Notices of New BooJcs. 287 Two obvious objections to the adoption of the Mont Cenis as the pass crossed by Hannibal, are found in the words Ttapa rbv 7roTap.bv, Polyb. in. 50, which it has been supposed could only refer to the Rhone, and in the necessity of placing the Allobroges in the valley of the Isere. But Mr Ellis shews that it is impossible to reconcile the supposition of the army having continued its march along the Rhone, with the distances stated : not to mention other difficulties, and the fact that we have no notice of the passage of the Isere ; which, if the whole army crossed the river at its junction with the Rhone, must have been an operation of risk and difficulty. Moreover, the expression irapa rbv irorapov, without any distinct intimation of what river was meant, is exactly parallel with nepl rfjp pdx lv > m c - 55 - And although it is true that we find the Allo- broges described by later authors as located north of the Isere, yet, according to Strabo it would seem that this river was their southern boundary, and it may easily have happened that they may have been driven somewhat towards the north in the intervening period. Indeed Mr Ellis argues with much plausibility, that the name Allevard is really identical with Allobrica, and thus contains in itself a proof of the occu- pation of the valley by this tribe. Mr Ellis believes, with De Luc, that Hannibal crossed the Rhone about Roquemaure, and descended, of course, into the territory of the Taurini. Such was the unanimous testimony of all Livy's authorities ; and the single passage in Polybius, which seems to contradict it, is inge- niously explained as one of the succinct accounts or summaries which Polybius has throughout his narrative prefixed to the details of each march or engagement. It had long ago been remarked that it was difficult to comprehend why Hannibal, if he had descended into the territory of the friendly Insubrians, should afterwards have gone out of his way to engage the Taurini : and in the passage of Strabo, iv. 6, rf]u dia Tavplvav tjv wifias difjXde, Mr Ellis finds a new argument in favour of the Mont Cenis. Though his conclusions are based, in the first instance, upon the text of Polybius, Mr Ellis has not neglected that of Livy, by whoso assistance he supplies, in one or two points, the omissions in Polybius' narrative : the last chapter is occupied with an examination of the ancient roads across the Alps, which serves to strengthen the case already made out. In fact such a mass of evidence is accumulated in favour of the theory, as appears almost to amount to an absolute demonstration of its truth. Not the least interesting part of the book are the etymologies of the Celtic names. Thus the reading "Zmpas [2Kcopas in one MS.] in Polyb. ill. 49, in the description of the island, is successfully vindicated, as being a compound of Isca and Aras, the former element the same as occurs in our own Esk, Usk, Ouse, Exe ; and the latter identical with Aar, Arar, &c. The compound Isc-aras will then be the name of the river formed by the two streams jointly after their union : like the Durance, the Dordogne in Auvergne, or our own Thames. Most of tho derivations indicated recommend themselves at once. We may be per-
