the mountains, was the noble Parnassius Apollo, with its cream-coloured semitransparent wings marked with scarlet and black-ringed ocelli, always associated in the Alpine naturalist's memory with the grand scenery of the lofty mountains which it loves to haunt. They abounded here, sailing gracefully from flower to rock, so fearless that when settled they would allow one to touch their wings with a hand net before they would move, and then darted with astonishing swiftness far down over the edges of precipices into the deep valley, making one long for their powers of flight. With the Apollos were thousands of Polyommati of five or six species (Corydon and Argus most plentiful), their beautiful bright blue wings literally studding every patch of bloom. Hipparchias were numerous considering the lateness of the season, especially Galathea, and I captured some thirteen species in the morning's walk. Lower down the valley in sunny nooks these were joined by Clouded yellows, brilliant red Coppers, and silver spotted Fritillaries, most of them freshly emerged from the chrysalis. In phænogamous plants we found nothing new, and there were extremely few ferns.

At Bonneval the two valleys of Bellaval and Bonneval unite, and the little village is romantically situated at their junction, surrounded by orchards and rich meadow slopes under lofty cliffs of darkly-waving pines. A mountain torrent issues from the narrow opening into the Val de Bonneval, fringed with fine birch-trees, a wooden bridge spanning its roaring course where it joins that from the other valley. Thus reinforced it becomes a fine stream, the bright blue waters foaming and sparkling in the sunshine, most refreshing to the eye in the increasing heat of the day. It is sometimes called the Isère, of which, however, it is only a branch, and in the maps is designated as the Glacier or Chapiu torrent. The path, which crosses to the right bank of the stream, descends very slightly compared with the

rapid fall of the valley, as it is so extremely contracted that it is necessary to keep high up; consequently the depth between the track and the bottom of the glen is constantly increasing, until, after a short hour's walk, the torrent is seen foaming at a vast depth below, at the bottom of a lofty precipice. From this commanding point the wide valley of the Tarentaise opens out, with a distant view of Scez and St. Germain, and the snows and glaciers of Tignes, Mont Iseran, and the Vanoise.

The transition in the vegetation is rapid; and passing through a little village and under the refreshing shade of fine walnut-trees, we dropped quickly down among terraced fields of ripe corn, potatoes, and even maize, to the swampylooking plain, studded with poplars. Reaching the broad level road which runs through it, we found it fearfully hot and close, and after about a mile's walk were glad to reach the little half French, half Italian looking town of Bourg St. Maurice, the capital of the upper Tarentaise. Here we halted at noon, preparatory to starting for the Little St. Bernard, so celebrated as the line of Hannibal's passage of the Alps.

CHAPTER III.

VAL DU PETIT ST. BERNARD.

Hannibal's passage - Controversy - Evidence in favour of Little St. Bernard - Bourg St. Maurice - Costume of Tarentaise - Ascent of pass — "Rocher Blanche" - Hospice - Antiquities on plain - Geology — La Thuile Avalanches in gorge-Mont Cramont - View from summit - St. Didier.

THOUGH the controversy has from time to time been revived as to whether the Little St. Bernard or the Mont Genèvre, the Col de Sestrières, Monte Viso, Mont Cenis, or others still less probable, were the point by which Hannibal effected his famous passage into Italy, the general mass of inferential evidence, since the investigations of General Melville, followed by the works of De Luc, Messrs. Cramer and Wickham, and numerous other writers, was so clearly in support of that of the Little St. Bernard, that the question might be considered as fairly settled. The able modern historian of Piedmont has however, and I think unfortunately, revived the controversy, by unhesitatingly giving the weight of his authority "to the Piedmontese version in favour of Mont Genèvre."*

The account of Polybius, clear and truthful, and aided by his personal knowledge of, and travels over, the same Alps only 60 years after Hannibal's passage, and, as he states, for the express purpose, is now almost universally admitted as the only reliable one; Livy's confused and irreconcileable narrative being manifestly but a garbled and interpolated paraphrase of Polybius, with the object in view of fixing the passage on

* History of Piedmont, Antonio Gallenga, vol. i. chap. 2.

Mont Genèvre. This opinion Gallenga also adopts, following M. Letronne and the Chevalier Folard.

The main question, as far as speculative arguments are concerned, appears to me to hinge on what route the emissaries of the Cisalpine Boii and Insubres would be likely to take on their way to meet Hannibal, and invite him to make common cause with them against the Romans. By the same route which these messengers took on their way over the Alps into Gaul, the probability amounts to a certainty, that, in the absence of local difficulties insuperable to an army, they would guide or direct Hannibal's advance. The question then so far depends on which line of country was both most convenient and most friendly to their design; and then, whether that route agrees in its features and position with the facts and descriptions of Polybius.

As the district of the Insubres extended along the north side of the Po up to the foot of the Alps, and they were only separated from the Salassi of the Val d'Aosta by another minor Gallic tribe, the Libui, the shortest and readiest means of communication for these tribes with Gaul was naturally by Ivrea, the Val d'Aosta, and the Graian Alps. But in addition to this reason was a stronger one for their choosing this pass, resulting from the relations of the several tribes who held the country between the Allobroges and themselves. The Cisalpine tribes of the Insubres and Boii of the plains of the Po and Milan, with the Libui also,—like the Allobroges, the nearest Transalpine Gallic tribe, and who inhabited the "island" included between the Rhone and the Isère, were all of them Gallo-Celts, who in their habits and choice of location were always Lowlanders. But the pure Ligurian tribes of the old Celtic family, such as the Salassi, Taurini, Centrones, Graioceli, and others who dwelt on the steeps on either side of the intervening Graian chain, were of the first migrationHighlanders, in the true sense of the word, hardy mountaineers, whose only limit on their native Alps was the snow-line.

On a review of the history of the eight years previous to the second Punic war we find the aboriginal Celtic tribes of North Italy, and among others the Ligurian tribes of the Taurini of Piedmont, and the Salassi of the Val d'Aosta, leagued together with these immigrant lowland Celtic tribes. of Gallic origin-amongst whom were the Insubres and Boii— against the Roman legions, in the Gallic war which lasted for four years, and terminated, B.C. 222, in the complete discomfiture of the allied Celts. After that period the Taurini and other Ligurians withdrew from this alliance, to the Romans, to whom from motives of fear or policy they attached themselves; and were so far hostile to the Insubres as not only to be no parties to their invitation to Hannibal, but Polybius* tells us that Hannibal afterwards warred against them because they were preparing to attack his Insubrian allies. But we have no evidence whatever of the Salassi having broken their alliance with the Insubres, and motives of self-interest would prevent it, their fastnesses in the Val d'Aosta debouching on the Insubrian plains. By the Val d'Aosta and the Graian Alps, therefore, the communication must naturally have been kept up between the latter and the Transalpine Gauls. The other Ligurian tribes, as the Centrones, Graioceli, and Medulli, who held the mountain fastnesses between the two, being of the same stock as the Salassi, would no doubt allow the messengers of the Insubres, the allies of the latter, safe and free conduct over the pass as heretofore.

Nothing could be more improbable, on the other hand, than that the messengers would endeavour to reach Hannibal, especially on such an errand, through the declared hostile Taurini, as must have been the case had they taken the route of Mont Genèvre, as accepted by Gallenga and others. Not only would it have been most impolitic for such a deputation to have attempted to pass through a people who were not only their known enemies, but also the allies of Rome; but * Polybius, lib. iii. 60.