The value of this passage turns upon the meaning of is saltus. It might refer solely to the pass of Hercules, but as the writer must have known the traditions of the Gallic hordes who crossed the Alps, it is more probable that he is specially referring to the Graian chain, as that over which both Hercules and Hannibal had passed.

4. Some lines of Varro, the learned writer on antiquities at the close of the Republic, are quoted for us by Servius in his commentary on Vergil x. 13: quas (Alpes) quinque viis Varro dicit transiri posse: una quæ est juxta mare per Ligures: altera qua Hannibal transiit: tertia qua Pompeius ad Hispaniense bellum profectus est: quarta qua Hasdrubal de Gallia in Italiam venit: quinta quæ quondam a Græcis possessa est, quæ exinde Alpes Graiæ appellantur.

5. Strabo, who wrote under Tiberius, cites Polybius as mentioning four passages across the Alps: τέτταρας ὑπερβάσεις ὀνομάζει μόνον· διὰ Λιγύων μὲν, τὴν ἔγγιστα τῷ Τυρρηνικῷ πελάγει· εἶτα τὴν διὰ Ταυρίνων, ἣν ̓Αννίβας διῆλθεν· εἶτα τὴν διὰ Σαλασσῶν· τετάρτην δὲ διὰ Ραιτῶν. IV. 6. 12.

Here

it should be noted that the important words v A. d. do not appear in a MS. of great value (Ep. Vat. 482), that if genuine, they may easily have been transplaced, or that they may be regarded as a comment of Strabo, rather than an extract from Polybius. Strabo himself shows elsewhere little interest in the route of Hannibal, and does not mention it where it would naturally occur.

Now if we turn to the third book of Polybius we shall find that the whole journey from Carthago Nova to the Italian plains is definitely measured. We are told that there were 2600 stadia to the river Iberus, and thence to Emporeion 1600, and 1600 more to the passage of the Rhone. From the Rhone to the beginning of the Alps (πρὸς τὴν ἀναβολὴν τῶν ̓Αλπέων) there were 1400, while the remainder of the way (oral al Twv A. Tepßolai) was 1200 stadia.

As far as the Pyrenees there is no doubt about the route, but much depends upon the place at which the Rhone was crossed, as that becomes the starting-point for future measurements.

The description of Polybius clearly indicates a passage near the town of Orange, about the village of Roquemaure, as that is halfway between the river's mouth and its junction with the Isère, while it should be according to the historian four days' march to either point. The actual distance of 75 miles to the Isère is in close harmony with the 600 stadia implied in his latter statements. There is also a long stretch of broad stream unbroken by any islands to suit with the words κarà Tηv

áπλν púow. It is moreover above the junction with the Durance, to cross over which would have been a needless labour for the army.

The other place suggested near Beaucaire and Tarascon is quite inconsistent with these data. In Livy there are no definite statements on the subject to point to any special place. After the passage of the Rhone however he says that Hannibal pushed on inland to avoid all contact with the Roman army, but his route was probably decided on beforehand, and he was guided by the Gauls, who had invited him to Italy, and who would naturally lead him through the passes which would bring him with most ease into their cantons.

Onward to the Isère his route is certain, after that all is matter of debate.

Our two authorities give a like account of the island enclosed between the Rhone and the Isère-the insula Allobrogum of later days and of the contests between the native powers, in which Hannibal took part. In Polybius we find besides the following data: (1) Hannibal having in ten days marched 800 stadia along the river, began the ascent of the Alps.' (2) We hear that the chieftain with whom he sided in the quarrel joined him in his march, and that the barbarians were kept in check in the plain country alike from fear of the cavalry, and of the native aid.

The words along the river,' rapà тòv Toтaμóv, are not in themselves definite. They have been taken to refer to the Isère which was spoken of not long before, and most critics accordingly trace the route of Hannibal along one or other of its banks. But there can be little doubt that the 'river' of the whole narrative is the Rhone, and in chapter 39 Polybius expressly says that they kept near it to the entrance of the mountain pass. Of course it is not to be supposed that it was tracked in all its windings, in the great bend for example which it makes at Lyons, but only that the general movement was in the direction of its stream. The country through which the march would lie was such that the cavalry could be used to good effect, while the left bank of the Isère would not at all meet this condition, and the right one would only partially fulfil it. In both cases the mountain country comes in sight too soon, and the ascent (dvaẞoλý) must have begun long before they had traversed 800 stadia of road, or made their ten days' leisurely advance, in constant fear of an attack,

Both these conditions are complied with by the route, which following the Rhone up to Vienne, leaves it for a while in its great bend, and meets it once more at St Genix, and thence to the pass of Mont du Chat where the ascent may be taken to begin. That point once reached the way would naturally lead along the upper course of the Isère, through the Tarantaise,

and over the Little St Bernard to the valley of Aosta. In favour of this route the following reasons may be urged.

1. The local features of the pass agree at least as well as any other with the general description of Polybius, and the detailed accounts of the measurements of space and time, with the 15 days of march, that is, and the 1200 stadia of way. The valley was a fertile one, and the native town which they attacked and pillaged may well have enriched them with its plunder. The XEUKÓTEтpov of the narrative may be probably identified with the Roche Blanche' on the Reclus. The pass is steeper on the Italian side, and the dangers therefore of the descent would have been naturally greater, and in the ravine below 'la Tuile' there is a place where the old snow might long remain unmelted, and the road for some way is much exposed to avalanches. Here therefore Hannibal might find the track completely swept away, and be obliged to halt until a new path was cut upon the mountain side. Though the pass itself is comparatively low, the season was advanced, and fresh snow had lately fallen. The climate was possibly more severe in those days than at present, and the hardships seemed more fearful to an army from the South.

2. The pass called afterwards the Graian Alp was one of the best known and earliest used across the mountains. By it, streams of invading Gauls had passed centuries before. Its neighbourhood was by far the most fertile of them all, and as such best suited to supply an army on the march.

3. Its outlet was nearest to the country of the Gauls who had sent to invite the Punic forces. Their envoys would naturally know it best, and be most likely to guide the invaders on that course. Intractable as the Romans found at a later date the tribe of the Salassi, who held the upper valley of Aosta, there was no reason why they should obstruct the passage of the enemies of Rome, and the other tribes, Libui or Lebeci, who were settled lower down, may well have followed the policy of the powerful Insubres, and sympathized in their alliance with the strangers. Polybius therefore did not stay to mention them, indeed from the island of the Allobroges he records no names until he makes Hannibal issue from the Alps among the Insubres, the leading state of the Gallic confederacy against Rome. He gives his reasons for this silence, in the general ignorance that prevailed of the exact position of the tribes and localities in question.

Livy wrote, however, at a later date, when the Alpine tribes and names were far better known to the Italian public. His information therefore is more definite in that respect, and

seems to point to an entirely different route from that which has been traced above. After speaking of the civil strife among the natives of the Island, and then of the supplies furnished to Hannibal in return for his decisive succour, he makes him turn ad lævam in his way towards the Alps. Of the various explanations offered of this phrase, two only give a likely meaning. The first assumes that he retraced his steps across the Isère and down the Rhone, and then turned to the left up the banks of the river Drôme. The other view leaves the bulk of the army on the South of the Isère, while a detachment only crosses to decide the civil war, which done, the whole continues on its march along the Eastern or left bank up to Grenoble. But the latter version can make little of the words in Tricastinos flexit which are coupled with ad lævam, for the Tricastini lay further to the South, and their chief town, called afterwards 'Augusta Tricastinorum,' may be most probably identified with Aoste on the Drôme, though by some placed lower down near St Paul trois châteaux' upon the Rhone. The advocates therefore of the march up the Isère, assume that the words in question have slipped out of their proper place in a passage which describes the march up the river towards the Island. Accepting the earlier explanation we may follow the track described by Livy along the Drôme up to Aoste, and thence to Die, which stands for Dea Vocontiorum, a powerful tribe here mentioned by him, whose northern borders reached up to the Isère and the Drac, while their frontier on the South-East extended far along the road to Gap and to Embrun, through which country Hannibal may have led his troops, skirting the lands of the Tricorii who were spread to the North-West. He would thus have reached the Durance, the Druentia of Livy, and have made his way to Briançon, and across the Mont Genèvre, known to the Romans of the time of Cæsar as the Alpes Julia, though afterwards called Cottiæ, after the native chieftain who did so much to improve the mountain roads about him to win the favour of Augustus. It would seem to have been the same route, though in a contrary direction, which Julius Cæsar followed in his march into Transalpine Gaul, as indicated in the words 'ab Ocelo, quod est citerioris provinciæ extremum, in fines Vocontiorum ulterioris provincia die septimo pervenit; inde in Allobrogum fines.' B. G. 1. 10. 5. It is the same track also in the main by which Livy v. 34 brings Bellovesus with his Gallic hordes through the Tricastini, and the Taurini Saltus into the plains of Lombardy where they settled, at the end of the regal period of Rome.

From the Island to the ascent itself, the narratives of Livy and Polybius have no points in common, the local names furnished by the former being entirely absent in the latter, while the other conditions of the march are quite distinct.

But in the description of the pass itself, in the struggles with the mountaineers, in the measurements of time, and in the accounts of the dangerous point of the descent where the road was swept away, there is often very close agreement in the language of the two, though Livy adds a few details such as those of the use of vinegar and fire to clear a passage through the rocks. The incidents, however, which they have in common are just those which can most easily be localized in any of the rival routes, and they must be regarded as the least important evidence upon the subject. But in chapter 38, when Livy has brought the Carthaginians to the plains of Italy, he pauses to notice the different opinions which were stated, and to give his reasons for the route which he had traced. The Pœnine Alps, the great St Bernard, seems to have been commonly regarded as the pass of Hannibal, and stress was laid on a false derivation of the word, as if it came from Pœnus. It was enough, he thought, to urge in answer that Germanic tribes held the entrance to this pass, and there could have been no motive to brave the stout resistance which they would probably have offered. The earlier writer Cælius Antipater, whose work on the Punic wars was largely used by Livy, brought the invaders through Cremonis Jugum,' a mountain unknown to other authors, but which may remind us in its sound of the Cramont, and at any rate closely corresponds to the Little St Bernard, called the Graian Alps by Roman writers. But this leads into the Italian Val d'Aosta, the upper part of which was occupied by the Salassi, while the Gallic Libui held the lower country. Tradition commonly, says Livy, knows nothing of these names in this connection, but makes Hannibal issue from the mountains through the tribe of the Taurini, with whom he first came into hostile contact. The Roman historian admits that there was no sure evidence before him, and that he relies mainly on tradition; the account of Polybius he did not notice. But tradition in this matter was a guide of little value. From the time when Scipio found himself too late upon the Rhone, till he faced his enemy on the Ticinus, the Roman government had entirely lost sight of the Carthaginian leader. The country through which he passed was quite unknown to them, and no trustworthy information could be forwarded to Rome, or lodged in the official archives. The Gallic mountaineers remained long unsubdued, and the eventful tramp of many a later army effaced from their minds the memory of the march of Hannibal. The popular legend of two centuries later was hardly likely to be accurate in such details. It was known indeed that the Taurini were attacked before the collision with the legions, and it was natural to suppose that they denied him passage when he moved along their valley, though Polybius tells us that he recruited first his