STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. AUGUSTUS' WAR IN SPAIN (26-25 B.C.) By David Magie Concerning the provinces of Spain one might have written in 26 b.c. what Tacitus wrote of the Germans more than a century later: triumphati magis quam victi sunt. Between 36 and 26 no less than six triumphs over the Spanish provinces had been granted to returning generals, 1 and yet, even before the last had been celebrated, the dauntless tribes of the northern part of the peninsula invaded once more the Romanized districts south of the Cantabrian Pyrenees. The war which ensued has received inadequate treatment at the hands of modern historians of Rome, none of whom has examined critically the sources for our knowledge of the campaigns, or studied them with the aid of inscriptions and the material afforded by ancient itineraries and modern maps. Schiller 2 repeated the meager narrative of the sources, merely adding a few modern place-names, without subjecting this narrative to criticism or applying to the vague statements of the ancients the evidence afforded by modern topo- graphical study. Still less satisfactory is the treatment which the war received from Gardthausen 3 and von Domaszewski. 4 Except for Gardthausen's enumeration of the legions composing the army and his attempt to correct an error of the sources in regard to the legates charged with the conduct of a particular campaign, these historians contented themselves with a paraphrase of the sources and omitted entirely any discussion of the topographical problems. And yet the Spanish war was the only one which Augustus, after his assumption of the imperium in 27, conducted in person, 5 and his subjugation of the peninsula stands foremost among his boasts of military prowess. 8 It seems justifiable, therefore, to attempt to retell the story of this 1 OIL, F, pp. 180 f. * Gesch. d. rOm. K.-zeit, I, 206 f. 3 Augustus u.s. Zeit, I, 675 f.; II, 367 f. 4 Gesch. d. r6m. Kaiser, I, 178 f. 5 Suet. Aug. 20. 6 Monumenlum Ancyranum, c. 26. [Classical Philology XV, October, 1920] 323 324 David Magie war, using the aid afforded by epigraphical and topographical sources and so correcting and supplementing the confused and meager narratives of the ancient historians. Save for some brief statements of Cassius Dio 1 and a few notices in Suetonius, 2 the literary sources for our knowledge of the war are solely Florus and Orosius. Their narratives, abridged from the Epitome Liviana, 3 a compendium of Livy's history made in the first century, contain material that in its original form was composed in the time of Augustus and so represents the version current in the period that immediately followed the completion of the war. Although the successive abridgments which the original has under- gone have mangled the text and confused the narrative, both authors relate the events in the same order and both have preserved the names of the places which were the centers of the campaigns. It seems possible, therefore, by means of a critical study of the texts and a comparison of the details which they afford, to distinguish the various campaigns of the war and to determine the locality of each. The narratives of Florus and Orosius read as follows: FLORUS ii. 33; (iv. 12. 46-59) OROSIUS vi. 21. 1-11 Hie duae validissimae gentes, Can- Anno ab urbe condita DCCXXVI tabri et Astures, immunes imperii imperatore Augusto Caesare sexies agitabant. Cantabrorum et prior et et bis M. Agrippa consulibus Caesar acrior et magis pertinax in rebellando parum in Hispania per ducentos animus f uit, qui non contenti liber- annos actum intelligens, si Cantabros tatem suam def endere proximis etiam atque Astures, duas f ortissimas His- imperitare temptabant, Vaccaeosque paniae gentes, suis uti legibus sineret, et Turmogos et Autrigonas crebris aperuit Iani portas atque in His- incursionibus fatigabant. panias ipse cum exercitu profectus est. Cantabri et Astures Gallaeciae provinciae portio sunt, qua extentum Pyrenaei iugum baud procul secundo Oceano sub septentrione deducitur. Hi non solum propriam libertatem " Dio liii. 22. 5—26. 1. 2 Suet. Aug. 20; 29. 3 See G. Reinhold, Das Geschichtswerk des Livius als Quelle spat. Historiker (Berlin, 1898), pp. 3f.; E. Kornemann, Die neue Livius-Epitome (Leipzig, 1904), =KKo, Beilrage z. alt. Gesch. Beiheft, II, 86 f.; Schanz, G.R.L., II 3 , 427 f. Augustus' War in Spain 325 Florus In hos igitur, quia vehementius agere nuntiabantur, non mandata expeditio, sed sumpta est. Ipse venit Segisamam, castra posuit, inde tripertito exercitu totam Canta- briam amplexus efferam gentem ritu ferarum quasi quadam cogebat indagine. Nee ab Oceano quies, cum infesta classe ipsa quoque terga hostium caederentur. Primum adversus Cantabros sub moenibus Bergidae (v.l. Belgicae) proeliatum. Hinc statim fuga in eminentissimum Vindium montem, quo maria prius Oceani quam arma Romana ascensura esse crediderant. Tertio Aracelium oppidum magna vi repugnat; captum tamen postremo fuit Medulli montis obsidio, quem perpetua XV milium fossa com- prehensum undique simul adeunte Romano postquam extrema barbari vident, certatim igne ferro inter epulas venenoque, quod ibi volgo ex arboribus taxeis exprimitur, prae- cepere mortem, seque pars maior a captivitate, quae morte gravior ad id tempus indomitis videbatur, vindicaverunt. Orosius tueri parati verum etiam finitimorum praeripere ausi, Vaccaeos et Turmo- gos et Autrigonas adsiduis erup- tionibus populabantur. Igitur Caesar apud Segisamam castra posuit, tribus agminibus totam paene amplexus Cantabriam. Diu fatigato frustra atque in periculum saepe deducto exercitu, tandem ab Aquitanico sinu per Oceanum incautis hostibus admoveri classem atque exponi copias iubet. Tunc demum Cantabri sub moeni- bus Atticae maximo congressi bello et victi in Vinnium montem natura tutissimum confugerunt, ubi obsi- dionis fame ad extremum paene consumpti sunt. Racilium deinde oppidum, magna vi ac diu repugnans, postremo cap- tum ac dirutum est. Praeterea ulteriores Gallaeciae partes, quae montibus silvisque consitae Oceano terminantur, Antistius et Firmius legati magnis gravibusque bellis per- domuerunt. Nam et Medullium montem Minio flumini imminentem, in quo se magna multitudo hominum tuebatur, per XV milia passuum fossa circumsaeptum obsidione cinxe- runt. Itaque ubi se gens homi- num trux natura et ferox neque toler- andae obsidioni sufficientem neque suscipiendo bello parem intellegit, ad voluntariam mortem servitutis timore concurrit. Nam se paene omnes certatim igne ferro ac veneno necaverunt. 326 David Magie Floeus Haec per Antistium Furniumque (codd. Firmumque) legates et Agrip- pam hibernans in Tarraconis mari- timis Caesar accepit. Mox ipse praesens hos deduxit montibus, hos obsidibus adstrinxit, hos sub corona iure belli venumdedit. Digna res lauro, digna curru senatui visa est; sed iam tantus erat Caesar ut triumpho augeri contemneret. Astures per id tempus ingenti agmine a montibus niveis descender- ant. Nee temere sumptus barbaris videbatur hie impetus; sed positis castris apud Asturam flumen tri- fariam diviso agmine tria simul Romanorum adgredi parant castra. Fuissetque anceps et cruentum et utinam mutua clade certamen cum tam fortibus, tarn subito, tarn cum consilio venientibus, nisi Brigaecini prodidissent, a quibus praemonitus Carisius cum exercitu advenit. Pro victoria fuit oppressisse consilia, sic quoque non incruento certamine. Reliquias fusi exercitus validis- sima civitas Lancia excepit, ubi cum locis adeo certatum est ut, cum in captam urbem faces poscerentur, aegre dux impetraverit veniam, ut victoriae Romanae stans potius esset quam incensa monumentum. Hie finis Augusto bellicorum cer- taminum fuit, idem rebellandi finis Hispaniae. Oeosius Astures vero positis castris apud Asturam flumen Romanos, nisi pro- diti praeventique essent, magnis consiliis viribusque oppressissent. Tres legates cum legionibus suis in tria castra divisos tribus aeque agminibus obruere repente moliti, suorum proditione detecti sunt. Hos postea Carisius bello exceptos non parva etiam Romanorum clade superavit. Pars eorum proelio elapsa Lan- cham confugit. Cumque milites cir- cumdatam urbem incendio adoriri pararent, dux Carisius et a suis ces- sationem impetravit incendii et a barbaris voluntatem deditionis exegit. Studiose enim nitebatur in- tegram atque incolumem civitatem victoriae suae testem relinquere. Cantabricae victoriae hunc hon- orem Caesar detulit, ut tunc quoque belli portas claustro cohiberi iuberet. Ita tunc secundo per Caesarem, quarto post urbem conditam clausus est Ianus. The war began, according to Orosius, in 28 b.c. This date is consistent with his further statement, Cantabrico bello per quinque Augustus' War in Spain 327 annos acto .... Caesar Romam rediit, 1 for Augustus, delayed by illness on his homeward journey, did not reach Rome until the spring of 24. 2 However, the assertion that the Emperor left Rome in 28 is obviously incorrect, for it was in 27 that the Senate conferred on him the new imperium and the name of Augustus. Gardthausen 3 has attempted to account for the error by the supposition that the Cantabri and Astures began their inroads in 28 and that the departure of Augustus was afterward brought into connection with this invasion. However, in view of the fact that when Augustus left Rome his destination was not Spain but Gaul and Britain, 4 Gardthausen's explanation is unsatisfactory. It seems more probable that either the Epitome Liviana or Orosius erred in the date, confusing Augustus VI et Agrippa II, consuls of 28, with Augustus VII et Agrippa III, consuls of 27. It was, then, in the course of the year 27 that Augustus left Rome. The exact date of his departure is uncertain, but it was after April 24, when Tiberius assumed the toga virilis, & and perhaps not until after September 25, when Messalla Corvinus celebrated his triumph over Aquitania. 6 According to current rumor, the purpose of his journey was the conquest of Britain, and the general expectation seems to have been that the new emperor would perform some great military exploit. 7 His first measure, however, was the organization of the Gallic provinces, including the recently conquered Aquitania, and orders were given to take a census. 8 When he formed the plan of conducting the Spanish war in person is unknown, but it can hardly have been prior to his departure from Rome. It seems most reason- able to suppose that the news of the outbreak in the much triumphed- over provinces reached him in Gaul, and that thereupon the astute l vi. 21. 21. 2 Dio liii. 28. 1; see also Fasti Feriar. Lat., CIL, I 2 , p. 58 =XIV, 2240. 3 II, 369, n. 17. * Dio liii. 22. 5. 5 Fasti Praeneslini, CIL, I 2 , p. 236. 6 CIL, I 2 , p. 180; see also Appian B.C. iv. 38. 7 An inference drawn from the poetry of the period; see G. Ferrero, Greatness and Decline of Rome, iv, 149, and T. Frank, Roman Imperialism (New York, 1914), pp. 348 f. 8 Dio loc. cil.; Liv. Per. cxxxiv ; see Marquardt, R8m. Staatsverwallung, I 2 , 204 f . 328 David Magie Emperor seized the opportunity of acquiring military glory by putting the final touch to the subjugation of the long rebellious and valuable peninsula. A successful campaign would satisfy the demand for a great national victory, and the conquest of Spain with its rich mines would be a more useful exploit than the invasion of a far-away island. His decision was acted on promptly, for on January 1, 26 B.C., he had reached Tarraco, where he assumed his eighth consulship. 1 His nephew Marcellus 2 and his stepson Tiberius 3 were with him, as he had himself been taken to this very province in 45 by Caesar the Dictator. The first move in the war was made in the spring of 26, for the Emperor's arrival in 27 was too late in the year to permit the under- taking of any campaign before winter set in. A general head- quarters was established by Augustus in person at Segisama, and from here three divisions of the army were sent out to overrun the whole of Cantabria. The name of the location of the headquarters is given by both our sources, but the identification of the site has been complicated by the fact that there were in ancient Spain two towns of similar name, Segisama, or Segisama Iulia, and Segisamo. 4 The latter has been identified by means of epigraphical evidence with the town of Sasamon in the province of Burgos, 6 and it was there- fore argued by Hiibner 6 that the site of the former must be sought elsewhere. However, the location of Segisama Iulia, enumerated among the towns of the Vaccaei by Polybius and Ptolemaeus and listed as a town of the Turmogidi by Pliny, approximates closely to ' Fasti Feriar. Lai., CIL, I ! , p. 58=XIV, 2240; Suet. Aug. 26. * Dio liii. 20. 1; Crinagoras in Anth. Pal. vi. 161; see Gardthausen, I, 682. » Dio loc. cit.; Suet. Tib. 9. 4 The evidence is as follows: Seyecrdjua, a town of the Vaccaei (Polybius in Strabo iii. p. 162). Segisamonenses and Segisamaiulienses, communities of the Turmogidi (i.e., Turmogi) (Plin. N.H. iii. §26). 2eyiVa/aa 'lovXia, a town of the Vaccaei (Ptol. ii. 6. 49). Seyiaa/Jiov (klyiaiftov, codd.), a town of the Murbogi (evidently Turmogi) (Ptol. ii. 6. 51). Segisama, a station on the main road from Tarraco by way of Caesaraugusta to Legio VII Gemina (Leon) and 92 miles from the latter (ltin. Ant. 394. 5; 449. 5; 454. 2). On the other hand, the site of Segisama Brasaca, mentioned in an inscription from Tarraco (CIL, II, 4157) cannot be identified. * CIL, II, 2915, and Suppl., pp. 932 f. 6 CIL, II, 2915. Augustus' War in Spain 329 the situation of Segisamo-Sasamon, for this was in the district of the Turmogidi a short distance east of the country of the Vaccaei. It seems clear, therefore, that the two towns, Segisamo and Segisama Iulia, were close to each other and that the general headquarters were not far from the modern town of Sasamon. The district is directly south of the Cantabrian Mountains and must have been near the region which had been invaded by the Spaniards. It was therefore an excellent point from which to proceed against the enemy. Situated as it is, moreover, in the broad plain watered by the river Brulles, a tributary of the Pisuerga, the locality was a favorable one for purposes of mobilization and procuring of supplies. Communi- cations with the Romanized portion of the province could be main- tained not only by the line of what was afterward the main route across the mountains into the valley of the Ebro, but also southward along the Pisuerga to the great plain of the upper Douro. From the headquarters at Segisama three divisions of the army were sent out, and the country of the enemy was invaded at three different points. The first result of the operations of one division was a battle just outside a town, the name of which, according to the commonly accepted text of Florus, was Bergida, according to the text of Orosius, Attica. Our modern historians have accepted, apparently without question, the traditional reading of the editions of Florus and narrate that the Romans fought with the Cantabrians at Bergida (or Bergidum) and that after the battle the enemy with- drew to Mons Vindius. This maneuver, however, is a geographical impossibility. There were two towns in ancient Spain which bore the name Bergidum, but neither is in the neighborhood of Mons Vindius, the western portion of the range of the Cantabrian Moun- tains. 1 It was supposed by Hubner 2 that the Bergidum which is listed by Ptolemaeus 3 among the towns of the Ilergetes is to be identi- fied with the Bergida where Florus locates this battle. Any connec- tion, however, between a town of the Ilergetes, in the modern province of Lerida in the northeastern corner of Spain, and a campaign in the 1 A. Forbiger, Handbuch d. alien Geographie, III 2 , 7. This is the "Opos OilvSiov of Ptol. ii. 6. 20. 2 Pauly-Wissowa, Realencycl., Ill, 292, s.v. " Bergistani." * ii. 6. 67. It is probably the modern Berga in the northern part of the province of Barcelona. 330 David Magie region of Mons Vindius is of course out of the question. The other alternative is Bergidum Flavium, listed by Ptolemaeus 1 as a town of the Astures, and situated, according to the Itinerarium Antonini, 2 on the road leading from Lucus Augusti (Lugo) to Asturica Augusta (Astorga). It has been identified with Castro de la Ventosa in the district of el Vierzo in the northwestern part of the province of Le6n. 3 This town is, indeed, much nearer to the seat of the war than is the Bergidum of the Ilergetes, but nevertheless it cannot be identified with the scene of the battle recorded by Florus and Orosius. Its site is quite incompatible with the statement of the sources that the enemy fell back on Mons Vindius, fci- even the most westerly portion of this range is a long distance to the northeast. Moreover, Bergidum Flavium is not in the district of the Cantabri but in the country of the Astures, and the sources distinguish clearly between this campaign against the Cantabri and the expedition conducted against the Astures by the third division of the Roman army. 4 The reading Bergidae, however, which has been adopted by most of the modern editors, is not found in all the manuscripts of Florus. The group headed by the Codices Nazarianus (Palatinus 894) and Vossianus 14 contains the reading Belgicae. b No Spanish town of this name is known, but this reading strongly suggests ObeWwa, listed by Ptolemaeus 6 as a town of the Cantabri. In fact, the reading Vellicae was adopted by Stadius in his edition and was defended by Gruter and Graevius. The name of this town appears on a tomb- stone found on Monte Cilda near Aquilar de Camp6o in the province of Palencia, in the southern foothills of the Cantabrian Mountains. 7 The mere mention of the name in this inscription does not, of course, prove that this locality is Vellica, 8 but the presumption in favor of i ii. 6. 28. 2 425. 4; 429. 2; 431. 1. *CIL, II, Suppl., p. 996; Hubner, Pauly-Wissowa, Realencycl.. Ill, 291; K. Miller, Itineraria Romana (Stuttgart, 1916), 167. • See below, pp. 333 f. 5 Rossbach has pointed out, to be sure, that the proper names in this group of manuscripts are, as a rule, in a much more corrupt condition than in the Codex Bamber- gensis; see his edition (Leipzig 1896), Intro., pp. xxxiv f. • ii. 6. 50. 7 CIL, II, 6297: Valerio Quadralo Boddi filio Vellic. an. XL, etc. • The identification is assumed by Kiepert, Formae Orbis Antiqui. xxvii; Hispania. On the other hand, Hubner (CIL, II, 6297) suggests that Vellica is to be identified with Oi«\«a of the Caristi (Ptol. ii. 6. 64) and with the Velienses, a community of the Augustus' Wak in Spain 331 this identification is strengthened by the fact that the place is well within the district of the Cantabri, and furthermore its actual posi- tion with regard to the town of ObaSivia (located near Cangas de Onis in the province of Oviedo) 1 corresponds exactly with that of Vellica as given by Ptolemaeus. If Vellicae is to be read in the text of Florus, the text of Orosius is to be emended similarly, for it is evident that both narratives refer to the same place, and, indeed, Orosius' reading Atticae is an easily made error for Vellicae. This correction in the text removes all the difficulties raised by the reading Bergidae and by the consequent attempt to reconcile the topography of this campaign with the situa- tion of Bergidum Flavium; for Vellica is well within the district of the Cantabri and among the hills southeast of the range of Mons Vindius. Furthermore, situated, as it is, not more than 50 km. northwest of Sasamon, it is just the locality where a force setting out from Segisama would come into conflict with the Cantabrian mountaineers. The first division, then, of the Roman army mobilized at Segisama, marched almost due north along the upper Pisuerga (along the line of the modern railway from Palencia to Santander) until it reached the foothills of the Cantabrian range at Vellica. Here a battle was fought under the walls of the town, and as a result the Spaniards retreated, withdrawing from Vellica in a northwesterly direction to Mons Vindius. 2 This retreat opened to the Romans the pass leading northward from Vellica and so enabled them to join forces with the troops which had landed on the north coast, and, according to Florus, had begun to attack the Cantabrians in the rear. The campaign would seem, according to the testimony of Orosius, to have been Carietes (Plin. N.H. iii. § 26) , which is evidently the same place as Beleia {Itin. Ant., 454. 8), now Irufia near Vitoria in the province of Alava (Mueller, adPtol., he tit.) and in the general region of the second expedition. But this town is so far to the east that a retreat from it to Mons Vindius would be altogether impossible. Similarly, the BiXXoi (Polyb. xxxv. 2. 3 f. and Appian, Hisp. 44 f.), also suggested by Hiibner as the site of Vellica, are altogether too far to the south to have any connection with the Cantabri. 1 Ptol. ii. 6. 50, and Mueller, ad loc. 2 According to the measurements of Ptolemaeus (ii. 6. 20), the eastern end of "Opos Oiiviiov is only a short distance from Vellica. 332 David Magie brought to an end, not by any general engagement, but by what in a terrain of such a kind as Mons Vindius was a much wiser policy, isolating the enemy and starving them into surrender. The account of the next campaign related by Floras is introduced by the word tertio. The presence of this word might lead to the supposition that the description of the operations of the second divi- sion has been omitted in the unskilful abbreviation of the source. It is true that the narrative, as found in both Florus and Orosius, has been badly mutilated by careless abbreviation; in the former the process has been carried to such a length that the sentence which contains the account of this campaign is, as it stands, altogether unintelligible. Nevertheless, since the account given by Orosius agrees in the order of events with that of Florus, and since his nar- rative, far from showing any indication of the omission of the opera- tions of the second division, connects the expedition against Aracelium with the campaign at Mons Vindius by the word deinde, it seems more reasonable to suppose that the campaign here described was that conducted by the second of the three divisions operating from Segisama. The tertio of Florus may be regarded as merely a bungling attempt to connect with one another the events so far recorded: (1) the battle at Vellica, (2) the retreat to Mons Vindius, (3) the capture of Aracelium. The method employed by Florus in abbreviating his source has led to the result that this portion of his narrative is, from the geographical standpoint, a veritable absurdity. According to his account, the town of Aracelium resisted the Romans stoutly, but it was finally captured by means of the blockade of Mons Medullus. That this assertion is, to say the least, inaccurate, becomes evident as soon as it is realized that Aracelium is in northeastern Spain, whereas Mons Medullus is in Callaecia in the northwest. Never- theless, this statement of Florus has been accepted without question by modern historians. A comparison with the narrative of Orosius makes it evident that Florus has combined in one sentence the achievements of the second division, which operated in northeastern Spain, and those of the third, which operated in Callaecia. Although the version of Orosius has also been mangled by abbreviation and although it contains the corrupt form Racilium, it nevertheless Augustus' War in Spain 333 presents a more correct narrative than that of Floras, for it carefully distinguishes the capture of Aracelium from the campaign in Cal- laecia and the blockade of Mons Medullus. Furthermore, it affords the additional item of information that Aracelium, after its capture, was completely destroyed. The situation of Aracelium can be determined from Pliny, 1 who lists the Aracelitani among the stipendiarii of the conventus of Caesaraugusta. Moreover, under the name Araceli it is definitely located in the Itinerarium Antonini 2 as on the main road leading from Asturica Augusta (Astorga) through Segisamo to Burdigala and 24 miles west of Pompaelo. Its ancient name is still preserved in the modern form Huarte Araquil. 3 The location of Araceli once determined, it is easy to indicate the line of march followed by the second division in this campaign. From Segisama the route lay along the upper Arlanzon and its tributary, the Vega, past Virovesca (Bribiesca) along the line of the main road to Deobriga (Miranda de Ebro), in the valley of the Ebro, and thence up the Zadorra along the line of the road through Tul- lonium (Alegria) to Aracelium, the whole distance, according to the figures of the Itinerarium Antonini* being 134 miles. Unfortunately, the meager details preserved in the sources do not afford any further information about this expedition against the eastern Cantabrians. The capture and destruction of Aracelium, it may be inferred, was the chief event of the campaign, and it may well have put an end to further resistance in this region. After the brief mention of the campaign around Aracelium the sources proceed to give a highly colored narrative of the expedition to Callaecia and the elaborate blockade of Mons Medullus. This campaign was evidently the task assigned to the third of the three divisions which set out from the general headquarters. The seat of its main operations was the district of Callaecia, which, strictly speaking, lay in the country of the Astures and not in that of the Cantabri, but the historian, in summing up the three campaigns, i N.H. iii. § 24. 2 455. 3. •A. Forbiger, Handbuch, IIP, 60; K. Miller, Itineraria Rom., 170. « 454 f. 334 David Magie may be permitted to include this one in the impressive phrase, totam Cantahriam amplexus. The march of the third division was longer than that of the others. As far as Asturica Augusta (Astorga) its route lay across the great plateau of Le6n and Old Castile, and here the enemy could have found no point of vantage from which to fall upon the Roman army. The attacks doubtless began soon after the Romans entered the mountainous country west of the plateau, for it cannot be sup- posed that the righting around Mons Medullus was the first engage- ment of the campaign. Even the meager account of Orosius includes the statement that the region was conquered magnis gravibusque bellis, and the indefinite allusions of Dio 1 to sudden attacks and ensuing hard fighting doubtless apply to this campaign as well as to the others. Finally the Spaniards made a stand on Mons Medullus, and here the Roman army blockaded them with elaborate siege- works, including a fossa of fifteen miles' circuit. The siege seems to have been long and desperate, but finally those of the enemy who did not commit suicide surrendered. The location of Mons Medullus, described in Orosius as Minio flumini imminens, has been sought in various places along the Mifio and its main tributary the Sil, all the way from the Portuguese province of Entre Duero y Mifio on the lower Mifio, to the district of el Vierzo on the upper Sil. The most commonly accepted sites are the range of the Sierra de San Mamed in the province of Orenze and the neighborhood of the village of Las M6dulas in southeastern Le6n. 2 The most northerly peak of the Sierra de San Mamed, La Picafta, is, indeed, not far from the Sil, and its situation cor- responds in general with the flumini imminens of Orosius. On the other hand, north of the village of Las M6dulas there is a ridge run- ning approximately east and west between the Sil and the Cabrera and terminating abruptly where the Sil turns sharply southward to join the Cabrera. The distance between the parallel courses of the two streams is not over 15 km., and the ridge, rising conspicuously > liii. 25. 5 f. 2 A. Forbiger, Handbuch, III 2 , 9; Las Medulas is on the north bank of the Cabrera just above its junction with the Sil. Augustus' War in Spain 335 between them, might well be described as flumini imminens. 1 In the establishment of this identification the name of Las M6dulas would have greater weight, were it not for the fact that it is an obvious one for a village in a country so rich in minerals. However, it is not impossible that the modern term is more than a mere coinci- dence and that it may preserve the ancient name of the locality. It must, indeed, be admitted that the identification of this range with Mons Medullus is open to the objection that it could not be sur- rounded by a fossa of 15 miles in length. The argument applies, however, with equal force to all the possible sites in this general neighborhood, for all the mountains are ridges and not isolated peaks. It would seem, therefore, that the statement that the moun- tain was surrounded by a trench cannot be accepted as literally correct. According to the narrative of Floras, Mons Medullus was finally captured in the course of the winter, and Augustus upon receipt of the news set out from Tarraco to receive the surrender in person. The Emperor seems to have spent the winter of 26-25 b.c. in the capital of the province, whither he had withdrawn after his bad health caused his departure from the seat of war, 2 and it was here that he entered upon his ninth consulship. 3 His official acceptance of the surrender took place certainly after his assumption of the consulship on January 1, 25 b.c, and probably in the early spring, after the rigors of the winter were over. The names of the Roman commanders who were charged with the conduct of this campaign have been subjected to a critical examina- tion by Gardthausen. 4 He pointed out that the statement of Floras that Agrippa was among them is an evident error, for Agrippa was in Rome during the winter of 26-25. 6 He then attempted to show 1 The flumen is, of course, the modern Sil, not the Mifio, but the use of the name Minius in designating it is not unnatural in view of the fact that the Sil forms nearly a straight line with the course of the lower Mifio. Kiepert identified Mons Medullus with the group of the Pefia Trevinca, which is south of Las Medulas and contains the headwaters of the Cabrera. This mountain, however, is so far from the main course of the Cabrera that the phrase flumini imminens cannot be applied to it. 2 Fasti Feriar. hat., CIL, I 2 , p.58=XIV, 2240; Suet. Aug. 81; Dio liii. 25. 7. 3 CIL, loc. tit.; Suet. Aug. 26. * ii. 374 f . * Dio liii. 23 and 27. 336 David Magie that Furnius, 1 the second of the legates named in the ordinary versions of our sources, was not in Spain at this time, basing his argument on the statement of Dio 2 that in the revolt of 22 b.c. the Spaniards despised him as airupov and on v&jsotI ckjhkto. Here, however, Gardthausen's reasoning is not so well founded, for Dio records distinctly that it was the Cantabri who despised their new governor and revolted, whereas the expedition under discussion was directed against the Astures. He would still have been a stranger to the Cantabri in 22 b.c. There is, however, another objection to Gardthausen's argument, namely, that it deals with what is merely an emendation in the text. In the manuscripts of Florus the name of this legate is given as Firmum, and Firmius is the reading of all the codices of Orosius. The emendation Furnium, suggested by Stadius and adopted by Frensheim and all later editors of Florus, has been accepted merely for the reason that no Firmius or Firmus is known to us, while Furnius was in Spain in 22. It seems much more reason- able to retain the reading of the manuscripts and admit frankly that the legate is otherwise unknown than to accept a conjecture and then attempt to show that the name is an erroneous addition to the text. In spite of the ceremonial surrender of Mons Medullus the resistance of the Spaniards continued during the year 25. The cam- paign of this year, as far as any information can be obtained from our sources, was fought solely against the Astures; the Cantabri seem to have had no part in it. A general attack was planned by the natives, and their forces were assembled on the river Astura. This river is mentioned only here and by Isidorus, 3 and there is therefore no certain clue to any identification. However, the state- ment of Florus that the Brigaecini betrayed their fellow-countrymen points to an identification of the Astura with one of the rivers of the great plain of Le6n, and the name of the town Asturica Augusta (Astorga) suggests that the river is in the same neighborhood. Accordingly, all the available evidence points to the identification of the Astura with the Orbigo, on one of the tributaries of which Asturica is situated, and which, together with its lower portion, the Esla, drains the plain of Le6n. 4 1 Identified with C. Furnius, consul 17 B.C. 3 liv. 5. 1. 8 Orig. ix. 2. 112; the name is given as the origin of the name of the Astures. 4 Pauly-Wissowa, Realencycl., II, 1862. Augustus' War in Spain 337 The exact location of the campaign can be inferred only from the statement that the Brigaecini betrayed the plan of the Spaniards. Their town Brigaecium (Brigecum) 1 was situated, according to the Itinerarium Antonini, 2 40 miles from Asturica on the roads leading to Ocelum Duri and to Rauda, i.e., in a direction generally east of south. It is usually identified with Benevente on the Orbigo or with Villabrazaro, a short distance further up the river. 3 The invasion of the plain of Leon by the Astures meant that the mountaineers, undaunted by the campaign on the Sil-Mifio and the capture of Mons Medullus, had taken the offensive and had come forth to challenge the Romans. 4 If the statement is correct that they planned to attack the three Roman camps, it would seem that the headquarters of the Romans had been transferred from Segisama to the plain of Leon, perhaps for the reason that the Cantabri were regarded as conquered and the Astures were now considered the only formidable enemies. Each Roman camp was under the com- mand of a legate. One of these was evidently Carisius, the others were perhaps Antistius Vetus 5 and Furnius (or Firmus), who may be supposed to have returned to general headquarters after the com- pletion of the campaign on the Sil. The Spaniards had arranged their forces in three divisions which were to attack the three Roman camps simultaneously, thus preventing any commander from going to the aid of his colleagues. Before this plan could be carried out, however, it was betrayed by the Brigaecini to Carisius. He was thus enabled to anticipate the threefold attack by taking the offensive himself. He seems to have surprised the enemy and to have inflicted on them a crushing defeat. Although the language of the sources suggests that it was with his own division solely that Carisius won this victory, the statements are so vague that it is quite possible that the whole Roman army took part in the battle. The Romans suffered serious losses, but the Astures were driven from the field 1 Bptyatiiov (Ptol. ii. 0. 29), Brigecum (Itin. Ant.). Brigiaecinus (adj.), (OIL, II, 0094). 2 439. S; 440. 1. 3 Mueller, ad Ptol. loc. cit.; Pauly-Wissowa, Rcalencycl., Ill, 843; K. Miller, Itin. Rom., 153. 4 Perhaps the vague language of Dio (liii. 25. 8) is to be referred to this offensive. 6 Suggested by Dio liii. 25. 7. 338 David Magie and retreated in confusion to Lancia (Cerro de Lancia), falling back along the upper Esla. They took possession of the town and attempted to make a stand against the Romans. Unable, however, to offer any effective resistance, they surrendered when Carisius threatened to set fire to Lancia. The capitulation of this town, according to Dio, 1 was followed by the capture of other strongholds and the conquest of the district was thus completed. The hero of the campaign of 25, according to the historians, was Carisius. Their testimony is borne out by the evidence of coins of Augustus bearing the legends P. Carisius, Leg. pro Pr. and P. Carisius, Leg. Augusti, 2 for the identification of this P. Carisius with the conqueror of the Astures seems assured by the emblems which appear on the coins, the characteristic Spanish sword and helmet with the angon, the trophy of Spanish arms with kneeling captive, and a section of the fortifications of a town with the name Emerita. The last is of particular importance in the identification, since, according to Dio, 8 the colony of Augusta Emerita was founded immediately after the close of the campaign. 4 After this successful campaign the Temple of Janus was cere- moniously closed. 5 It was officially announced that the war had been completed and the Spaniards conquered. The conquest, how- ever, was a superficial one, for in 24 6 and again in 22 7 the Asturians 1 liii. 25. 8. 2 Babelon, Monnaies de la R6p. rom., I, 318 f., Nos. 14-30. 3 liii. 26. 1. 4 According to Dio, the name of the legate was Titus Carisius. Babelon {op. cit., p. 317) explained the discrepancy by supposing that Dio confused P. Carisius with his father T. Carisius, a triumvir monetalis under Caesar, and he has been followed by Hilbner (CIL, II, p. 52), Schiller (Gesch. d. rom. K.-zeit, I, 207, n. 4), and Groag (Pauly-Wissowa, III, 1592). On the other hand, Gardthausen (II, 375) made an unsuccessful attempt to explain the difference in names by the hypothesis that P. Carisius was the son of the victor over the Astures, and that at some later date, when he too had become a legatus Augusti, he issued coins commemorating the exploits of his father. 6 Dio (liii. 26. 5) agrees with our sources; see also Mommsen, Res gestae Din. Aug.*, 50. There seems to be little reason for supposing with Schiller (I, 206, n. 8) that the temple was closed before the campaign on the Astura. Certainly the rhetori- cal sentence of Florus narrating the "capture" of Mons Medullus cannot be taken as evidence for the belief that the temple was closed after this exploit. Orosius and Dio record the event in its proper sequence after the capture of Lancia and the other strongholds. 5 Dio liii. 29. 1 ; Cassiod, Chron. an. 730. ' Dio liv. 5. Augustus' War in Spain 339 and Cantabrians renewed their attacks on the Romans. They were finally repressed only by hard fighting and a policy of terrorization, and it was not until 19 b.c. that these mountain peoples were finally subjugated by the resolution and persistence of Agrippa. 1 The campaign of 26-25 b.c, however, enabled Augustus, with comparatively little expenditure of men and money, to gain the glory of conquering a rebellious province and to divert the thoughts of the Romans from the long-talked-of conquest of Britain. More- over, his victory gave him an opportunity to close the Temple of Janus, and thereby to declare formally that his policy was one of peace and that wars had ceased throughout the Roman Empire. 2 Princeton University 1 Dio liv. 11. 2 Mon. Anc, c. 26: Gallias el Hispanias provincials .... pacmi].