been split in two, and set up with the flat face towards the ‘outland.’ Their average thickness is 23–29 centimetres and breadth 37–54 cm. They were, as Spartianus says, ‘funditus jactis’ in a ditch 4½ feet deep, and ‘conexis’ (bound together) on the inside by strong cross beams.

^{page 37 note 3} The theory that this ditch merely represented the original line of demarcation drawn by the surveyors, and then filled in, has now been abandoned. See Kohl, , Limes-Bl. p. 487Google Scholar.

^{page 37 note 4} Along the Odenwald line the palisade runs in front of the line of forts at an average distance of about 30 yards. Ibid. p. 465.

^{page 37 note 5} In this section the later ‘barrier,’ the ‘Pfahlgraben,’ runs as a rule in front of the forts, but behind the palisade (see O.R.L. pt. 20, ‘Gross Krotzenburg,’ p. 3), which it may have been intended to supplement, not to supersede (Limes-Bl. p. 720). For an exception to this rule see Arch. Anz. 1893, p. 3. For an earlier view see Ibid. 1899, p. 78, but Hettner(Ibid. p. 79) regards the palisade as ‘allezeit ein unentbehrlicher Theil d. Grenzwehr.’

^{page 38 note 1} Tac. (Germ. 41), speaking of the exceptional privileges granted to the Hermunduri: ‘non in ripa commercium sed penitus passim et sine custode transeunt.’

^{page 38 note 2} Dio. 73. 2. The Germans were to meet only once a month, at a fixed place and under the supervision of a Roman centurion.

^{page 38 note 3} E.g. at Markobel, O.R.I., 3, pp. 16, 17; Arnsburg, Ibid. 17, p. 18.

^{page 39 note 1} Fabricius, , Baden, pp. 78–82Google Scholar.

^{page 39 note 2} Mommsen, , Hermes, 19, 219Google Scholar.

^{page 39 note 3} Now also, if not earlier, the small earth forts on the Odenwald were re-placed by stone ones, which, however, are also small forts with only three gates. Litnes-Blatt pp. 474, 550, 772. The inscriptions give the dates 145 and 146 A.D. C.T.L. 13, 6501, 6502.

^{page 39 note 4} E.g. near the fort of coh. III. Aquitanorum at Neckarburken was a small ort (79 m. × 79 m.)garrisoned bya ‘numerus Brittonum Elantiensium.’ O.R.L. pt. 9.

^{page 39 note 5} The ‘Brittones Triputienses’ are found along the Odenwald line, C.I.L. 13, 6502, 11; the Elantienses (? from the river Elz) near Neckarburken, Ibid. 13, 6490, 8; the Murrenses (from the river Murr) near Benningen, Ibid. 13, 6471; cf. ‘Vicani Murrenses’ 6454; for Gr(inarionenses) (Kongen) and L(inenses) (? Lyn), see C.I.L. 2, 1, p. 254, Limes-Bl. p. 824.

^{page 39 note 6} Baden p. 80.

^{page 39 note 7} ‘Ad tutelam lirnitis;’ cf. the inscr. Archaol.-epig. Mittheil. 17, p. 103.

^{page 40 note 1} C.I.L. 13, 6501.

^{page 40 note 2} Ibid. 13, 6561.

^{page 40 note 3} Walldiirn is the exception, being only 96 m. × 84 m.

^{page 40 note 4} See the official reports on Osterburken (O.R.L. pt. 2), Öhringen (ibid. pt. S), Murrhardt (ibid. pt. 1), Welzheim (ibid. p. 21). Whetherthere were at first wooden towers along this line is still a matter of dispute (Arch. Anz. 1898, p. 5), but that Pius' line consisted of forts, watch towers, and palisade seems certain (ibid. 1901, 2, p. 83). The towers on this frontier are at intervals of from 250 to 400 m.

^{page 40 note 5} Brambach, , C.I.R. 1643Google Scholar.

^{page 40 note 6} O.R.L. pt. 21. The ‘ala I.’ whose tiles were found at Welzheim is almost certainly the ‘ala I. Flavia,’ which is given as stationed in Upper Germany id the diplomata of 74, 82, 90, and 116.

^{page 40 note 7} C.I.L. 13, 6077.

^{page 40 note 8} Ibid. 6542–3.

^{page 41 note 1} For Urspring see Fabricius', report in the Bericht iiber d. Fortschritte d. Röm. Germ. Forschung, 1904, p. 31Google Scholar; for Faimingen, , Limes-Blatt, p. 918Google Scholar.

^{page 41 note 2} O.R.L. pt. 13. It measures 271 m. × 195 m., is nearly as large as Aalen, and was an important road centre. Arch. Anz. 1896, p. 188.

^{page 41 note 3} O.R.L. pt. 23. It was garrisoned by ‘alall. Flavia miliaria.’ For the importance of its position see above, p. 35, note 2.

^{page 41 note 4} This route is probably as early as Hadrian. Arch. Anz. 1900, 2, p. 95. Hadrian repaired the road from Rottenburg to Kongen. W.D.Z. 1900, Korr.-Bl. p. 16.

^{page 41 note 5} Possibly an early (Trajan) frontier line existed from a point N. of Lorch to Haghof and thence to Benningen. Arch. Anz. 1900, 2, 94. The frontier line from Lorch to Haghof has not the rigid straightness of that from Haghof to Walldiirn and is, therefore, probably older. Arch. Anz. 1903, p. 115.

^{page 42 note 1} Vit. Marci, 21, 22.

^{page 42 note 2} Ibid. 8; Vit. Did. Jul. I.

^{page 42 note 3} O.R.L. part 14; C.I.I.. 3, Suppl. 11933, a dedication to Commodus by coh. I. Breucorum, the garrison of Pfünz.

^{page 42 note 4} Limes-Blatt, p. 884, ‘vexillarii leg. III. Italirae vallum fecerunt…item portas cum turribus IIII’ (181 A.D.)

^{page 42 note 5} Regina, Castra C.I.L. 3Google Scholar, Suppl. 11965. Leg. III. Italica was stationed there.

^{page 42 note 6} Ibid. 13, 2, 1, p. 276, and nos. 6578, 6582.

^{page 42 note 7} To Welzheim, Öhringen, Walldürn, Miltenberg. See C.I.L. 13, 2, 1, p. 264.

^{page 42 note 8} Ibid. 13, 2, 1, p. 499; Limes-Blatt, p. 890.

^{page 42 note 9} Act. Fr. Arv. (ed. Henzen, ), p. 81Google Scholar, ‘per limitem Rsetiae ad hostes extirpandos.’

^{page 43 note 1} Act. Fr. Arv. (ed. Henzen, ), p. 81Google Scholar.

^{page 43 note 2} Fit. Carac. 10; Aur. Viet. De Cæ. 21, ‘Alamannos prope Mœnum devicit.’

^{page 43 note 3} At Gilsland, in Cumberland, the ‘vallum’ is carried across a piece of boggy ground on ‘a layer, averaging 12 feet wide, of large freestones, cobble, and gravel.’ Haverfield, , Hep. on Wall of Pius, append, iii. p. 171Google Scholar.

^{page 43 note 4} Arch. Anz. 1899, p. 80.

^{page 43 note 5} Ibid. p. 79.

^{page 44 note 1} Vit. Alex. 59, ‘Germanorum vastationibus Gallia diripiebatur.’

^{page 44 note 2} Vit. Maximini, 10, 11.

^{page 44 note 3} Brambach, , C.I.R. 1963Google Scholar.

^{page 44 note 4} C.I.L. 13, 6552.

^{page 44 note 5} Ibid. 6566.

^{page 44 note 6} Ibid. 7467.

^{page 44 note 7} Ibid. 7352.

^{page 44 note 8} Ibid. 7424.

^{page 44 note 9} Ibid. 7754.

^{page 44 note 10} Ibid. 7440.

^{page 45 note 1} W.D.Z. 1903, p. 137.

^{page 45 note 2} A centurion of the 8th legion commanded the 1st Helvetian cohort at Bockingen in 148 A.D., C.I.L. 13, 6472. A centurion of the 22nd commanded the 1st Sequanian cohort in the Odenwald, ibid. 6509. A centurion of the 8th commanded a detachment of Britons at Welzheim, ibid. 6526.

^{page 45 note 3} On these ‘beneficiarii’ and their stations see Domaszewski, in W.D.Z. 1902, p. 158Google Scholar.

^{page 45 note 4} Our chief source of information is, of course, the series of military diplomata for the years 74, 82, 90, 116 and 134. They are printed in C.I.L. 3, 2, and Suppl.

^{page 45 note 5} Cohors II. Augusta Cyrenaica.

^{page 45 note 6} Cohors I. Flavia Damascenorum sagittariorum.

^{page 46 note 1} Cohortes I. and II. Mattiacorum, and the cohors Usiporum; see above, p. 30.

^{page 46 note 2} See above, p. 38.

^{page 46 note 3} C.I.L. 13, 6448.

^{page 46 note 4} Limes-Blatt, p.762; C.I.L. 13, 7441; or ‘Nida’ may have been the ancient name of Heddernheim. , Fabr. Baden, p. 64Google Scholar.

^{page 46 note 5} C.I.L. II, 3104. Under the head of the local militia should perhaps be included the ‘juventus’ at Sumelocenna, 6468, and Öhringen, 6549, and also the ‘hastiferi sive pastores’ at Kastel, 7317; comp. Hermes, 22, 557.

^{page 46 note 6} C.I.L. 13, 6592.

^{page 47 note 1} Three were moved to Raetia, five to Pannonia and Mœsia; one cavalry regiment, the ‘ala Picentiana,’ was transferred to Britain after 82 A.D., possibly in connection with Agricola's campaigns.

^{page 47 note 2} And thence to Pannonia before 133. Limes-Bl. p. 579.

^{page 47 note 3} Tiles of the cohort have been found at Rottweil. It was at Langenhaim in the time of Commodus. C.T.L. 13, 7436.

^{page 47 note 4} Ibid. 13, 7735.

^{page 47 note 5} Ibid. 13, 6407, 7342.

^{page 47 note 6} See above, p. 42.

^{page 47 note 7} C.I.L. 13, p. 499.