1 Dio, li. 23, expressly says this as to the year 725: τέως μὲν οὖν ταῦτ᾿ ἐποίουν (i.e. so long as the Bastarnae attacked only the Triballi—near Oescus in Lower Moesia, and the Dardani in Upper Moesia), οὐδὲν σφίσι πρᾶγμα πρὸς τοὺς Ῥωμαίους ἦν· ἐπεὶ δὲ τόν τε Αἷμον ὑπερέβησαν καὶ τὴν Θρᾴκην τὴν Δενθελητῶν ἔνσπονδον αὐτοῖς οὖσαν κατέδραμον κ. τ. λ. The allies in Moesia, of whom Dio, xxxviii. 10 speaks, are the coast towns.

2 When Dio says (li. 23): τὴν Σεγετικὴν καλουμένην προσεποίησατο καὶ ἐς τὴν Μυσίδα ἐνέβαλε, the town spoken of, doubtless, can only be Serdica, the modern Sofia, on the upper Oescus, the key to the Moesian country.

3 After the campaign of Crassus the conquered land was probably organised in such a way that the coast went to the Thracian kingdom, as Zippel has shown (Röm. Illyricum, p. 243), and the western portion was, just like Thrace, assigned in fief to the native princes, in place of one of whom must have come the praefectus civitatium Moesiae et Triballiae (C. I. L. v. 1838), who was still acting under Tiberius. The usual assumption that Moesia was at first combined with Illyricum, rests only on the circumstance that in the enumeration of the provinces apportioned in the year 727 between emperor and senate in Dio, liii. 12 it is not named, and so was contained in “Dalmatia.” But this enumeration does not extend at all to the vassal–states and the procuratorial provinces, and so far all is in due keeping with our assumption. On the other hand, weighty arguments tell against the usual hypothesis. Had Moesia been originally a part of the province of Illyricum, it would have retained this name; for on the division of a province the name was usually retained, and only a defining epithet added. But the appellation Illyricum, which Dio doubtless reproduces l.c., was always in this connection restricted to the upper (Dalmatia) and the lower (Pannonia). Moreover, if Moesia was a part of Illyricum, there was no room left for that Prefect of Moesia and Triballia, or in other words for his kingly predecessor. Lastly, it is far from probable that in 72727. a command of such extent and importance should have been entrusted to a single senatorial governor. On the other hand, everything admits of easy explanation, if small client–states arose in Moesia after the war of Crassus; these were as such from the outset under the emperor, and, as the senate did not take part in their successive annexation and conversion into a governorship, this might easily be unnoticed in the Annals. It was completed in or before the year 74311., seeing that the governor, L. Calpurnius Piso then waging war against the Thracians, to whom Dio (liv. 34) erroneously assigns the province of Pamphylia, can only have had as his province Pannonia or Moesia, and, as at that time Tiberius was acting as legate in Pannonia, there is left for him only Moesia. In 6 A.D. there certainly appears an imperial governor of Moesia.

4 The official title of Cottius was not king, like that of his father Donnus, but “president of the cantonal union” (praefectus civitatium), as he is named on the still standing arch of Susa erected by him in honour of Augustus in the year 745–69–8.. But the position was beyond doubt held for life, and, under reservation of the superior’s right to confirm it, also hereditary; so far therefore the union was certainly a principality, as it is usually so termed.

5 We know this road only in the shape which the emperor Claudius, the son of the constructor, gave to it; originally, of course, it cannot have been called via Claudia, but only via Augusta, and we can hardly regard as its terminus in Italy Altinum, in the neighbourhood of the modern Venice, since, under Augustus, all the imperial roads still led to Rome. That the road ran through the upper Adige valley is shown by the milestone found at Meran (C. I. L. v. 8003); that it led to the Danube, is attested; the connection of the making of this road with the founding of Augusta Vindelicum, though this was at first only a market–village (forum), is more than probable (C. I. L. iii. p. 711); in what way Augsburg and the Danube were reached from Meran we do not know. Subsequently the road was rectified, so as to leave the Adige at Bautzen, and to lead up the Eisach valley over the Brenner to Augsburg.

6 The locality “in which the Bessi honour the god Dionysos,” and which Crassus took from them and gave to the Odrysians (Dio, li. 25), is certainly the same Liberi patris lucus, in which Alexander sacrificed, and the father of Augustus, cum per secreta Thraciae exercitum duceret, asked the oracle respecting his son (Suetonius, Aug. 94), and which Herodotus already mentions (ii. III; compare Euripides, Hec. 1267) as an oracular shrine placed under the protection of the Bessi. Certainly it is to be sought northwards of Rhodope; it has not yet been discovered.

7 That the battle at Arbalo (Plin. H.N. xi. 17, 55) belongs to this year, is shown by Obsequens, 72, and so the narrative in Dio, liv. 33, applies to it.

8 That the fall of Drusus took place in the region of the Saale we may be allowed to infer from Strabo, vii. 1, 3, p. 291, although he only says that he perished on the march between Salas and Rhine, and the identification of the Salas with the Saale rests solely on the resemblance of name. From the scene of the mishap he was then transported as far as the summer camp (Seneca, Cons. ad Marciam 3: ipsis illum hostibus aegrum cum veneratione et pace mutua prosequentibus nec optare quod expediebat audentibus), and in that camp he died (Sueton., Claud. 1). This camp lay in the heart of the barbarian land (Valerius Max. v. 5, 3) and not very far from the battlefield of Varus (Tacitus, Ann. ii. 7, where the vetus ara Druso sita is certainly to be referred to the place where he died); we may be allowed to seek it in the region of the Weser. The dead body was then conveyed to the winter–camp (Dio, lv. 2) and there burnt; this spot was regarded, according to Roman usage, also as the place of burial, although the depositing of the ashes took place in Rome, and to this is to be referred the honorarius tumulus with the annual obsequies (Sueton. l.c.). Probably we have to seek for this place at Vetera. When a later author (Eutropius, vii. 13) speaks of the monumentum of Drusus at Mentz, this is doubtless not the tomb, but the elsewhere mentioned Tropaeum (Florus, ii. 30: Marcomanorum spoliis et insignibus quendam editum tumulum in tropaei modum excoluit).

9 What we learn from Dio, lv. 10, partly confirmed by Tacitus, Ann. iv. 44, cannot be apprehended otherwise. Noricum and Raetia must have been put under this governor as an exceptional measure, or the course of operations induced him to pass beyond the limit of his governorship. The assumption that he marched through Bohemia itself, which would involve still greater difficulties, is not required by the narrative.

10 To a connection in rear of the camp on the Rhine with the port Boulogne we might perhaps take the much disputed notice of Florus, ii. 30, to refer: Bonnam (or Bormam) et Gessoriacum pontibus iunxit classibusque firmavit, with which is to be compared the mention by the same author of forts on the Maas. Bonn may reasonably have been at that time the station of the Rhine–fleet; Boulogne was in later times still a fleet–station. Drusus might well have occasion to make the shortest and safest land–route between the two stations for the fleet available for transport, though the writer, probably bent on striking effect, awakens by his pointed mode of expression conceptions which cannot be in that form correct.

11 As to the administrative partition of Gaul there is, apart from the separation of the Narbonensis, an utter absence of accounts, because it rested only on imperial ordinances, and nothing in reference to it came into the records of the senate. But the first information of the existence of separate Upper and Lower German commands is furnished by the campaigns of Germanicus, and the battle of Varus can hardly be understood under that assumption; here, doubtless, the hiberna inferiora appear, viz. that of Vetera (Velleius, ii. 120), and the counterpart to it, the superiora, can only have been formed by that of Mentz; but this was not under a colleague of Varus, but under his nephew, who was thus subordinate to him in command. Probably the partition only took place, in consequence of the defeat, in the last years of Augustus.

12 The praesidium constructed by Drusus in monte Tauno (Tacitus, Ann. i. 56), and the φρούριον ἐν Χάττοις παρ’ αὐτῷ τῷ Ῥήνῳ associated with Aliso (Dio, liv. 33), are probably identical, and the special position of the canton of the Mattiaci is evidently connected with the construction of Mogontiacum.

13 That the “fort at the confluence of the Lupias and the Helison,” in Dio, liv. 33, is identical with the oftener mentioned Aliso, and this must be sought on the upper Lippe, is subject to no doubt; and that the Roman winter–camp at the sources of the Lippe (ad caput Lupiae, Velleius, ii. 205), the only one of the kind, so far as we know, on German ground, is to be sought just there, is at least very probable. That the two Roman roads running along the Lippe, and their fortified places of bivouac, led at least as far as the region of Lippstadt, the researches of Hölzermann in particular have shown. The upper Lippe has only one confluent of note, the Alme, and as the village of Elsen lies not far from where the Alme falls into the Lippe, some weight may be here assigned to the similarity of name. To the view, supported among others by Schmidt, which places Aliso at the confluence of the Glenne (and Liese) with the Lippe, the chief objection is that the camp ad caput Lupiae must then have been different from Aliso, and in general this point lies too far from the line of the Weser, while from Elsen the route leads directly through the Dören defile into the Werra valley. Schmidt, who does not adhere to the identification of Aliso and Elsen, remarks generally (Westfälische Zeitschrift für Gesch. und Alterthumskunde, xx. p. 259), that the heights of Weser (not far from Elsen), and generally the left margin of the valley of the Alme, are the centre of a semicircle formed by the mountains in front, and this highlying, dry region, allowing an exact look–out as far as the mountains, which covers the whole country of the Lippe and is itself covered in front by the Alme, is well adapted for the starting–point of a march towards the Weser.

14 This and not more is what Velleius says (ii. 110): in omnibus Pannoniis non disciplinae (= military training) tantummodo, sed linguae quoque notitia Romanae, plerisque etiam litterarum usus et familiaris animorum erat exercitatio. These are the same phenomena as are met with in the case of the Cheruscan princes, only in increased measure; and they are quite intelligible when we bear in mind the Pannonian and Breucian alae and cohortes raised by Augustus.

15 If we assume that of the twelve legions who were on the march against Maroboduus (Tacitus, Ann. ii. 46), as many as we find soon after in Germany, that is, five, went to form the army there, the Illyrian army of Tiberius numbered seven, and the number of ten (Velleius, ii. 113) may fairly be referred to the contingents from Moesia and Italy, that of fifteen to the contingents from Egypt or Syria, and to the further levies in Italy, whence the newly raised legions went no doubt to Germany, but those thereby relieved went to the army of Tiberius. Velleius (ii. 112) speaks inaccurately, at the very beginning of the war, of five legions brought up by A. Caecina and Plautius Silvanus ex transmarinis provinciis; firstly, the transmarine troops could not be at once on the spot, and secondly, the legions of Caecina were of course the Moesian. Comp. my commentary on the Mon. Ancyr. 2d ed. p. 71.

16 Velleius (ii. 118) says so; adsiduus militiae nostrae prioris comes, iure etiam civitatis Romanae eius equestres consequens gradus; which coincides with the ductor popularium of Tacitus, Ann. ii. 10. Such officers must have been of no infrequent occurrence at this time; thus, there fought in the third campaign of Drusus inter primores Chumstinctus et Avectius tribuni ex civitate Nerviorum (Liv. Ep. 141), and under Germanicus Chariovalda dux Batavorum (Tac. Ann. ii. 11).

17 The effigy of Varus is shown on a copper coin of the African town Achulla, struck under his proconsulate of Africa in the year 747–8, B.C. 7–6 (L. Müller, Num. de l’ancienne Afrique, ii. p. 44, comp. p. 52). The base which once supported the statue erected to him by the town of Pergamus has again been brought to light by the excavations there; the subscription runs: ὁ δῆμος [ἐτίμησεν] Πόπλιον Κοινκτίλιον Σέξτου υἱὸν Οὐάρ[ον] πάσης ἀρετῆ[ς ἕνεκα].

18 The report of Dio, the only one which hands down to us a somewhat connected view of this catastrophe, explains the course of it sufficiently, if we only take further into account—what Dio certainly does not bring into prominence—the general relation of the summer and winter camps, and thereby answer the question justly put by Ranke (Weltgeschichte, iii. 2, 275), how the whole army could have marched against a local insurrection. The narrative of Florus by no means rests on sources originally different, as that scholar assumes, but simply on the dramatic accumulation of motives for action, such as is characteristic of all historians of this type. The peaceful dispensing of justice by Varus and the storming of the camp are both known to the better tradition, and that in their causal connection. The ridiculous representation of the Germans breaking in at all the gates into the camp, while Varus is sitting on the judgment–seat and the herald is summoning the parties before him, is not tradition, but a picture manufactured from it. That this is in utter antagonism to the description by Tacitus of the three bivouacs, as well as to sound reason, is obvious.

19 The normal strength of the three alae and the six cohortes is not to be calculated exactly, inasmuch as among them there may have been double divisions (miliariae); but the army cannot have numbered much over 20,000 men. On the other hand, there appears no reason for assuming a material difference of the effective strength from the normal. The numerous detachments which are mentioned (Dio, lvi. 19) serve to account for the comparatively small number of the auxilia, which were always by preference employed for this duty.

20 As Germanicus, coming from the Ems, lays waste the territory between the Ems and Lippe, that is, the region of Münster, and not far from it lies the Teutoburgiensis saltus, where Varus’s army perished (Tacitus, Ann. i. 61), it is most natural to understand this description, which does not suit the flat Münster region, of the range bounding the Münster region on the north–east, the Osning; but it may also be deemed applicable to the Wiehen mountains somewhat farther to the north, parallel with the Osning, and stretching from Minden to the source of the Hunte. We do not know at what point on the Weser the summer camp stood; but in accordance with the position of Aliso near Paderborn, and with the connections subsisting between this and the Weser, it was probably somewhere near Minden. The direction of the march on the return may have been any other excepting only the nearest way to Aliso; and the catastrophe consequently occurred not on the military line of communication between Minden and Paderborn itself, but at a greater or less distance from it. Varus may have marched from Minden somewhat in the direction of Osnabrück, then after the attack have attempted from thence to reach Paderborn, and have met with his end on this march in one of those two ranges of hills. For centuries there have been found in the district of Venne at the source of the Hunte a surprisingly large number of Roman gold, silver, and copper coins, such as circulated in the time of Augustus, while later coins hardly occur there at all (comp. the proofs in Paul Höfer, der Feldzug des Germanicus im Jahre 16, Gotha, 1884, p. 82, f.) The coins thus found cannot belong to one store of coins on account of their scattered occurrence and of the difference of metals, nor to a centre of traffic on account of their proximity as regards time; they look quite like the leavings of a great extirpated army, and the accounts before us as to the battle of Varus may be reconciled with this locality. As to the year of the catastrophe there should never have been any dispute; the shifting of it to the year 10 is a mere mistake. The season of the year is in some measure determined by the fact that between the arrangement to celebrate the Illyrian victory and the arrival of the unfortunate news in Rome there lay only five days, and that arrangement probably had in view the victory of 3d Aug., though it did not immediately follow on the latter. Accordingly the defeat must have taken place somewhere in September or October, which also accords with the circumstance that the last march of Varus was evidently the march back from the summer to the winter camp.

21 Tacitus, Ann. i. 9, and Dio, lvi. 26, attest the continuance of the state of war; but nothing at all is reported from the nominal campaigns of the summers of 12, 13, and 14, and the expedition of the autumn of 14 appears as the first undertaken by Germanicus. It is true that Germanicus had been proclaimed as Imperator probably even in the lifetime of Augustus (Mon. Ancyr. p. 17); but there is nothing to hinder our referring this to the campaign of the year 11, in which Germanicus commanded with proconsular power alongside of Tiberius (Dio, lvi. 25). In the year 12 he was in Rome for the administration of the consulate, which he retained throughout the year, and which was still at that time treated in earnest; this explains why Tiberius, as has now been proved (Hermann Schulz, Quaest. Ovidianae, Greifswald, 1883, p. 15), still went to Germany in the year 12, and resigned his Rhenish command only at the beginning of the year 13, on the celebration of the Pannonian victory.

22 The hypothesis of Schmidt (Westfäl. Zeitschrift, xx. p. 301)—that the first battle was fought on the Idistavisian field somewhere near Bückeburg, and the second, on account of the morasses mentioned on the occasion, perhaps on the Steinhudersee, near the village of Bergkirchen, which lies to the south of this—will not be far removed from the truth, and may at least help us to realise the matter. In this, as in most of the accounts of battles by Tacitus, we must despair of reaching an assured result.

23 The statement of Tacitus (Ann. ii. 45), that this was properly a war of the republicans against the monarchists, is probably not free from a wish to transfer Hellenico–Roman views to the very different Germanic world. So far as the war had an ethico–political tendency, it would be called forth not by the nomen regis, as Tacitus says, but by the certum imperium visque regia of Velleius (ii. 108).

24 There triumphed over Spain—apart from the doubtless political triumph of Lepidus—in 718 36, 40.
34.
38, 34, 29.
38.
38, 28.
39, 26.
29. Cn. Domitius Calvinus (consul in 714), in 720 C. Norbanus Flaccus (consul in 716), between 720 and 725 L. Marcius Philippus (consul in 716) and Appius Claudius Pulcher (consul in 716), in 726 C. Calvisius Sabinus (consul in 715), and in 728 Sex. Appuleius (consul in 725). The historians mention only the victory achieved over the Cerretani (near Puycerda in the eastern Pyrenees) by Calvinus (Dio, xlviii. 42; comp. Velleius, ii. 78, and the coin of Sabinus with Osca, Eckhel, v. 203).

25 As Augusta Emerita in Lusitania only became a colony in 72925. (Dio, liii. 26), and this cannot well have been left out of account in the list of the provinces in which Augustus founded colonies (Mon. Ancyr. p. 119, comp. p. 222), the separation of Lusitania and Hispania Ulterior must not have taken place till after the Cantabrian war.

26 Callaecia was not merely occupied from the Ulterior province, but must still in the earlier time of Augustus have belonged to Lusitania, just as Asturias also must have been at first attached to this province. Otherwise the narrative in Dio, liv. 5, is not intelligible; T. Carisius, the builder of Emerita, is evidently the governor of Lusitania, C. Furnius the governor of the Tarraconensis. With this agrees the parallel representation in Florus, ii. 33, for the _Drigaecini_ of the MSS. are certainly the Βριγαικινοί, whom Ptolemy, ii. 6, 29, adduces among the Asturians. Therefore Agrippa, in his measurements, comprehends Lusitania with Asturia and Callaecia (Plin. H. N. iv. 22, 118), and Strabo (iii. 4, 20, p. 166) designates the Callaeci as formerly termed Lusitani. Variations in the demarcation of the Spanish provinces are mentioned by Strabo, iii. 4, 19, p. 166.

27 These were the Fourth Macedonian, the Sixth Victrix, and the Tenth Gemina. The first of these went, in consequence of the shifting of quarters of the troops occasioned by the Britannic expedition of Claudius, to the Rhine. The two others, although in the meanwhile employed elsewhere on several occasions, were still, at the beginning of the reign of Vespasian, stationed in their old garrison–quarters, and with them, instead of the Fourth, the First Adiutrix newly instituted by Galba (Tacitus, Hist. i. 44). All three were on occasion of the Batavian war sent to the Rhine, and only one returned from it. For in the year 88 there were still several legions stationed in Spain (Plin. Paneg. 14; comp. Hermes, iii. 118), of which one was certainly the Seventh Gemina already, before the year 79, doing garrison–duty in Spain (C. I. L. ii. 2477); the second must have been one of those three, and was probably the First Adiutrix, as this soon after the year 88 takes part in the Danubian wars of Domitian, and is under Trajan stationed in upper Germany, which suggests the conjecture that it was one of the several legions brought in 88 from Spain to upper Germany, and on this occasion came away from Spain. In Lusitania no legions were stationed.

28 The camp of the Cantabrian legion may have been at the place Pisoraca (Herrera on the Pisuerga, between Palencia and Santander), which alone is named on inscriptions of Tiberius and of Nero, and that as starting point of an imperial road (C. I. L. ii. 4883, 4884), just as the Asturian camp was at Leon. Augustobriga also (to the west of Saragossa) and Complutum (Alcalá de Henares to the north of Madrid) must have been centres of imperial roads, not on account of their urban importance, but as places of encampment for troops.

29 With this we may connect the fact that the same legion was, though only temporarily and with a detachment, on active service in Numidia.

30 The expression used by Josephus (contra Ap. ii. 4), that “the Iberians were named Romans,” can only be referred to the bestowal of Latin rights by Vespasian, and is an incorrect statement of one who was a stranger.

31 Probably the most recent monument of the native language, that admits of certainty as to its date, is a coin of Osicerda—which is modelled after the denarii with the elephant that were struck by Caesar during the Gallic war—with a Latin and Iberian legend (Zobel, Estudio historico de la moneda antigua española, ii. 11). Among the wholly or partially local inscriptions of Spain several more recent may be found; public sanction is not even probable in the case of any of them.

32 There was a time when the communities of peregrini had to solicit from the senate the right to make Latin the language of business; but for the imperial period this no longer held good. On the contrary, at this time probably the converse was of frequent occurrence. For example, the right of coining was allowed on the footing that the legend had to be Latin. In like manner public buildings erected by non–burgesses were described in Latin; thus an inscription of Ilipa in Andalusia (C. I. L. ii. 1087) runs: Urchail Atitta f(ilius) Chilasurgun portas fornic(es) aedificand(a) curavit de s(ua) p(ecunia). That the wearing of the toga was allowed even to non–Romans, and was a sign of a loyal disposition, is shown as well by Strabo’s expression as to the Tarraconensis togata, as by Agricola’s behaviour in Britain (Tacitus, Agric. 21).

33 These remarkable arrangements are clear, especially from the lists of Spanish places in Pliny, and have been well exhibited by Detlefsen (Philologus, xxxii., 606 f.). The terminology no doubt varies. As the designations civitas, populus, gens, belong to the independent community, they pertain de jure to these portions; thus, e.g. there is mention of the X civitates of the Autrigones, of the XXII populi of the Asturians, of the gens Zoelarum (C. I. L. ii. 2633), which is just one of these twenty–two tribes. The remarkable document which we possess concerning these Zoelae (C. I. L. ii. 2633) informs us that this gens was again divided into gentilitates, which latter are also themselves called gentes, as this same document and other testimonies (Eph. Ep. ii. p. 243) prove. Civis is also found in reference to one of the Cantabrian populi (Eph. Ep. ii. p. 243). But even for the larger canton, which indeed was once the political unit, there are no other designations than these, strictly speaking, retrospective and incorrect; gens in particular is employed for it even in the technical style (e.g., C. I. L. ii. 4233 Intercat[iensis] ex gente Vaccaeorum). That the commonwealth in Spain was based on those small districts, not on the cantons, is clear as well from the terminology itself as from the fact that Pliny in iii. 3, 18, places overagainst those 293 places the civitates contributae aliis; moreover it is shown by the official at census accipiendos civitatium XXIII Vasconum et Vardulorum (C. I. L. vi. 1463) compared with the censor civitatis Remorum foederatae (C. I. L. xi. 1855, comp. 2607).

34 As the Latin communal constitution is unsuited for a community not organised as a town, those Spanish communities, which still after Vespasian’s time lacked urban organisation, must either have been excluded from the bestowal of Latin rights or have had special modifications to meet their case. The latter may be regarded as having more probability. Inscriptions, even of the gentes, subsequent to Vespasian’s time, show a Latin form of name, as C. I. L. ii. 2633, and Eph. Ep. ii. 322; and if isolated ones from this period should be found with non–Roman names, it must always be a question whether this is not simply due to actual negligence. Presumptive proofs of non–Roman communal organisation, comparatively frequent in the scanty inscriptions that certainly date before Vespasian (C. I. L. ii. 172, 1953, 2633, 5048), have not been met with by me in inscriptions that are certainly subsequent to Vespasian.

35 The direction of the via Augusta is specified by Strabo (iii. 4, 9, p. 160); to it belong all the milestones which have that name, as well those from the region of Lerida (C. I. L. ii. 4920–4928) as those found between Tarragona and Valencia (ibid. 4949–4954), and lastly, the numerous ones ab Iano Augusto, qui est ad Baetem, or ab arcu, unde incipit Baetica, ad oceanum.

36 At Clunia there was found a dedication to the Mothers (C. I. L. ii. 2776)—the only Spanish example of this worship so widely diffused and so long continuing among the western Celts—at Uxama, one set up to the Lugoves (ib. 2818), a deity that recurs among the Celts of Aventicum.

37 The choliambics (i. 61) run thus:—

38 The domain of Iberian coins reaches decidedly beyond the Pyrenees, though the interpretation of individual coin–legends, which are among others referred to Perpignan and Narbonne, is not certain. As all these coinings took place under Roman authorisation, this suggests the question whether this portion of the subsequent Narbonensis was not at an earlier date—namely before the founding of Narbo (636 U.C.)118.—under the governor of Hither Spain. There are no Aquitanian coins with Iberian legends any more than from north–western Spain, probably because the Roman supremacy, under whose protection this coinage grew up, did not, so long as the latter lasted, i.e. perhaps up to the Numantine war, embrace those regions.

39 This is shown by the remarkable inscription of Avignon (Herzog. Gall. Narb. n. 403): T. Carisius T. f. pr[aetor] Volcar[um] dat—the oldest evidence for the Roman organisation of the commonwealth in these regions.

40 Noviodunum (Nyon on the lake of Geneva) alone perhaps in the three Gauls may be compared, as regards plan, with Lugudunum (iv. 254)iv. 242.; but, as this community emerges later as civitas Equestrium (Inscrip. Helvet. 115), it seems to have been inserted among the cantons, which was not the case with Lugudunum.

41 The persons earlier driven forth from Vienna by the Allobroges (οἱ ἐκ Οὐιέννης τῆς Ναρβωνησίας ὑπὸ τῶν Ἀλλοβρίγων ποτὲ ἐκπεσόντες), in Dio, xlvi. 50, cannot well have been other than Roman citizens, for the foundation of a burgess–colony for their benefit is intelligible only on this supposition. The “earlier” expulsion probably stood connected with the rising of the Allobroges under Catugnatus in 693 61. (iv. 223)iv. 213.. The explanation why the dispossessed were not brought back, but were settled elsewhere, is not forthcoming; but various reasons prompting such a course may be conceived, and the fact itself is not thereby called in question. The revenues accruing to the city (Tacitus, Hist. i. 65) may have been conferred upon it possibly at the expense of Vienna.

42 The ground belonged formerly to the Segusiavi (Plin. H. N. iv. 15, 107; Strabo, p. 186, 192), one of the small client–cantons of the Haedui (Caesar, B. G. vii. 75); but in the cantonal division it counts not as one of these, but stands for itself as μητρόπολις (Ptolem. ii. 8, 11, 12).

43 This was the 1200 soldiers with whom, as Agrippa the king of the Jews says in Josephus (Bell. Jud. ii. 16, 4), the Romans held in subjection the whole of Gaul.

44 Nothing is so significant of the position of Treves at this time as the ordinance of the emperor Gratianus of the year 376 (Cod. Theod. xiii. 3, 11), that there should be given to the professors of rhetoric and of the grammar of both languages in all the capitals of the then subsisting seventeen Gallic provinces, over and above their municipal salary, a like addition from the state chest: but for Treves this was to be on a higher scale.

45 In Caesar there appear doubtless, taken on the whole, the same cantons as are thereafter represented in the Augustan arrangement, but at the same time manifold traces of smaller client–unions (comp. iv. 237)iv. 226.; thus as “clients” of the Haedui are named the Segusiavi, the Ambivareti, the Aulerci Brannovices, and the Brannovii (B. G. vii. 75), as clients of the Treveri the Condrusi (B. G. iv. 6), as clients of the Helvetii the Tulingi and Latobriges. With the exception of the Segusiavi, all these are absent from the Lyons diet. Such minor cantons not wholly merged into the leading places may have subsisted in great number in Gaul at the time of the conquest. If, according to Josephus (Bell. Jud. ii. 16, 4), three hundred and five Gallic cantons and twelve hundred towns obeyed the Romans; these may be the figures that were reckoned up for Caesar’s successes in arms; if the small Iberian tribes in Aquitania and the client–cantons in the Celtic land were included in the reckoning, such numbers might well be the result.

46 This is indicated not only by the inscription in Boissieu, p. 609, where the words tot[i]us cens[us Galliarum] are brought into connection with the name of one of the altar–priests, but also by the honorary inscription erected by the three Gauls to an imperial official a censibus accipiendis (Henzen, 6944). He appears to have conducted the revision of the land–register for the whole country, just as formerly Drusus did, while the valuation itself took place by commissaries for the individual districts. A sacerdos Romae et Augusti of the Tarraconensis is praised ob curam tabulari censualis fideliter administratam (C. I. L. ii. 4248); thus doubtless the diets of all provinces were invested with the apportionment of the taxes. The imperial finance–administration of the three Gauls was at least, as a rule, so divided that the two western provinces (Aquitania and Lugudunensis) were placed under one procurator, Belgica and the two Germanies under another; yet there were probably not legally fixed powers for this purpose. A regular taking part in the levy may not be inferred from the discussion held by Hadrian—evidently as an extraordinary step—with representatives of all the Spanish districts (vita, 12).

47 For the arca Galliarum, the freedman of the three Gauls (Henzen, 6393), the adlector arcae Galliarum, inquisitor Galliarum, iudex arcae Galliarum, no other province, so far as I know, furnishes analogies; and of these institutions, had they been general, the inscriptions elsewhere would certainly have preserved traces. These arrangements appear to point to a self–administering and self–taxing body (the adlector, the meaning of which term is not clear, occurs as an official in collegia, C. I. L. vi. 355; Orelli, 2406); probably this chest defrayed the doubtless not inconsiderable expenditure for the temple–buildings and for the annual festival. The arca Galliarum was not a state–chest.