1^ 'OOHOSHBIH S3139WV SQ\ ^^\ ^ HANNIBAL ONCE MORE HANNIBAL ONCE MORE BY DOUGLAS W. FEESHFIELD, M.A. VICE-PBB8IDENT OF THE ROYAX GEOGRAPHICAL SOCIETY SOMETIME PRESIDENT OF THE ALPINE CLUB AUTHOR OF 'TBAVBLS IN THE CENTRAL CAUCASUH AND BASH AN' ' ITALIAN ALPS ' ' THE EXPLORATION OF THE CAUCASUS ' 'ROUND KAKGCHENJUNOA ' ETC. LONDON EDWARD ARNOLD 1914 lAU rights rtstrvfd] -\\s-5 6 FEB 1918 < -< o An, p PREFACE The following pages do not pretend to offer a summary, or analysis, of recent attempts to elucidate the historical episode they deal with. They are rather an endeavour to treat the classical texts relating to Hannibal's Passage of the Alps from the point of view of an Alpine traveller and I^ topographer. In undertaking this task I have '^i been impressed by the importance of a fragment I from Varro which has come down to us, and by j^ the curious way in which it has been either K ignored, or what seems to me its plain meaning CT) distorted and obscured, by most of my prede- cessors in this discussion. Incidentally I have ^ been led to examine somewhat closely the argu- ments recently brought forward by several French mihtary ^vriters in their effort to prove that Hannibal crossed the Col du Clapier, a lofty and difficult by-pass near the Mont Cenis. I have been urged to do this by the fact that their theory has been adopted and put forward in this country by the Professor of Military History in the University of Oxford, Mr. Spenser Wilkinson. DOUGLAS W. FRESHFIELD. March 1914. " CONTENTS CHAP. PAOB I. THE CLASSICAL TEXTS ... 1 IL THE DURANCE PASSES . . . .23 in. THE VARS-ARGENTIEEE ROUTE ... 31 IV. THE MONT CENIS AND ITS VARIANTS . . 58 APPENDICES— I. Francis the First's Passage of the Col db L'Arqenti^re ..... 100 11. M. Chappuis' Theory .... 107 HI. The Roman Pass from Sdsa to the Maurienne 110 IV. Mr. Marindin on Fochs and Osiander . . 118 V. A Short Bibliography of Recent Publications 119 MAPS The Pas de la Reyssole . . . opposite page 36 The Col du Clapier and Mont Cenis Passes ,, „ 78 The Western Alps . . . . . at end ILLUSTRATIONS Larche and the Ascent to the Col de l'Argentiere .... The Valley of the Ubayette, looking west ..... opposite page 23 The Lac de la Madeleine, looking west \ The Col de l'Argentiere, looking to- J- „ „ 41 wards Italy . . . , J A Scene near Sambuco, from Brocke- don's 'Passes of the Alps' . „ „ 48 We were hinted by the occasion, not catched the opportunity to write of old things, or intrude upon the antiquary. We are coldly drawn unto discourses of antiquities, who have scarce time before us to com- prehend new things, or make out learned novelties. Sir Thomas Browne. HANNIBAL ONCE MORE CHAPTER I THE CLASSICAL TEXTS More than thirty ^^'inters have passed since I was first tempted to disregard Dr. Johnson's depreca- tion of any further mention of the second Punic War and to take part in the old-world dispute as to the pass by which Hannibal crossed the Alps. From time to time, as occasion arose, I have made some further contributions in the way of criticism of the theories set out by more recent commen- tators.^ I am now induced by a desire to put all I have found to say into a more convenient and final form to return, somewhat unAvillingly, to the familiar story, and to give my contemporaries and the critics of the future an opportunity, if they care to use it, of weighing my arguments, positive as well as negative, in the balance. For like most, if not all, of those who have been dra\vn into this captivating controversy, I have been led to hazard a solution of my owai of the problem under discussion. It is one practically new to English readers. In the course of my 1 See Alpine Journal, vol. xi. p. 267 (Xo. 82), vol. xiii. p. 28 (Xo. 93). Proceedings R.G.8., n. s., vol. viii. pp. 138-188, October 1886; Oeo- grafkical Journal, vol. xxxvii. Xos. 4 and 6. A 2 HANNIBAL ONCE MORE Alpine wanderings I have learnt local facts which satisfy me that, to put it at the lowest, as good a case can be made for the pass the claim of which I desire to bring forward as for any of those which have hitherto found favour in the eyes of historians and scholars. Further, in the course of my literary inquiry I have been led to question in several not unimportant points the statements as to physical facts and features hazarded by historians, as well as to dispute several of the conclusions drawn from the classical texts by scholars from whose opinion I differ with both respect and reluctance. My in- vestigations, I ought to add, have in more than one instance induced me materially to modify, or vary, my own views as expressed in earlier articles. It was accident that first drew me into the troubled waters of the Hannibalian controversy. ' Some there be,' wrote Sir Thomas More, ' that have pleasure onelye in olde rustie antiquities, and some onelye in their owne doynges.' It is a commonplace to accuse the tribe of mountain climbers, of which I am counted a member, of falling under More's second category. I set up no claim to be an antiquarian ; nor have I any natural inclination to join in heated discussions of minor problems of history. But when one day I read the pages which the late Mr. Bosworth Smith, in his interesting and popular work on Carthage and the Carthaginians ^ devoted to the Pass of Hannibal, I found that my alpine and geographical sense were both outraged. The narrative appeared to me as unsatisfactory as it was vivid and picturesque. ^ Carthage and the Carthaginians, by R. Bosworth Smith (2nd edition). Longmans, 1879. THE CLASSICAL TEXTS 3 Physical facts were adapted to its requirements without apparently any scruple or hesitation. My own knowledge of the Alps, general and particular, convinced me that the author's attempt to fit the story told in the classical texts to the Little St. Bernard was a complete failure. The adventures and incidents described by the ancient historians, however credible in themselves, could not possibly, it seemed to me, have occurred in the locahties assigned to them by the modern writer. On further investigation, I discovered, somewhat to my surprise, that this view of mine was far from being the popular one ; that, on the contrary, the partisans of the Little St. Bernard in this country, sustained no doubt by the great authority of Mommsen, had taken of recent years to speak and write as if the question had been finally decided in their favour. And this, notwith- standing the weighty dissents of a geographer so competent as the late Sir E. Bunbury, and of Alpine experts such as Professor Bonney and the late John Ball. Where I had expected to find myself introduced to a number of more or less plausible hypotheses, contending amongst them- selves with becoming modesty, I was met by what seemed to me the least plausible, stalking abroad with all the airs and assurance of accepted truth. At such an assumption I naturally rebelled. And my indignation grew when I found the popular theory illustrated and supported by a series of collateral statements, many of which were in direct opposition either to any history and geography I had been taught in my youth, 4 HANNIBAL ONCE MORE or to such acquaintance with the Alps as I had gained in more than twenty summers. The victors, so it seemed to me, had won the battle not so much by finding a way to bring the narra- tives into agreement with the physical facts, as by accommodating the facts, and sometimes the nar- ratives, to their theories.^ In short, the more I saw of the handling of the evidence by which a judg- ment had been obtained in favour of the Little St. Bernard the less inclined did I feel to acquiesce in that judgment without at least a very full and independent rehearing of the whole case. I did not, however, at the outset promise myself to arrive at any certain conclusion as to the pass traversed by Hannibal. For I noticed that the critics most eager to assert such a conclusion were frequently those of least general insight and local knowledge. But it seemed to me worth while to take the first step in any pursuit after truth, and to attempt by the elimination of mitenable hypotheses to narrow the field of search for others. In order to undertake this task with as little 1 The late Mr. Capes in his school edition of the xxi. and xxn. Books of Livy supplied a note, in which, while summarising fairly the principal theories of previous writers, he collected in a concentrated form some of their most palpable perversions. Witness the following sentence, in which the Mont Genevre is compared to its disadvantage with the Little St. Bernard : ' General probabihties are also in fevour of the easier, the lower, the better known, and the more favourably placed of the two passes.' Thi-ee of these statements are the reverse of true, the fourth is contestable. The Little St. Bernard is steeper and longer, more remote for those who come from Spain, and higher by 1100 feet than the Mont Genevre. A new edition of Book xxi., with Mr. Capes' original notes revised by J. E. Melhuish, has been published in Macmillan's Elementary Classics, 1910. The notes still show a very imperfect acquaintance with the Western Alps. The writer even retains a belief in ' Mount Iseran ' ! (p. 104). THE CLASSICAL TEXTS 5 prejudice as possible, I determined to keep out of sight all the modern literature on the subject until such time as I should have personally ex- amined the classical authorities, and made up my own mind on their evidence alone. Every year geographers are taught by some fresh instance, that while some of the best-abused statements of ancient authors are literally true,^ others are exaggerations rather than absolute in- ventions. I approached therefore the classical authorities, expecting to find consistency — if con- sistency there were — in outline rather than in detail. I had little hope of being able to harmonise every point in the narratives of Polybius and Livy. But I was not disposed to assume in the face of a general agreement that minor discrepancies would necessarily prove the two accounts to be based upon irreconcilable authorities, intended to refer to widely different routes, and, therefore, one or the other fictitious. In discussions of this sort historians and school- masters seem to me not infrequently to create half their difficulties by a certain failure in historical imagination. They omit an essential preliminary step. In place of adequately realising the geo- graphical standpoint of the author they are dis- cussing, and of the contemporary audience he addressed, they, unconsciously, impute to both their own knowledge. For instance, we cannot understand and fairly appreciate Polybius, unless we first ascertain not only what his general con- ^ For example, the statement of Ptolemy as to the Nile deriving some of its waters from snow mountains. I have myself seen one of the Nile sources issuing from a glacier on Ruwenzori. 6 HANNIBAL ONCE MORE ception of the Alps was, but also how far he might reasonably expect his readers would be able to follow him. Now the Greek historian, we find, pictured the Alps as mountains extending from their junction with the Apennines, east of Marseilles and west of Genoa, across the whole head of Italy. They were, he writes, drained by two great rivers — by the Rhone on one side, flowing generally south- west in a deep valley parallel to the mountains from a source north of the Adriatic ; on the other by the Po, running in its early course southwards into the Cisalpine plain. Except in so far as it exaggerated the course of the Rhone, this picture was roughly correct ; for it is only convention that fixes on the stream from Monte Viso rather than on the river from Mont Blanc as the true source of the Po. But it was exceedingly vague, and it did not lend itself to any nice detail. In fact we shall find that Polybius, in telling the story of Hannibal's march, recognises only three broad divisions — the Rhone Country, the Passes of the Alps, and the Plains of the Po. The greatest of the difficulties in fitting the story told by Polybius to local facts has been found in the length he assigns to Hannibal's march along the river. It is, I admit, a serious difficulty for those who are not prepared to make large allowance for the loose way in which the Greek historian is apt to use geographical expressions. In my behef, the only satisfactory way out of it is to mider- stand Polybius as including in the phrase ' along the river ' the whole Rhone basin. Support for this solution may be found in his broad use of the THE CLASSICAL TEXTS 7 same, or similar,^ phrases elsewhere, and in the fact that no march of 1400 stades along the river bank could have brought Hannibal to any Alps. I may note one somewhat parallel case. The explanation which has satisfied Mr. Strachan- Davidson ^ with regard to Appian's assertion that Pompey'sPass was close to the sources of the Rhone and the Po is based on the principle I am here apply- ing to Polybius. The Master of Balliol refers to Strabo's statement that the sources of the Durance and the Dora Riparia are close to one another, and considers it reasonable to believe that Appian borrowed a statement made respecting the sources of the two tributaries and applied it to the main streams. Now if classical authors thought and wrote thus loosely, if they held it legitimate to treat the springs of any important tributary of the Rhone as the Rhone sources, a similar licence must surely be allowed them with respect to the basins of the tributaries. If this be admitted the difficulty vanishes, and even such a phrase as Trap avTov Tov TTOTafxov may be taken as equivalent to ' in the river basin.' ^ Into this vague framework Polybius did his best to fit the details he had collected from eye- witnesses of Hannibal's march, or from previous chroniclers, and so to fit them that they should be intelligible to readers who had no maps, whose orography was even vaguer than his own, and to ^ Compare 6 Kara t6v 'Po8av6v avXwv, and ol Kara TOV 'Po8avov roiroi. 2 See Appian, Book »., with notes and maps by J. L. Strachan- Davidson (now Master of Balliol College, Oxford), The Clarendon Press, 1902, pp. 131-2. 2 See on this point Mr. Marindin in Classical Review, vol. xiii. p. 243. 8 HANNIBAL ONCE MORE whom the local or tribal names he had gathered on his own Alpine travels w ould, as he thought, convey no definite meaning. For their sake, with a very perverse indifference to posterity, he left out these names. ^ The merit of Polybius's narrative has, in my opinion, been strangely overrated ; it seems now in some danger of being as unjustly depreciated. He was content to tell the story in the form he believed to be best suited to the readers of his own time. He took no thought of the better-instructed public of the future, and he must consequently bear the blame of having thereby caused an infinite waste of time, temper, and ink to many genera- tions of students. But in my judgment we do him some wrong, and labour to our own confusion, if we take his round-figure distances, his loose geographical expressions, his rough indications of direction, or of the broad sections into which he divides the march, and read them strictly and narrowly in the light of our own intimate acquain- tance with the Western Alps. I would earnestly beg future students not to lose sight of these, as they seem to me, fundamental considerations while following the subsequent argument. Polybius is, as I have just pointed out, notoriously scanty in local nomenclature. We have to make the most of such few indications as we can extract from his narrative. The Carthaginians, he tells us, crossed the Rhone at a spot about four days' ordinary march from the sea, and at a distance it took Hannibal four days to march (the distinction * I agree with much that Commandant Colin has written in the volume I comment on in another chapter, on the literary method of Polybius. See also for a full discussion of the matter, W. T. Arnold's Second Punic War. THE CLASSICAL TEXTS 9 must be noted) from the confluence with the main river of another stream. The name of this second stream is in the MSS. S/copa?, but the emendation 'Icra/oa? has been generally adopted in the best modern editions. (See Mr. George Long on ' Isara,' in Dictionary of Greek and Roman Geography.) ^ Geographical details are next furnished as to the district lying between the rivers, and it is said to be called ' The Island,' and compared as to form and size with the Egyptian delta. The following distances are given for the Alpine march : from the Passage of the Rhone to the Ascent of the Alps leading to Italy, 1400 stades ; The Passes of the Alps, 1200 stades. In another place we are told that Hanni- bal marched 800 stades along the river from the Island to ' the Ascent towards the Alps.' Most commentators have naturally considered the Ascent of the Alps leading to Italy (c. 39) and the Ascent towards the Alps (c. 50) to refer to the same spot, and as a consequence the 800 stades (88 miles) to be part of the 1400, leaving 600 stades (66 miles) between the Passage and the Island. But the two writers I shaU have to deal with presently. Commandant Colin ^ and his follower Professor Spenser Wilkinson, will have none of this. For them the 800 stades begin from their Island, Bedarrides ; there are two dis- tinct Ascents of the Alps, and the remaining 600 ^ This is the reading given in the latest edition of Polybius issued in this country. ^ Commandant Colin translates (p. 338) ava^oKi] eiUrte in one passage, and montee in another to suit his convenience ! Professor Wilkinson's argument may be found, Geog. Jmirnal, vol. xxxvii. No. 6, p. 674. 10 HANNIBAL ONCE MORE stades have to be made up with the distances from Fourques to Bedarrides (32 miles), and from St. Nazaire, their 'Ascent towards the Alps,' to the Bee d'Echaillon (23 miles), which are inadequate ! [Where not otherwise specified, ' miles ' in my text implies English miles.] To me this seems a fan- tastic perversion of the plain sense of the text of Polybius. ' Times ' also are given us. Hannibal marched in four days from the Passage to the Island. He occupied ten days in his march up the river and fifteen in crossing the Alps. In the view of Commandant Colin, Hannibal covered 32 miles in the first four days, and about 90 in the following ten. In the mountains, however, he estimates his marches as from 13 to 17 miles a day.^ Further indications of the route adopted have to be sought in the description of a battle with the AUobroges, and in the statements — some- times held to be conflicting — that Hannibal de- scended boldly into the territory of the Insubres, and that he took by storm the chief town of the Taurini. There remains to be considered a reference to Polybius found in Strabo which runs as follows : ' He mentions four passes only ; that through the Ligurians, nearest the T3nrrhene Sea ; then that through the Taurini, which Hannibal crossed ; then that through the Salassi ; and fourth, that through the Rhaetians, aU rugged.' On this I would remark, that the fact that Polybius here speaks of Hannibal's Pass as the ^ On p. 424 four of the day's marches are reckoned at 25 kilometres each ; on p. 369 as 20 ; on p. 383 another estimate is given. THE CLASSICAL TEXTS 11 Taurine Pass may fairly be taken as a confirma- tion of the belief that from his point of view he was including the fighting with the Taurini in the Passage of the Alps, and consequently as a help to the understanding of the phrase as to Hannibal's descent on the Insubres. Another inference may fairly be drawn from this passage, that Hannibal's Pass was not the Mont Genevre. For if the Mont Genevre had been already known as Hannibal's Pass in Polybius's time, Pompey would hardly have ventured to assert that by crossing it he had opened a new route over the Alps. This claim to have found a new Alpine pass was put forward by Pompey himself in a letter to the Senate pre- served by Sallust. He writes : ' I have opened a road over the Alps different from Hannibal's, and for us more convenient.' Pompey' s object was Spain, his immediate starting-point and military base was presumably Mutina (Modena), whence he had set out to visit Rome not long before. It was at Mutina that the three roads from Rome to GaUia Cisalpina, the Aurelia, the Cassia, and the Flaminia, met (Cicero, Phil., xii. 9. 22). Pompey's natural approach to the Western Alps in this case would lie north of the hills of Montferrat, and the vaUey of the Dora Riparia and Mont Genevre would cut ofiE thirty-five miles in his march to the Durance and to Spain. I find an additional reason for beUeving the IMont Genevre to have been Pompey's Pass in the fact that twelve tribes in the kingdom of Cottius, which extended on both sides of the Mont Genevre, were omitted from the Ust of the tribes subdued by Augustus inscribed on the trophy of Turbia 12 HANNIBAL ONCE MORE above Nice (Tropaeum August!) , on the ground that they had been previously tranquillised.^ Is it not likely that Pompey's route lay in this direction rather than among the turbulent tribes to the southward whose names are recorded on the trophy ? Was it not this road which Cottius at a later date made more passable at great exertion and cost ? I am glad to find Mr. Strachan - Davidson adhering without hesitation to Mommsen's conclusion, that Pompey selected the Mont Genevre route. ^ On the other hand, Mr. Marindin's suggestion that Pompey's new pass may have been the Mont Genevre plus the Col de Lautaret, as contrasted with the Mont Genevre plus the Col Bayard, does not at all commend itself to me.^ An army at ^ These twelve tribes were granted the Jus Latii by the Lex Pompeia, carried B.C. 89 by Pompeius Strabo, the father of Pompey the Great. See Ellis, On the Ancient Routes helireen Italy and Gaul. Cambridge, 1867, 2 I agree with all Mr. Strachan -Davidson writes about the passages in Strabo (iv. 6. 5), and Appian (i. 109). They seem to me to fall in with the view expressed above. ^ See Classical Review, vol. xiii. No. 5, on this point. Mr. W. T. Arnold wrote to me, 14th June 1899, the note here quoted. It may, I hope, help to convince Prof. Wilkinson that our late friend formed an independent judgment on the texts and cannot fairly be described as having been my ' shadow ' in the controversy. ' The Romans,' wrote Mr. Arnold, ' had a clear idea of the main Alpine wall, and would not rank a new way over secondary obstacles as a new pass " per Alpes." Moreover I have always felt that SaUust's " nobis opportunius " distinctly pointed to the low and easy Mont Genevre as opposed to a rather more difficult pass (i.e. your Col de I'Argentiere) not far off. As I see the whole matter, the evidence converges on a Durance pass. The best Durance pass is certainly the Mont Genevre. But Varro and Sallust point rather to the Argentiere — and so the conclusion is that of my Second Punic War.' Again, four years later, he wrote (16th June 1903) : ' There are two points about the Second Punic War in the April English Historical Review which tempt me to send it a letter — Hannibal's Pass, p. 330, and the Elephants, p. 329. I want to take the occasion to give Osiander a bit of a knock, especially as aU the German Classical Reviews treat him more seriously than he deserves. I have a reference to his clumsy attempt to explain away the Varro passage.' THE CLASSICAL TEXTS 13 Brian9on bound for Aries and Spain would not find in the Lautaret a shorter and easier road than that down the valley of the Durance. ]\Iore- over, the Roman habit was to look on the routes through the Alps as a whole, and to take little account of separate minor ridges and consequent ups and downs. So far Polybius. Now let us take up Livy. But before doing so, I must quote some remarks of Mr. Marindin which seem to me much to the point. He writes : ' In studying the question I have been struck by one or two phases of con- temporary writing. One is the unceremonious way in which any ancient writer who traverses the course selected by the modern essayist is swept aside by the most fanciful arguments, as if he were no obstacle at all. ... It is surely time to protest against the zeal for rewriting history, and also against the modern method of criticising through a microscope which shuts off a view of the surroundings. It is the fashion so to deal with Livy. We are told he need not be reckoned, because he only cared to write picturesquely. We may alter the names of places, admire cimi risu his descriptions ; v.hen he says ad laevam, we may, with one commentator, say he means to the right. . . . Now I believe this is in a great measure nineteenth- century arrogance and to some extent nineteenth- century haste and impatience. I feel sure that many are far too ready to discount Livy's evidence ; that, at any rate, if he did not always use his materials critically, he used them honestly, and he used them more critically than is sometimes supposed. On this occasion it is pretty clear that hejiad the narrative of Polybius before him 14 HANNIBAL ONCE MORE (though some have denied it) and he tells us himself that he was using the account written by L. Cincius Alimentus, a Roman officer who had been a prisoner in Hannibal's camp at one time, and was more likely to know the details of the cam- paign than even the most recent critics of Livy.' I may add that Livy, probably, had also before him the works of Silenus and Sosilus, two Greeks who were in Hannibal's camp during the march, and of L. Coelius Antipater, a writer contem- porary with the Gracchi, and esteemed, though criticised, by Cicero, of whom only a few frag- ments remain. I have quoted these sentiments of better scholars than myself in order heartily to endorse them. While admitting to the full the difference in value between the statements of Polybius, the almost contemporary chronicler and traveller, and Livy, the later compiler and man of letters, I must give Livy the credit he is at pains to claim. In more than one passage he insists that he has taken especial care to sift the evidence in order to tell the story of Hannibal's Alpine adventures as accurately as he could. That he may have fallen into mistakes and confusions in the attempt to combine his scattered literary material seems to me more than probable. But I am not prepared to foUow those who believe and assert that he introduced, or copied, a string of tribal names without reason or reflection ; that he invented an Alpine route for Hannibal and then compiled an itinerary to fit it. I cannot think it reasonable to assume that a writer who had under his eyes the narratives of chroniclers who had been with THE CLASSICAL TEXTS 15 Hannibal in his Alpine march, and of one who had been a prisoner in Hannibal's camp a few years later, was completely mistaken as to the river basin by which Hannibal approached the Alps. The descriptions of scenery interspersed in Livy's narrative, so contumehously discarded by some modern writers, I fmd — and on this point I may, perhaps, be allowed to speak with some authority — coloured, if coloured at all, only to such an extent as to represent the local features of the Western Alps from the point of view, not of a tourist of the twentieth century, but of an African soldier in 218 B.C. The accounts of the difficulties encountered from autumn snowfalls correspond both with my own experience and local information, and therefore invite my belief. Moreover, they indicate a close acquaintance with Alpine phenomena which the comments made on them by many historical students have taught me is far from universal even at the present day. And when we come to the famous story of the rocks split by Acetum, the laugh is whoUy on Livy's side. It is surely time that the ridicule cast on the historian on the strength of a familiar line of Juvenal was thrown back on his critics. Had these learned scoffers been at the pains to use their bookshelves, they might have learnt that ' acetum ' was in practical use in Roman times for the very purpose for which it is said to have been employed by Hannibal.^ One is tempted to quote ^ See Pliny, xxiii. 27 and xxxiii. 21, where ' acetum ' is said to be used for splitting the rock in Spanish mines. ' Aceto summa vis est in refrigerando, non tamen minor in discutiendo . . . saxa rumpit infusum quae non ruperit ignis antecedens.' Dion Cassius and Apol- lodorus are cited to the same effect. In the diploma granted to the Marquis of Saluzzo by the Emperor Frederick vi. ' acetum ' is men- 16 HANNIBAL ONCE MORE Carlyle, ' Truth is better than error, were it even on Hannibal's vinegar.' As far as the general condition of Alpine tracks in late autumn is concerned, Livy's narrative strikes me as not only credible, but convincing. It is accurate in technical details, which could hardly have been described so graphically except from actual experience of the Alps on the part of the writers from whom Livy borrowed. It is mainly the lack of any such experience that has led many modern writers to misunderstand and depreciate the classical narratives, to make Hannibal, like a modern Alpinist, go out of his way to cross a glacier by a 'new route,' and to commit themselves to sundry other curious incongruities. But there are further questions of detail to be dealt with. How far can such local indications as are supplied by Livy's nomenclature be followed and trusted ? In what points, if any, is his narra- tive irreconcilable with that of Polybius ? Now Livy tells us that the quarrelsome chiefs ' near the Island ' were Allobroges, and that after settling their dispute, Hannibal turned to the left into the district of the Tricastini. After passing through the further border of the Vocontii, he traversed the land of the Tricorii, and met with no difficulty until he came to the Durance. Beyond that river he fought his first battle, then marched for tioned as one of the means employed to make the Traversette tunnel. See Vaccarone, Le Periuis du Viso. Turin, 1881. I may venture to mention a personal experience. WhUe driving down Val di Stura, I was interested to find a smallholder clearing his field of a big boulder by an expedient resembhng Hannibal's. He lit a bonfire round it and then threw on pails of a hquid I am not prepared to say may not have been acetum. I am inclined, however, to believe that it was cold water. THE CLASSICAL TEXTS 17 three days through comparatively open country, and had a second and very severe engagement in a defile. For the rest of the route Livy agrees in the main with Polybius. He brings Hannibal down upon the Taurini before he reaches the territory of the Insubres. Our gains in nomenclature from Livy are the three tribal names and the mention of the Durance. The chief discrepancies between the Roman author and Polybius are the statements (1) that the rival chiefs near the Island were Allobroges, (2) that the first conflict took place beyond the Durance, and (3) that the foes in it are described as Gauls, or barbarians, and not as Allobroges. Mr. Strachan-Davidson, I find, tells his readers that ' there is no reason to believe, though it is, of course, a possible supposition, that Livy took the names of the Druentia, and of the various tribes mentioned, from any previous writer.' He surmises that Livy chose his pass first and then provided local names to fit in with it. But here the soldier corrects the scholar. Commandant Colin (p. 191) cites a passage from Timagenes, an Alexandrian born twenty-five years before Livy, who wrote history in a turgid style in the days of Augustus and secured the immortality of a cursory mention from Horace {Epistles, i. 19. 15). The passage (preserved by Ammianus Marcellinus XV. 9. 10) gives a summary of Hannibal's march in which the same tribal names occur, and the Durance and the ' acetum ' find mention. A fresh touch of local colour is added by a reference to ' the gorges of the Tricorii,' since the deep trenches in which the streams issuing westward from the B 18 HANNIBAL ONCE MORE Alps of Dauphine flow are the only interruptions to the generally gentle gradients of the roads in this otherwise open and fertile district. The Durance is brought in oddly at the end. ' Having crossed the Durance and its shifting rapids he took possession of the region of Etruria.' Com- mandant Colin surprises us by the remarkable suggestion that the Durance in this passage may be identified with the Tanaro ! Now can it reasonably be doubted that Livy and Timagenes both drew from a common source ? It would be inconsistent with Livy's own account of his method, if he copied directly from one who was almost a contemporary. The conclusion I incline to after examining both Polybius and Livy is that Polybius's nar- rative (for reasons that will be given presently) excludes the Little St. Bernard, and that Livy by his description of the valley of the Durance confines our search to a pass leading out of the basin of that river. For whatever inconsistencies in detail Livy may have faUen into, I find it impossible to assume that he was entirely misled on so vital a point as the river-basin by which the Punic host approached the Alpine water- shed. But have we no further evidence to caU ? What if we can summon a Roman general who him- self knew the Western Alps,^ who was not only the most prolific but the most learned author of his age, whom Cicero caUed ' diligentissimus investigator antiquitatis,' who was born only a few years after the death of Polybius, was the * See Preface to Varro on Farmitig, translated by L. Starr, 1912. THE CLASSICAL TEXTS 19 leading literary authority at Rome, the custodian of its great library in the time of Caesar and when Livy was writing his history, whose fame reached down the centuries, so that a mediaeval scholar hailed him as ' veterum omnium doctis- simus,' Dante inquired affectionately for him in the after-world, and Petrarch called him 'il terzo gran lume Romano,' coupling his name with Virgil and Cicero.^ How if we can ask Marcus Terentius Varro what he knew about the passes of the GalUc Alps ? He is ready at our call. By a happy chance the following passage from his writings has been pre- served in the commentaries on Virgil {apud jEneid. X. 13) of Servius, a grammarian of the fifth century. ' The Alps,' writes Varro — he is speaking of the Gallic Alps — ' can be crossed by five passes, one near the sea, through the Ligurians ; the second by which Hannibal crossed ; the third by which Pompey went to the Spanish war ; the fourth by which Hasdrubal came from Gaul into Italy ; ^ the fifth which was formerly occupied by the Greeks and is hence called the Graian Alps.' No sensible person, I should hope, will argue that Varro did not know which was Pompey' s Pass, but his opinion as to Hannibal's and Hasdrubal' s is of course no more than the best 1 Lassels, in his Voyage of Italy (Paris, 1670), calls Varro ' the all- knowing Varro.' Lassels gives a description of the Mont Cenis and other passes, ' the Sampion, S. Godarde, Splug, hill enough for any traveller,' ' the great hill BerUno and Mount Aurigo, a smart hill.' The last two are the Bernina and Aprica. ^ I may mention here that Hasdrubal is stated by Appian to have passed the Pyrenees at their western end. He traversed the territory of the Arverni (Auvergne). He would therefore naturally strike a more northern pass than his brother. 20 HANNIBAL ONCE MORE opinion of his time. Still, while admitting this, it is hardly hkely — I might use a stronger ex- pression— that Varro, Pompey, and Sallust were, as Mr. Strachan- Davidson suggests, all sharers in what Livy is at pains to describe as a 'vulgar error,' the belief that Hannibal crossed one or other of the Saint Bernards and came down the Val d'Aosta. Had this been the case, Livy would surely have used a less disrespectful expression in alluding to the believers in one of the Val d'Aosta Passes. It seems to me that to any one who sets honestly and without prejudice to work to identify Varro' s Passes, there can be no legitimate room for doubt about the first and fifth; they can only be the Coast-road and the Little St. Bernard. I have already shown what I must hold to be conclusive reasons for assigning the Mont Genevre to Pompey, reasons which are accepted by almost aU scholars from Mommsen onwards. Is it not then reason- able to assume that Varro in his catalogue took all the five passes in their geographical order, begin- ning from the sea ? Is it not the only plausible assumption ? If this be so, we have to look for two passes — one south of the Mont Genevre for Hannibal, and one north for Hasdrubal. Now there are to this day two main routes across the Alps in the situations required, the Col de I'Argentiere and the Mont Cenis. These are modern highroads, and, though the existence of a modern road does not in itself prove an ancient route, it does show that the natural features of the gap in the chain traversed by it favour traffic. The argument for taking Varro' s Ust of the THE CLASSICAL TEXTS 21 passes according to their geographical order is, I think, forcibly strengthened by the straits to which the able critics who have tried to take it in any other way have been reduced. Consider for example Mommsen and Mr. Strachan-Davidson. Mommsen, who deals with the matter in his Inscriptiones Galliae Cisalpiiiae Latinae, suggests that Varro's second pass must be the Great St. Bernard. In order to reconcile this guess with his previous adoption of the Little St. Bernard for Hannibal, he is forced to impute to Varro the vulgar error denounced by Livy. For Varro takes Hamiibal over his second pass and mentions the Graian Pass fifth. Varro's fourth pass Mommsen is at a loss to identify, but tentatively guesses it may be the ' Col de Monte Viso,' or ' by Val Chisone.' ^ jMr. Strachan-Davidson also takes the view that Varro shared in what Livy calls a ' vulgar error,' and led Hannibal over the Great St. Bernard. This belief leads him into some obvious incon- sistency. He writes, on the same page, ' Varro has just before said that Alpes proprie montiuin Gallicorum sunt, and it is clear that he is here speaking only of the ways from Italy into Gaul.' But Livy has told us that the Great St. Bernard led among half-German tribes, and Varro was already sixty when in B.C. 57 one of Caesar's lieutenants made an abortive attempt to open the pass by subduing the tribes who held its 1 See Iiucr. Gall. Cis. Latin., vol. v. Part 2, p. 809. Mommsen pro- bably meant to refer to the Col de la Traversette near Monte Viso, and the Col de Sestri^res, a branch from the Mont Genevre, the descent of which towards Italy leads through Val Cliisone. Neither of these is a great pass. 22 HANNIBAL ONCE MORE northern approaches. It was not till a much later date that it was improved and finally made practicable for wheeled vehicles.^ Why then should Varro count it as among the practicable routes into Gaul ? But this is far from being the worst dilemma into which the surmise that Hannibal may have crossed the Great St. Bernard, or that Varro thought he did, lands those who indulge in it. Two pages further on, Mr. Strachan Davidson refutes his own argument. In writing of Pompey he says : ' The Great St. Bernard Pass would have taken him round an unnecessary circuit of at least two hundred miles on his way to Spain and that by a very difficult route.' An obvious and incontestable geographical fact ; but equally true, of course, of Hannibal going in the opposite direction. Whatever may have been the popular belief, the suggestion that Hannibal's Gallic guides misled his army so shamelessly as to take it from Spain to Italy over the Great St. Bernard is too preposterous to have been treated seriously by Roman generals or geographers. I have, I hope, sufficiently shown what has happened to all who endeavour to read Varro' s catalogue in any other than a geographical order. I challenge any fresh critic to provide us with a satisfactory alternative. Unless this can be done, the problem of Hannibal's Pass is not indeed solved, but we are entitled to claim that we know in what part of the chain the Romans best qualified to judge a hundred and fifty years after the second Punic War believed his Pass to have lain. ^ See Inscr. Gall. Cis. Latin., vol. v. Part 2, p. 761. The \'alley of the Ubayette looking west. Larche and the ascent to the Col de l'Arc.entiere. CHAPTER II THE DURANCE PASSES We have found ourselves led, in order to satisfy the requirements of the classical texts, to look for an Alpine route from the basin of the Durance into Piedmont, lying between the Coast-road of Roman and modern times and the Mont Genevre, and practicable for an army. There are several tracks over this part of the range that may serve in summer for foot-passengers, or small detach- ments of troops. Mr. Coolidge, who from his wide reading and many perambulations of the Western Alps is a formidable authority in any matter con- nected with them, is reluctant to allow that the Col de FArgentiere belongs to a category different to that of the many loftier passes that lead out of Val de Queyras, or the upper Ubaye valley, or of the several circuitous and less accessible gaps near the Mont Genevre. But the point of view of a stubborn pedestrian is apt to assimilate itself to that of a smuggler or a peasant rather than to that of the leader of an army. I must maintain my position that the Col de I'Argentiere was rightly spoken of and accepted by Brockedon as one of Nature's great passes, and that the rest are by- ways. Strategical considerations, religious perse- cutions, greedy nobles and prelates with their 24 HANNIBAL ONCE MORE exactions and tolls, political frontiers and custom- houses, no doubt drove from time to time small bodies of men, refugees and military or com- mercial travellers, to leave the main tracks. My contention is that, whenever these influences or hindrances were removed, traJB&c resumed its natural channels, the passes that by their low- ness and relative easiness of approach recommend themselves to the average traveller. Napoleon, holding for a time both France and Italy, set himself to reconstitute the Alpine roads of the Caesars. He decreed that a route * from Spain to Italy ' should cross the Col de 1' Argentiere. For besides the Mont Genevre there is only one fre- quented historical highway through the mountains from the basin of the Durance to that of the Po. It is that which, leaving the Durance valley where it bends north at Mont Dauphin, crosses the ridge of the Col de Vars (6939 ft.) to the head-waters of the Ubaye, and then the main chain at the Col de r Argentiere (sometimes called Col de Larche or Col della Maddalena) i (6545 ft.) to the vaUey of the Stura. In primitive times and countries tracks as a rule prefer a hiU to a defile,^ and the first pass was a means of avoiding the lower gorges of the Ubaye. These passes were weU known to the Gauls. A Keltic cemetery has recently been discovered beside the modern chapel on the top of the Col de Vars, and the objects found in it removed (I am told) to the Museum at Digne. The word Vars is still alive in local patois, and designates a tomb or ^ Larche is the highest large village on the French side. There is a chapel dedicated to .St. Mary Magdalen on the top of the pass. * The old track over the Forclaz to Chamonix is a faniihar Alpine example of this tendency. THE DURANCE PASSES 25 burial-place.^ That these passes were used by the Romans is proved by the traces of a Roman camp on the hillside above Guillestre commanding the northern approach to the Col de Vars, as well as by numerous discoveries of remains in the Val di Stura and the Ubaye district. ' Quoi qu'il en soit,' writes M. Rey in his Royaume de Cottius (p. 248), ' on ne saurait emettre des doutes apres les nombreux vestiges recueillis dans le sol par le Docteur OUivier sur 1' existence a I'epoque Romaine d'lme voie regidiere qui par le Col de la Madeleine joignait la vallee de 1' Ubaye a celle de la Stura.' In A.D. 1515 these passes enter prominently into military history. The adventures and troubles of Francis I. in getting his army and artillery over them are picturesquely recorded in a letter to his mother -written from St. Paul, a village between the passes, and in the contemporary works cited below. ^ I shall not interrupt my main argument here to tell the story. Those who wish to compare the French king's experiences with Hannibal's will find the material on a later page. It is not my purpose in this place to write a mihtary history of the passes. They were crossed by armies again and again; in 1692, 1710, and 1744. The Marquis de St. Simon has told the story of the campaign of 1744, in which he was ^ See La ValUe de Barcelonnette, by F. Amaud. Grenoble, 1900. For Vars see also Le Royaume de Cottius, by R. Rey. Grenoble, 1898 ; and Dr. 01U\ier, Une voie Romaine dans la ValUe de V Ubaye, 1889. ^ Signot Jacques : ' La totale et vraie description de tous les Passages Lieux et Destroictz par lesquelz on peut facilement passer et errer des parties de Gaules es parties d'ltalieet signamment par oii passerent Han- nibal, Jules Cesar, Charles vrn., Louis xn. et Francois. Paris, 1507. Re- printed in 1515 (see Appendix). The British Museum copy is of 1518. Pasquier le moine : ' Le Couronnement du Roy Francois,' etc. Paris, 1520. 26 HANNIBAL ONCE MORE himself concerned. Among more recent authors, Sismondi blunders sadly over the local orography (see Brockedon, Passes of the Alps). Michelet, well called by M. Anatole France, ' 1' enfant terrible de I'histoire,' so revels in the picturesque that, con- trasted to his extravagances, Livy may well seem the most chastened of historians. Turning to modern authorities, the author of Le Alpi che cingono F Italia considerate militar- mente (Turin, 1845), a work published by the Piedmontese staff, expresses himself as follows (p. 525), Vie principali che valicavano le Alpi: ' La prima che entrata nella Valle di Stura saliva il facil Passo dell' Argentiera, da dove forse erano discesi i primi Galli in Italia. Finalmente la stessa via calava della parte opposta de' monti cosi neir Ambrunese come nella Provenza dei tempi moderni.' BaU, in the Alpine Guide, writes of the Col de I'Argentiere : ' It is the lowest and most accessible pass between the Col di Tenda and Mont Genevre.' Joanne says : ' Le col que doit franchir une Route Internationale (now made) est un des plus frequentes des Alpes, des milliers de Piemontais le traversent chaque ete.' Signori Martelli and Vaccarone, in their Guida delle Alpi Occident ali, 1889, say : ' This pass is the easiest and most important of all those leading to France between the Col di Tenda and Mont Genevre ; it was known to the Romans, and their passage into Gaul by this route is proved by the inscriptions found at Bersezio and Argentiera.' ^ ^ M. Rey suggests that certain inscriptions, possibly the same, have been shown to be forgeries. But Mommsen, in Inscr. Gall. Cis. Latin., vol. V. Part 2, gives five inscriptions found at, or near, Demonte, and one from Bersezio. THE DURANCE PASSES 27 Napoleon, as I have already pointed out, recog- nised the facility and importance of this Col, and decreed the construction of a road over it with the title ' Route Imperiale de I'Espagne en Itahe.' I saw some years ago near Nismes a milestone with the inscription, ' Route de Montpellier a Coni.' I have myself crossed the Col de Vars once, and the Col de I'Argentiere twice. They are both, to borrow an adjective applied by the late Mr. Bullock Hall ^ to the Mont Genevre, very ' human ' passes. There are villages a mile or two below their summits, cornfields extend to within 500 feet of their crests. The upper part of their track winds among glades or over open pastures. What difficulties there are in the nature of the ground "^ lie in the approaches to the passes, in the gorge of the Ubaye below St. Paul, and on the Italian side in the defile of the Stura known as The Barricades. The only sterile part of the whole route between Mont Dauphin and Cuneo is the upper part of the Stura valley.^ This point, the relative fertility and adapta- bility to culture of the districts traversed by the different routes, has been too httle considered. ' If the army must carry its food through the mountains, the shorter the route through them, the smaller the number of days for which tlie ^ The Romans on the Riviera. London, 1898. A very interesting and too little known work by a sound scholar. ^ There are none, of course, on the modern roads. The Col de Vars forms part of the ' Route des Alpes ' from fivian to Nice, and is daily traversed in summer by two pubhc motor-brakes at some peril to the occupants, owing to the windings and narrowness of the track. ' ' Arida e cosparsa di roccie prive di vegetazione.' JIartelli and Vaccarone, Guida delle Alpi Occidentali di Picmonte. 28 HANNIBAL ONCE MORE supply will be needed, and the less, therefore, the burden to be carried,' ^ writes a military critic. Both the Mont Genevre and the Vars-Argentiere routes have an advantage over the more northern passes in this respect. There are villages and cornfields, cow-pastures and hay-meadows all the way ; the road runs through a country adapted for human uses. Doubtless its resources are not capable of feeding an army, and were still less so two thousand years ago, but they may well have served as a very considerable aid to the commissariat, and we are told by the historians that they did so. These two considerations, the resources of the district traversed and the length of the pass, have to be weighed and balanced in deciding the route of an army across a great mountain range. The other passes leading out of the Durance basin (other than the Mont Genevre) are more rugged, and with one exception higher. ^ It is, of course, possible to assert the claims of such by- ways as the Col de Vallante (9269 ft.), or the Col deir Agnello (9003 ft.). The latter was used by a detachment of cavalry belonging to Francis i.'s army, and is frequented by Piedmontese labourers seeking employment in Provence. The Col de la Traversette (9560 ft.), famous for its mediaeval tunnel, has also been mentioned. But no sound argument in favour of any of them is as yet forth- coming. ^ Prof. S. Wilkinson, HannibaVs March ihrowjTi the Alps, 1911. 2 The Col de I'lfichelle (5744 ft.), the lowest in the Western Alps, lying north-west of the Mont Genevre, never appears to have been crossed by anybody worth mentioning, owing perhaps to its circuitousness and inconvenience compared with the neighbouring Mont Genevre, and to the steep cliff which has given it its name. THE DURANCE PASSES 29 As I have already indicated, I am not disposed to lay too great stress on exact correspondences either in distances or in local features. Ancient and modern roads are apt to differ. An accurate measurement in a mountain country is a late result of civilisation, while rough estimates are, as a rule, exaggerated. Similar caution may, in my judgment, profitably be used in considering the more picturesque incidents in the accounts of the Alpine march. The pass to be looked for must not too exactly correspond with our texts, but should offer, even in its natural state, difficul- ties less than those described in them. The general tendency of historians in dealing with mountain campaigns is to exaggerate the difficulties of the ground ; and difficulties are always greater to an army than to a tourist. Moreover the in- vestigator must put the modem highroad out of view ; he must follow, or try to imagine, the old tracks. And in Hannibal's case, he must allow for the advanced season, the middle of October, if not later. So far as I can judge from the literature of the subject, most of the wTiters who deal with Hannibal's march have failed altogether to realise the vagaries and the frequently premature rigours of an Alpine autumn. Commandant Colin even seems to think (p. 42) that the ' first snowfall on the mountains ' may be treated as a fixed date. The record of the autumn of 1912 ought to be enough to disabuse contemporaries of any such delusion. In the upper valleys of the French Alps the first snow generally falls some time in September. But it can do so, as in 1912, in August. In the first week of November snow 30 HANNIBAL ONCE MORE fell as far down as Amberieux at the foot of the Jura, and in the environs of Grenoble, and at that date there was already ten feet of snow on the mountains. On the map published by the Tour- ing Club of France, the Argentiere is marked as one of the passes closed to wheels in November. At St. Paul I was told the average winter snowfall was 7 metres (22 feet). I shall have to return to this subject of the snowfall. It is of primary importance in any argument as to Hannibal's adventures ; and yet very few of those who have written about them have shown any clear con- ception of the climatic conditions of the Western Alps. There is really no need to drag Hannibal up to one of the rare glaciers in the Cottian Alps in order to account for his recorded adventures. Gorges and white rocks can be found on every pass, and each partisan in turn soon persuades himself that the pass he inclines to has the most suitable set. I do not pretend to rise altogether superior to this very human weakness. But I am content not to argue more from such coinci- dences than that the Col de 1' Argentiere is not put out of the running by any fatal incompatibility between the classical texts and its physical features and conditions. CHAPTER III THE VARS-ARGENTlfiRE ROUTE I MUST now ask the reader who has got thus far to join me in an attempt to go over the ground with the classical texts in our hands, and to ascertain what are the points in which they wiU support, and where they may weaken, our identification of the Vars-Argentiere route as Hannibal's. For the present I shall assume that my readers are prepared to accept the generally received in- terpretation of Polybius and Livy and to recognise as the Island the district between the Hhone and the Isere. Those who are disposed to contest this view, I must refer to the final chapter. From the Confluence near Valence we have to find a fairly easy route up to the scene of the first battle. It ought, if we read the text of Polybius in the usual way, to be about eight hmidred stades in distance from the Confluence. Let us take up Livy. According to his account, after settling the local disputes in the Island, Hannibal turned to the left from the Confluence, leaving the direct road to Italy on his right. One direct road to Italy, in Livy's time a well-known Roman highroad, is clear on the map. It leads from the Rhone up the vaUey of the Drome, past Die and over the Col de Cabre. If he marched SI 32 HANNIBAL ONCE MORE up the Isere, Hannibal faced, as Polybius says, eastward and turned to the left.^ He would soon reach the junction of the Drac with the Isere, and find a track up the valley of the former which would lead him towards the more southern Alpine passes. It is a track which has been used, as I shaU show later, by modern armies bent on the invasion of Italy. The nature of this region, the basin of the Drac, has been singularly misrepresented by several of my predecessors in this discussion. I have been through it myself more than once. It is by no means barren. It consists of a series of broad, cultivated slopes and platforms, cut across by the trenches of the torrents which flow out of the snowy heights of Dauphin e. This is the descrip- tion given by Joanne of the country between Le Mure and Corps : ' Cette contree est appelee Le Beaumont. En effet il est peu de pays de mon- tagne plus beau. Les ruisseaux sont hordes de frenes et de saules, les hauteurs sont couronnees de chenes, les productions les plus varices s'y succedent : vignobles, vergers, pres, jardins, chenevieres, bouquets d'arbres, maisons encadrees de feuillage, en attestent la fertilite.' ^ Livy goes on to give the names of three tribes Hannibal passed through; the Tricastini, Vocontii, and Tricorii. The Tricastini lay, according to most com- mentators and the atlases, far south, near St. Paul-Trois-Chateaux. What authority there is ^ I see no reason to suppose that Hannibal marched up the left bank of the Isere. * Joanne, Jura et Alpes Frangaises, 1877, p. 726. THE VARS-ARGENTIERE ROUTE 33 for this location of them, except a possible indica- tion in the name of the town, I have not yet been able to find. Mr. Long, in Smith's Dictionary of Ancient Geography, places them between the Drome and Isere, following in this Ptolemy, who located them next to the Allobroges and east of the Segalauni on the banks of the Isere. In the notes to the fine edition of Ptolemy (Paris, 1887), edited by Herr MiiUer, this position for the tribe is asserted, and the causes by which many writers had been led to place them further south are fully explained. It may be noted that the old Swiss geographer Simler places them in the Trieves district between Grenoble and Gap.^ Now let us see about ' the furthest parts of the territory of the Vocontii.' Here we have no difficulty. Joanne fixes Vizille on the Drac as their frontier, and it would be their further frontier as opposed to their nearer frontier on the road up the Drome. ' Vizille (Vigilia, Castra Vigihae) . . . surveillait les belliqueux Voconces etablis sur la rive gauche du Drac' ^ Following up the same vaUey, we find in St. Bonnet the chief town of the Tricorii, ' ville principale des Tricoriens a I'epoque gauloise.' ^ They are supposed to have been clients of the Allobroges.* We have thus found for Hannibal a road to the foot of the Alps fairly satisfying the require- ments of both narratives ; for the eight hundred ^ See Coolidge, J. Simler et les origines de VAlpinisme, j). 199. See also Mr. T. Rice Holmes in Caesar's Conquest of Gatil (Oxford University Press, 1912), on the Allobroges and Vocontii. 2 Alpes Franqaises, p. 72U. ^ Alpes Franchises, p. 729. * See Rice Holmes, of. cit. C 34 HANNIBAL ONCE MORE stades from the confluence near Valence would have brought Hannibal to a point near Corps, to the foot of the first pass, that between the Drac and the Durance.^ It is a broad- topped ridge, some- what steep on both sides, and now well cultivated. In the basin to its south lies the town of Gap, the Gallo-Roman Vapincum. On this ridge, the Col Bayard, we may place without any straining of the narrative the incidents of the first battle, fought according to Polybius with Allobroges, but according to Livy with unnamed ' barbarians.' But here we encounter a difficulty which has been made a great deal of : if the Allobroges dwelt, as Livy says, in the neighbourhood of the Island, what business had any of them so far south ? At a later period their centre was Vienne.^ On this point, I may quote Mr. Marindin : ' It cannot be assumed that Celtic tribes had precisely the same limits of territory in 218 B.C. and in the time of Julius Caesar." The chief seat of the Allo- broges in the days of Polybius may well have been in the district south of the Isere. According to his narrative it extended over plain and moun- tain for a distance of several days' march. Since in later times, however, their chief town was Vienne, Livy may very easily have been led to write according to the location of the tribe in his own ^ Some writers have, as I have ah-eady pointed out, attempted to distinguish between the Ascent of the Alps of c. 50 and the Ascent of the Alps leading to Italy of c. 39, and allege that it is permissible to count 800 stades from the Passage (Cohn, p. 339). This reading seems to me contrary both to the letter and spirit of Polybius, who looked on the Alps as a whole. Compare otto roiv Kara rbv ''Pobavov roiroiv (ve^aXfv (Is 'IraXiav, of ch. 47 and ch. 56, where the fifteen days clearly cover what has passed from the beginning of the''AX7recov ava^oXr) of ch. 49. 2 Ptolemy, Geo. 11. 10. Commandant Cohn has laid down their limits as extending southwards of Grenoble for some ten miles. THE VARS-ARGENTIERE ROUTE 35 day. I am unable to agree with Commandant Colin that here are * two blunders, two enormous blunders which spoil the whole picture.' ^ We must remember that the term Allobroges was used very widely in later Roman times ; minor tribes accepted their hegemony, and the people thus designated played a leading part in the later history of Gaul.^ I may, perhaps, suggest a parallel case in recent history which may possibly help my readers to take a more charitable view of Livy's short- comings. I have sometimes amused myself by imagining the difficulties and confusion writers two thousand years hence may fall into should they attempt to unravel the allusions to the tribes of the Caucasus in English books and newspapers of the time of the Crimean War. The term ' Circassian ' was used habitually with the most reckless inaccuracy. Schamyl, a native of Daghestan, was spoken of as a Circassian. The true seat of the tribe is two hundred miles further west mider Elbruz and along the north-west flanks of the chain. It is separated by the land of the Ossetes and the Darial Pass from the eastern highlands and their denizens. I have found an acquaintance with the Caucasus, as it was forty- four years ago, and in some respects is still, some help in my attempt to realise the state ^ See also pp. 72-4. Livy calls the quarrelsome tribe Allobroges, the guides supplied to Hannibal Gauls ; the friendly people of the low country — in Polybius Allobroges — are to him also Gaids ; the hostile tribe are distinguished as ' mountaineers,' but he adds that there was ' httle difference in language and customs ' between them and the Galhc guides. Polybius calls the hostile tribe Allobroges and barbarians alternately. "^ The anonymous author of HannibaVs Passage of the Alps (London : Whittaker and Co., 1830) aptly quotes Strabo: 'AXXo/S/joyej fivpiaat TToXKaii irpoTfpov fxiv farpaTevov, vvv fie yewpyovai to, nfdia koI tovs 36 HANNIBAL ONCE MORE of the Alps in the time of the Caesars, if not in Hannibal's. The same battle is described by Livy — the incidents are obviously identical — but placed beyond the Durance. On this point all attempts to establish a complete accordance between the two historians seem to me to break dowai. There is a discrepancy between Polybius and Livy wliich can only be accounted for by some failure of the later writer in collating his authorities. I, at any rate, have no better explanation to offer. In this particular I feel obliged to follow Polybius and reject Livy. Having revictualled his forces in the neighbour- hood of Gap, Hannibal set out on his further march into and up the Durance valley. The traveller who now takes this direction, either by road or rail, soon leaves behind the pleasant farms and meadows of the Gap basin, and enters a valley where an impetuous stream runs riot in a wide stony bed, where each tributary torrent flows out of the hills on the top of an enormous fan of gritty soil, torn from the gashed and tottering flanks of the gaunt mountains overhead. It is an abomina- tion of desolation which will remind the eastern traveller of some remote region, such as the vale of the Araxes below Julfa, rather than of any scenery he is familiar with in the Alps. The land- scape before his eyes answers in every detail to the description of Livy. How some writers can venture to dispute this correspondence, I am, I confess, at a loss to understand.^ ^ M. Chappuis, a native of the French Alps, here entirely agrees with me. See his tracts referred to in the Appendix. THE VARS-ARGENTIERE ROUTE 37 Beyond Embrun the scenery becomes less for- bidding, and at Mont Dauphin the Durance valley bends sharply north. Before reaching the angle Hannibal's guides made him cross the river, which would not have been needful had he been making for the ]\Iont Genevre. From this point there is no difficulty in fitting the texts with the ground up to the second battle. We can imagine the Carthaginians turning their backs on the eternal snows of Dauphine and tramp- ing along the projecting crest of the hog's-back, which leads up to a broad recess in the hills now occupied by villages and cornfields. The track next mounts over gentle wooded slopes to the green ridge of the Col de Vars, and then descends transversely across pasturages to a stream some way above St. Paul, at the spot where an upper hamlet now stands. The local topography will be better understood by the aid of a diagram. Below the junction of the water from the Col de Vars and the Ubaye, the united streams flow through a narrow ravine, the Pas de la Reyssole, only made practicable for wheels since 1849. The old mule path, after crossing the Vars torrent at the upper hamlet, traversed on a high shelf for at least half a mile a hillside, precipitous in places and exposed in others to the frequent fall of stones from lofty cliffs of exceptionally loose and crumbling rock. The track is alternately carried on galleries, or cut through spurs : a ruined tower defends its lower end. It would be impossible to find a defile more exactly corresponding with Livy's description of the scene of the second engagement. Professor Wilkin- son has stated that, from a military point of view, 38 HANNIBAL ONCE MORE he considers it one of the worst defiles he knows. I would add a detail (not, I think, hitherto noticed) suggested to my mind by Livy's text ; his narrative appears to imply that this action was fought during a descent, which would be the case here. The spot, a natural gateway, seems to me a likely one for the tribes of the Ubaye to choose for their stand ; they may well have taken advan- tage of this favourable opportunity to resist the invader. It is about fifty miles from Gap to the Pas de la Reyssole, and fourteen on to the top of the Col de I'Argentiere. Distances, as I have already noted, are complicated by the difference between old paths and new roads — new roads in the moun- tains are as a rule longer — and are not to be too closely compared, but these will fit well enough with the days allowed by Polybius and Livy. Nor is there any difficulty in making the total distances assigned for the march from the passage of the Rhone to the Italian plain fit in with the route we are following. The next important point in the texts is the account of what took place on or near the top of the pass over the Alpine watershed. The summit was broad enough for an army to camp on, and not too lofty for it to halt for two days in late autumn. Polybius and Livy both give us, in terms so nearly identical as to show that they either drew from a common source, or that the later historian copied his predecessor, a report of a speech made by Hannibal to his troops during their halt. Polybius abounds in reports of speeches, which imply either an able reporter on THE VARS-ARGENTIERE ROUTE 39 the spot, or an artistic appreciation of the situa- tion in its human aspect on the part of the historian. We have in these few chapters deahng with the Passage of the Alps, not only the speech on the mountain, but the harangue of the Gallic chieftains at the passage of the Rhone, Hannibal's consequent address of encouragement to his troops, and Hannibal's and Scipio's speeches to their armies on the Ticino. That we have in any of these cases a literal reproduction, or anything like a precise record of what was said, is out of the question. Discussions as to the force of the Greek prepositions employed ; whether Hannibal ' pointed out ' the plains of Piedmont, but only ' indicated the direction ' of Rome, would strike me, therefore, as somewhat trivial, even did not a competent critic like ]\Ir. Marindin deny that there is any scholarly basis for most of them. It may be convenient to bring back to the minds of my readers the passages in question. Polybius must come first : ' Hannibal, seeing his men discouraged both on account of the hardships already undergone and those they anticipated, called them together and strove to cheer them up, having a single ground for this, the evidence ^ (ivapyeia) of Italy. For it lies so close mider the aforesaid mountains that to those who contemplate both in one view the Alps appear to bear the relation of an Akropolis to the whole of Italy. Wherefore, ^ See Law, The Alps of Hannibal (London, 1866), vol. i. p. 254, for grounds for this translation of the Greek word when used by Polj'bius, He cites many other passages to the point. Commandant Colin supplies his readers vdih a French translation of the whole passage somewhat coloured by his own convictions. 40 HANNIBAL ONCE MORE pointing out to his men the plains about the Po, and reminding them of the general friendliness of the Gauls who inhabited them, and at the same time indicating the position of Rome itself, he to a certain extent raised their spirits.' So far Polybius. Let us now take up Livy : ' Hannibal, having advanced the standards to a promontory whence there was a view far and wide, and ordered the soldiers to halt, points out Italy and the plains about the Po lying under the Alpine range (telhng them), '' they were sur- mounting not only the walls of Italy, but those of Rome itself, that the rest would all be level or downhill, that by one or at the most two battles they would become the possessors and masters of the citadel and head of Italy." ' Now to my mind all that these passages convey is that Hannibal made a picturesque speech suitable to the occasion. The phrase Livy employs, ' the rest would be level or downhill,' suggests to me a sufficient key to all that follows. To any born leader of men, the downward slope of path and stream, the warmer glow over an Italian horizon, would hint the occasion for an appeal to the imagination of his followers. I cannot bring myself into the frame of mind of those who main- tain that each reported phrase must be taken ' at the foot of the letter,' and that an extensive view over Piedmont is ' the crucial requirement ' for Hannibal's Pass. If I am wrong here, at least I am wrong in very good company. For not one of the great passes of the Western Alps enjoys such a view. The supporters of the Little St. Bernard, the Mont Cenis, and the Mont Genevre are there- The Col nE l'Argentiere lookim; towards Italy. The Lac de la .Madelalne looklnc, west. THE VARS-ARGENTIERE ROUTE 41 fore all sharers in my error. It has been left to some French officers to put a literal interpretation on the classical texts, to push it to its logical consequence, and to have the good fortune to convert an Alpine expert in Dauphine, M. Ferrand, and a distinguished military critic in our own country, Professor Spenser Wilkinson. On a subsequent page I shall pursue these bold adventurers in their search among the loftier, more rugged, and obscure byways of the Alps for a pass which comphes with their two ' crucial requirements ' : a plain on the top large enough for an army to camp on, and a view extending over Piedmont and suggesting the road to Rome. But for the moment we will return to the Col de r Argentiere. That pass, like most of the other great passes into France of the Western Alps, offers on, or near, its summit a wide extent of available camp- ing ground. But it commands no view whatever of the Piedmontese plain ; the eyes range over mountain spurs until they are caught by the highest crests of the Maritime Alps some thirty- five miles off. There are sunny distances, there is a warmer colour in the atmosphere, but these are the only signs that the traveller stands in one of the Gates of Italy. If the speech reported is to be taken literally, Hannibal and his host cannot have passed this way. But must it be so taken ? The difficulties on the descent have now to be dealt with, and here the statements in the texts will require from us unusually careful considera- tion. The majority of critics, whether from lack of mountain experience, or of knowledge of the climate of the Western Alps, have fallen into 42 HANNIBAL ONCE MORE strange confusion in the discussion of details which suggest httle or no difficulty to the Alpine expert. Snow, says Polybius, had already fallen on the summits, and in the first descent the track proved steep and slippery. After a time a point was reached where it had been carried away for some three hundred yards. An attempt was made to get round the gap by a higher route, involving the crossing of a snowfield or avalanche bed. This, however, had to be abandoned : ' for what happened,' explains Polybius, ' was singular and unexpected.' On the top of the old snow that remained from the previous winter was a layer of new snow, which being fresh, still soft, and having no depth, was easily penetrated. So that the men, when they trod through it on to the harder snow below, made no impression, but slipped with both feet at once. The animals, however, broke into the lower snow, and many of them stuck as if trapped. Consequently the attempt was abandoned and men were set to work to repair the proper path. Livy gives in the main the same account, some sentences being practically common to both authors. But he adds that in the attempted descent along the path the soldiers at first helped themselves by holding on to stumps and bushes. When, however, an attempt was made to find a higher and more circuitous track over the snov»', they found no such aid at hand. He tells us further that in order to improve the old path ' huge trees ' were cut down, a bonfire lighted, and the heated rocks split by pouring on them acetum. The rock was then broken up with tools, and the THE VARS-ARGENTIERE ROUTE 43 gradient lessened by moderate zigzags, so that the elephants as well as the baggage could pass. During the halt round the cliff the transport animals were on short commons, for the summits were bare, and snow covered whatever pasture there might have been earlier in the season. The Alpine traveller is struck at once on examin- ing in detail and comparing the two texts by the extremely accurate and graphic account here given of what happens when men and animals have to cross new snow lying on old. I have often found such a state of things ; in 1881 on the Col du Bonhomme the effects described were exactly reproduced, the first snow having fallen on 15th August, the second a fortnight later. ^ There is not in my opinion the least excuse for dragging in, as some writers do, a glacier. The surface of the lower end of a glacier is too hard to be broken into by traffic ; the snow of an avalanche bed, on the other hand, lends itself to the inconveniences described. One set of ques- tions, however, may fairly be asked. Could snow, avalanche or other, from the previous winter or spring have lain through a whole summer below or near the tree-level in the Maritime Alps ? Again, could a snowfall, which had lain long enough to consolidate, happen in the autumn at a level so low that ' big trees ' were close at hand ; or could an avalanche have fallen as early as late October ? My own former and very recent experiences lead me to believe that there is no impossibility in either occurrence. ' Singular and unlooked for,' * It is worth noting that some tourists I met on the pass believed, the lower snow to be that of the previous winter. 44 HANNIBAL ONCE MORE Polybius calls the incident, and the phrase seems to me to meet the occasion. As I have already pointed out, in bad seasons, such as 1912, snow falls in the Alps much earlier and more often than most men of letters, or even summer tourists, realise. In 1912 snow fell down to Brian9on in the last days of August. Early in September six inches lay on the Little St. Bernard. Twice in September it fell and lay for days down to 5500 feet on the hills of Como. In the middle of that month, in another part of the Alps I saw fresh avalanches lying at 4000-5000 feet. In the first week of November six feet of snow lay on the mountains round Grenoble, and it fell as far down as the Gresivaudan, and at Amberieux at the foot of the Jura on the Lyons- Geneva Railway. I do not therefore think it improbable that the difficulties described might be met with below the tree-level on any of the passes of the Western Alps, even without the in- tervention of an avalanche, at the end of October. The description given by St. Simon, one of the leaders engaged, of the retreat through the Barricades in Val di Stura of the French troops on 15th November 1744, may be apposite here. It should be contrasted with the two following state- ments, which have recently been put forward. ' The sole difficulty of the defile of the Barricades in war is that which has sometimes been caused by an enemy's resistance. . . . There is on the descent from the Col de Larche (the Col de I'Argentiere) no possibility of men or mules falling over a precipice.' So writes Professor Wilkinson. But this is St. Simon's story of his own experience : THE VARS-ARGENTIERE ROUTE 45 ' II n'est pas possible de former une idee de ce que I'armee eut a souffrir vers la fin du jour lorsqu'elle passa le defile des Barricades. Le vent le plus froid jetait une poussiere fine melee d'un gresil dur et piquant qui s'attachait a tout ce qu'il rencontrait et s'y convertissait en un moment en verglas ; le chemin devenait comme ime glace sur laquelle ni les hommes ni les chevaux pouvaient plus se soutenir, plus de cent soldats perirent du froid et de ce vent qu'il est mortel de respirer. Le Camp occupa une petite plaine couverte de neige ou il n'y avait ni arbres, ni asile, ni ressource. Les soldats chercherent d'abord quelque paille ou fourrage dans les environs, mais ne trouvant absolument rien,' etc. ; and again, ' Les equipages et les hommes precipites roulaient des mon- tagnes.' ^ Some of my readers may suggest that the Vars- Argentiere route, on account of its more southern course, is likely to be the least exposed to such premature obstruction. Meteorological records do not confirm this objection. The physical geo- grapher will point out that the Maritime Alps, being the first heights to catch the moisture-laden sea- winds, enjoy an exceptionally heavy snoAvfall. Beaumont in his Travels (a.d. 1795) tells us that the snowfall on the Maritime Alps is heavier than on the Mont Cenis. He mentions, as his own ^ The following further facts as to the past autumn are quoted from La Montagne, the official pubUcation of the Club Alpin Fran9ai3 for this year. In the last days of August and first of September snow fell down to 5000 feet in the Tarentaise, and the cattle had to descend from the Alps. In October two shght snowfalls early in the month were followed on the 21st by a storm which forced the cattle to be brought from the fields into their stables. At the end of the month snow was lying down to 4600 feet, and there were many avalanche.«. - 46 HANNIBAL ONCE MORE experience, a fall of ' full three feet ' on 7th August on the Col di Tenda ; on 22nd September he saw a soldier frozen to death on the same pass. On 9th June 1769 six inches of snow fell in the town of Demonte (2553 ft.) in Val di Stura.i i^ the third week of June 1883 I saw plenty of avalanche snow lying in the gullies below Pietra Porzio in the same valley. On another occasion I saw snow falling among the chestnut woods of the Maritime Alps on 24th- 25th September, and Ij^ing heavily on all the ridges and slopes above 5500 feet. And if anything, the climate was moister 2000 years ago than now, for the destruction of forests, in the French Alps, must have diminished deposition. If x^^f^^^ could in this case legitimately be taken to mean ' storm,' I should see no difficulty whatever. If, however, the more general inter- pretation ' winter ' must prevail, and I do not overlook the fact that a little further on Polybius uses the word in this sense, then, unless a mistake as to the date of the lower snow was made by the chroniclers with Hannibal — which is by no means improbable — the occurrence described was un- doubtedly of the exceptional character attributed to it by the Greek writer. But I repeat there is nothing impossible in the story as told us by the classical historians.^ So much for the climatic conditions. We must ^ See a curious and entertaining volume : ' A true and particular account of the most surprising Preservation and happy DeUverance of three women who were buried thirty-seven days in the ruins of a stable by a heavy fall of snow from the mountains at the village of Bergemoletto in Italy, with curious remarks by Ignazio Somis, Professor of Physic in the University of Turin,' etc. London, 176S. 2 It may be worth noticing that in the upper Val di Stura, both THE VARS-ARGENTIERE ROUTE 47 now consider how far the nature of the ground in the upper Val di Stura will meet the requirements of the texts. From the lake on the summit of the Col de I'Argentiere there is a steep descent of a thousand feet to the upper valley. This is for a time a narrow and arid, but fairly open glen. After about four miles, the traveller reaches a gorge where the old mule path and the modern road have been carried on galleries, or cut out of the cliff that overhangs the torrent on its right bank. High up on the hillside, an alternative path follows a contour line, and crossing the low projecting spur which divides the main valley from a tributary glen, descends again to the Stura below the gorge, the bottom of which lies between 4000 and 5000 feet. This is the defile known as ' The Barricades,' famous in military history in the sixteenth and succeeding centuries for the passages of the French under Francis i., and of other armies. I shall treat of these historical incidents on a later page.^ I am not prepared to affirm unhesitatingly, as I was able to do with regard to the Pas de la Reys- sole, that the Barricades correspond exactly to the classical descriptions. The snow conditions are capable of explanation, but they were, if correctly described, unusual in such a locality ; the upper path which gets round, or over, the gorge is a good deal easier than one would expect to find it. Yet it seems to me reasonable to allow for a certain near and below the Barricades, a number of grooves in the steep hill- sides are each called on the government map Cialancia with distinguish- ing adjectives. Cialancia on the Italian, calanche on the French side, are patois terms for avalanche. ^ See Appendix i. 48 HANNIBAL ONCE MORE exaggeration and looseness, and perhaps confusion in detail, in the classical narratives, and I claim that a very moderate allowance (much less than one has to make for many modern historians when they deal with questions of orography, or mountain craft) is needed in this case to reconcile the texts with the localities. The primitive paths, it seems to me, may very well have furnished material for such a narrative. When we turn to the story told by Brockedon in his Passes of the Alps of his experience at the same spot in the last century before the carriage road was made, we are able to realise the vast discrepancy between the sensa- tions of the modern traveller conveyed along an admirable road and those of the foot or horse-man of former centuries. Brockedon' s plate, here reproduced, of * A Defile near Sambuco,' shows the impression made on a hardened Alpine tourist ninety years ago by the old mule path in a relatively open part of the valley. We have rocks, a steep and narrow track with an abyss on one side of it. This view is taken near the spot, some miles below the Barricades, where the government map shows the series of avalanche channels previously mentioned. Finally, I am not surprised that forage should be scarce in the upper glen of the Stura in late October. It is the only part of the Vars-Argentiere route which does not lie across fields, woods, or pastures.^ The change — from Alpine sternness to Italian luxuriance — in the character of the landscape in Val di Stura which occurs below the mouth of 1 See the Marquis de St. Simon's narrative quoted above as to the lack of forage in the valley above the Barricades. THE VARS-ARGENTIERE ROUTE 49 the glen leading to the Baths of Vinadio is immediate and very striking. The valley widens, the slopes become gentler, we recognise the vales and sunny hillsides, the woods and streams, the cultivable region, of Livy's picturesque descrip- tion. But the same change is found on almost every descent into Italy, and no inference in favour of any particular pass can fairly be drawn from the sentence referred to. I have now conducted the Carthaginian army to the edge of the Piedmontese plain. Here I have an important point to take, to which I desire to call the attention of students and antiquarians. Great stress has been laid by almost every writer on the statement that Hannibal, when he came out of the Alps, marched through the Taurini ' who lived close to the mountain skirts,' and took their chief town. But for some reason which I fail to understand the commentators have first refused to consider what were the probable limits of ' the Taurini living close to the mountam skirts,' through whose territory Hannibal marched, and next they have unhesitatingly identified their chief town with the Augusta Taurinorum of later date, founded B.C. 24 by the Romans on the site now occupied by modern Turin. They have pro- ceeded to rely on this identification as a principal argument to prove that Hannibal came out of the mountains by the valley of the Dora Riparia ; in other words, that he crossed either the Mont Genevre or the Mont Cenis (or one of the passes from the Arc valley parallel with the I\Iont Cenis). Now the territory of the Taurini, as defined by Ptolemy, and recognised by Mommsen, was no 50 HANNIBAL ONCE MORE narrow plot at the mouth of the Susa valley, but the long expanse of plain lying under the Alps and reaching from below Ivrea on the north, down to and beyond the Tanaro on the south. It extended both south and east of the Stura, right into the corner where Alps and Apennine meet. Hannibal, if after resting his troops in the fertile basin of Demonte he came out of the hills near Borgo San Dalmazzo, came out into the country of the Taurini. A further consideration which seems to me both obvious and weighty has been either overlooked or put aside. How, had the Taurini been a tribe of a single canton, would they have dared to face, and that on the plain, an army of at least 25,000 men — and elephants ? Nor, as Mr. Marindin supposes, would Hannibal have had to cross any great river in marching northwards to get round the hills of Montferrat. The southern tributaries of the Po flow out of the Alps in a number of streams which, owing to the scarcity of glaciers on the Cottian Alps, are of relatively small importance. Mr. Marindin further objects ^ that the Argentiere route takes Hannibal too far south. It is true that it is, if ' the neigh- bourhood of Turin ' be taken as the end of the Alpine march (and Hannibal probably went thus far north to avoid the hills of Montferrat), longer in distance by 35 miles than the Mont Genevre (120 to 85 miles). But should we not also con- sider the relative length of the portion of the roads in the mountains. On this point I may cite Professor Wilkinson : ' An army on the march always loses men from fatigue. The best way * Mr. Strachan-Davidson endorses this objection as his main one. THE VARS-ARGENTIERE ROUTE 51 to keep down losses from this cause would be to make the fatiguing marches as few as possible ; in other words, to choose the shortest way through the mountain region.' ^ From Mont Dauphin to Demonte in the Stura vaUey is fifty-four miles, to Avighana on the Dora Riparia seventy miles. Further, I make bold to say that no sufficient grounds have yet been showai for the identification of Turin as the chief toAvn of the Taurini, and that the argument founded on it must consecj^uently fall to the ground. A learned antiquarian, jM. Promis (a fellow-worker of Mommsen in his Corpus) has written at length a history of Turin. ^ An ardent and patriotic citizen, he assumes, as others have, the identity he cannot prove between the tribal centre and the Roman military town. He fails altogether, however, to bridge the gap be- tween Hamiibal and JuUus Caesar. Another high authority, Nissen, teUs us that no traces of a pre- Roman town have been found on the site. It is, in fact, indefensible by nature on two sides, the south and west, and therefore thorouglily unsuited for the stronghold of a warlike tribe. Its foundation was probably due, as Livy suggests, to the facihties for water transport given by the Po, and to its position near the exit of the Roman roads leading to Gaul opened by Pompey and Julius Caesar. It watched the central passes, as Aosta, founded about the same date and for the same purpose, ^ A large-scale physical map such as the Italian Surrey (1,200,000) is needed to enable the reader to see the reason why an army in primitive times would prefer this circuit to a march through the maze of hiUs south-east of Turin. 2 Storia delV Aiitico Torino. See also Xisscn, Historische Landes- Jcunde. 52 HANNIBAL ONCE MORE watched the more northern ones. Some town Uke Cavour, lying close to the foothills, and described by Promis as a natural fortress raised four hundred feet above the plain, is far more likely to have served as the city of refuge of the Taurini. In the territory of the Taurini we reach the end of the Alpine march. Looking back on the texts, I recognise that there are difficulties and discrepancies in the narratives both of Polybius and Livy ; for example, the length of the march ' along the river ' (Polybius), the dis- crepancy as to the position and action of the Allo- broges, the location of the first battle with regard to the Durance (Livy), the panoramic view from the pass (both), the old snow and the big trees (Livy), any explanations of which on any route are open to criticism and question. On the other hand, I accept Livy's repeated statements that where his authorities differed, he did his best to collate and select between the various versions.^ If he has once and again failed in this task, I am not prepared on that account to throw over his narrative as a whole. And when I am called on to choose between discrediting his definite state- ment as to the river-valley by which the Car- thaginians approached the Alps, or treating the account of the view and the speech on the pass common to him and Polybius as a picturesque embellishment, I prefer the latter alternative. I recognise that the difficulties of Alpine defiles and Alpine paths are likely to have been greater two thousand years ago than they are now ; nor can ^ 'Certe variat memoria actae rei,' — Livy, lib. xxi. c. 28. 5. See also c. 38. THE VARS-ARGENTIERE ROUTE 53 I close my eyes to the fact that any physical obstacles encountered by armies are apt to gain rather than lose in the hands of historians no matter of what century. On the whole, the Vars-Argentiere route seems to me, as far as the nature of the ground goes, to satisfy the requirements of the classical nar- ratives at least as well as any other pass.^ But on its physical features alone, I might hesitate to prefer it definitely to the Mont Genevre. The decisive consideration to my mind is that the evidence for the IMont Genevre being Pompey's Pass is overwhelming. Again, I cannot believe that the tradition in Caesar's time, the tradition accepted by Varro and Pompey, that Hannibal crossed a different pass from Pompey, was false. Finally, I claim to have shown that the territory of the Taurini included the whole western end of the plain of Piedmont from near Lanzo doA\Ti to the foot of the Apennines, and that there is no solid ground for identifying Turm as their chief town. In so doing, I have swept away one of the main arguments both for the Mont Genevre and the Mont Cenis group of passes. It now remains for me to advert as briefly as may be to the considerations which seem to me decisive against the more northern passes, the Little St. Bernard, and those leading from the Arc valley to the Dora Riparia, the Mont Cenis and its substitutes. In the case of the Little St. ^ The question how to reconcile the statements that Hannibal descended first into the territory of the Taurini, and that he came down boldly on the Insubres, has, I hold, been met satisfactorily by the suggestion that the Taurini incident was treated in this passage by Polybius as part of the Alpine march. 64 HANNIBAL ONCE MORE Bernard, the most decisive of these considerations (to those who loiow the gorges of Val d'Aosta between Chatillon and Ivrea) is the short time allowed in the narratives for the descent from the broken road close to the pass to the plain of Piedmont, and the absence of any mention of difficulties in this descent. The failure to recognise the fact, which Napoleon learned by experience, that the lower part of the Val d'Aosta is a pro- longed defile dangerous to armies, and that the descent to the plain of Piedmont by this route is double in distance what it is by way of the more southern passes, is a fatal flaw in the case for the St. Bernards. Mommsen, no doubt, has led astray many by his adoption of the Little St. Bernard. He begged the question in a manner not rare among learned historians when they come to deal with topo- graphical questions which lie outside their main interests. He jumped to a conclusion, and then invented premisses for it. The Little St. Bernard is very far from being what he calls it, ' the easiest Alpine pass.' What proof is there that it was ' from earliest times the great military route from the Celtic to the Italian territory ' ? We learn from Chanoine Berard that the great historian in 1873, after the publication of his history, made an excursion in Val d'Aosta in order to study its Roman antiquities. Did M. Berard convert his guest to his own opinion ? I will quote it : ' Rien ne prouve que le general Carthaginois ait traverse les Alpes Graies. Son passage par la Vallee d' Aoste ne ferait pas honneur a son talent militaire. En effet, etant a Ivree, pourquoi se serait-il dirige THE VARS-ARGENTIERE ROUTE 55 sur Turin, sachant que le Consul Scipion etait a Plaisance.' ^ Nor, despite the efforts of the commentators, can the country on the Savoy side of the Little St. Bernard be brought into any satisfactory corre- spondence with the texts. Those who lead Hamiibal over the Mont du Chat, plunge both him and themselves into additional and super- fluous difficulties. It would be cruel once more to pick to pieces IMr. Bosworth Smith's fanciful suggestions as to glaciers and elephants' bones. The key to the pass on the west lies, not where a White Rock has been found at the foot of the final climb, but in the gorges about ^Moutiers. As Professor Bonney writes : ' The Rocher Blanc below St. Germain is an insignificant thing, and an army would have gone up the opposite bank by way of the present mule road without much difficulty.' 2 While dealing with the Little St. Bernard, I must allow myself one mmor comment, though it is on a detail which has no direct bearing on the present controversy. I am at a perpetual loss to conceive how critics of the highest authority, such as Dr. Diibi and Mr. Strachan- Davidson, can continue to lay stress on, and perplex themselves over, the ' Cremonis jugum ' of Livy xxi. 38. Commandant Cohn also is ready to accept it, and to furnish a risky transla- tion: 'Cremonis jugum,' he writes, ' c'est-a-dire, sans doute, Le Col du Precipice ' ! The reading ^ Antiquites Romaines, etc., dans la ValUe (TAoste, p. 84. * Alpiyie Guide, vol. i. p. 57. The claims of the Mont Cenis group of passes have lately been brought forward with much persistence, and their consideration must be reserved for a separate chapter. 56 HANNIBAL ONCE MORE is a doubtful one — it has been disputed since the days of Heinrich Moriti, otherwise known from his birthplace, Glarus,^ as Glareanus, a celebrated scholar and geographer of the Renaissance, who in A.D. 1552 declared that several MSS. of Livy read ' Centronis ' (sic). There are three passages in the same short chapter in which the critics accept without a scruple very bold and brilliant emendations, such as ' Taurini Semigalli ' for ' Taurisne galli,' ' per Salassos montanos ' for ' per saltus montanos,' and lastly, ' Sedimi et Veragri ' for ' sed uno vel acri.' In the face of these in- stances and of the general character of the MSS. of Livy, of which IMr. Capes says, ' it would be quite hopeless to adhere even to the best MS. authority,' it is surely time that attempts to base any hypothesis on the identification of the ' Cremonis jugum ' with the Cramont near Pre St. Didier were given up. The Cramont, Gramont, or as it is spelt in the ItaHan Survey, Grammont, is a minor crest, no part of the watershed, and seven to eight miles as the crow flies east of the pass of the Little St. Bernard. It might never have been heard of had not De Saussure given a most eloquent description of the view of Mont Blanc from its summit. The name, found else- where in the Alps, is probably a corruption of ' grand mont,' the upper slopes being a cow- pasturage.^ Alpine passes are not called after ^ I owe this reference to Mr. Coolidge's Josias Simler et les origines de VAlpinisme, p. 108. Glareanus writes ' ita babent vetustiores codices.' See an article on this distinguished geographer in the Geographical Journal, vol. xxv. No. 6, 1895. ' In the Savoyard and Piedmontese Alps the primary signification both of Mont and Alp is a summer pasture. As a rule, the ' alps ' are the higher pastures, the ' monts ' the lower ones. THE VARS-ARGENTIERE ROUTE 57 relatively insignificant crests miles away from the watershed. The contrary is sometimes the case. Passes have given their names to peaks, as in the cases of the Piz Bernina and Piz Julier. The Little St. Bernard is spoken of as ' iter per Centrones ' elsewhere ; ' Centronum jugum ' would be a reasonable reading here. Anyhow ' Cremonis ' has, I fancy, a good deal more connection with Cremona than with Cramont. It is just the sort of blunder likely to be made by a copyist who had heard of the city, and to whom a little know- ledge and a familiar collocation of letters served as a snare. The modern compositor not infrequently sins after the same fashion. CHAPTER IV THE MONT CENIS AND ITS VARIANTS In the following pages I propose to consider some of the passes connectmg the Arc valley with that of the Dora Riparia. Of these, the best kno\^ai is the Mont Cenis (6893 ft.). Its claim to be Hamiibal's Pass has had several distinguished advocates both in this comitry and abroad.^ For it has served as a convenient refuge to critics, who, knowing enough of the Alps to realise the impossibility of the Little St. Bernard, have yet been unwilling to follow Livy to the Durance. They have been supported in their choice by the fact that the Mont Cenis leads into the district of the Taurini, which has been commonly assumed to be identical with ' the neighbourhood of Turin.' Some writers have suggested that the Petit Mont Cenis (7166 ft.), a short cut in distance over a slightly higher ridge, may have been the pass used : while in recent years, a small but earnest band of French officers has urged the claims of another and still loftier bjrvvav, the Col du Clapier (8172 ft.). I shall deal in an Appendix with the right of the Mont Cenis to figure as a Roman pass, and its claim to have been opened by Julius Caesar. I ^ Amongst them Ukert, Nissen, Larauza, Maissiat, Ellis, Ball, Bun- bury, and Bonney. 68 MONT CENIS AND ITS VARIANTS 59 will only pause here to point out that I am unable to endorse the suggestion made by some of the adherents of the Clapier that the steepness of the final descent on the Italian side to Novalesa is likely to have prevented the use of the Mont Cenis. It is hardly for those who favour the Col du Clapier to urge this argument. For the pre- cipitous descent below that pass is double the height (4000 ft. in place of 2000 ft.) and far steeper than any on the old track of the Mont Cenis. And if they adopt for this route the circuitous path to Exilles and the long and uneven track through the valley of the Dora thence to Susa, their argument based on the Clapier being the shortest and most direct pass falls to the ground.^ Nor can I find any evidence that in mediaeval times the Little Mont Cenis was as a rule preferred to the Great, as Commandant Colin seems to believe. On the contrary, Signer Vaccarone in the fascinating article {Boll. Club Alp. Ital. XXXV. No. 68), in which he has re- counted in picturesque detail the crossings of the Alps by the Princes of Savoy, shows that travellers from the beginning of the thirteenth century habitually passed by Lanslebourg, and therefore over the Great Mont Cenis. The role of the Petit would seem to have been — at any rate from that date — a secondary one. As to the Col du Clapier (including for the moment in this term the High- level Route to Exilles) I have failed to discover any evidence — except Commandant Colin' s alleged 1 Commandant Colin asserts that the Clapier is 30 to 35 kilometres, or • a long day's march,' shorter than the Mont Cenis. But this is not true, if his Exilles route is adopted. The distance in this case is practically the same, the height to be cUmbed 1279 feet more. 60 HANNIBAL ONCE MORE recognition of the remains of a Roman road near the top — that it was ever more than a smugglers' byway and a cattle-track. ^ What grounds Com- mandant Colin can find for his assertion that it was only after the opening of a good carriage- road over the Mont Cenis that the Clapier was abandoned I cannot discover. The first to propose this in so many respects unsuitable candidate, the Col du Clapier, was, I believe, Colonel Perrin, who in 1887, impressed by the fact that the brow above it commanded a view of the Italian plain which answered to the require- ments of the classical texts,^ put forward its claim. He has been followed by his brother- officers Lieutenant Azan and Commandant Colin. ^ The last-named, a distinguished member of the French staff, has devoted to the subject a substantial volume of 428 pages, profusely pro- vided with extracts from government surveys and original maps. He has had the good fortune to convert to his theory — as far, at least, as the pass crossed is concerned — M. Ferrand, a member of the Grenoble Bar and the author of several handsome and popular works on the French Alps. Further, he has secured as a mouthpiece for his views in this country a well-known writer, Mr. Spenser ^ It niust be noted that Commandant Colin in his preface himself throws doubt on this recognition. He speaks of *une ancienne chauss6e pavee, peut-etre une voie romaine.' 2 Marche d'Annibal des Pyrenees au Po. Paris, 1887. Of course Colonel Perrin did not ' discover ' the view. It had been very fully described by Brockedon in the first edition of Murray, and afterwards by John Ball in the Alpine Ouide. ^ Revue des Etudes Ancienries, vol. x. No. 1 ; vol. xiii. No. 3, 1911. liecherches du Col des Alpes franchi par Annibal. Casale, 1912. Lieutenant Azan, Annibal dans les Alpes. Paris, Picard, 1902. Com- mandant Colin, Annibal en Oaule. Paris, Chapelet et Cie, 1904. MONT CENIS AND ITS VARIANTS 61 Wilkinson, now Professor of Military History in the University of Oxford. I am very ready to admit that a much better case may be argued for the Mont Cenis or any of the Arc passes, than for the Little vSt. Bernard, and that in many respects their natural features can be brought to agree fairly with the require- ments of Polybius. But the fatal drawback to the route, in my mind, is that no efforts can bring it into any relation with Livy's account. Setting aside, however, for the moment this objection, I will proceed to examine the case made out for the Col du Clapier by its ablest advocate, Commandant CoUn, and the adaptation of his view recently put before English readers by Professor Wilkinson.^ Commandant Colin differs from his brother- officers in laying down two bases for his argument. He agrees with them in the belief that the 'crucial re- quirement ' for Hannibal's Pass is an extensive view over Italy. But he finds a second key to the problem in the distance, * about 1600 stades,' assigned by Polybius for the march from Emporion on the Spanish side of the PjTenees to the passage of the Rhone. This leads him to shift the passage to a point nearer Aries and the fork of the river, whereas his predecessors had been content to accept the common belief that the passage of the Rhone took place at Roquemaure, some eight miles above Avi- gnon, and had fixed the sites of th*" chief incidents of the march accordingly. It is only, therefore, when he approaches the crest of the Alps that he comes into any close agreement with his comrades. ^ HannihaVs March through the Alps, by Spenser Wilkinson. Claren- don Press, Oxford, 1911. 62 HANNIBAL ONCE MORE In order to do justice to Commandant Colin's very able argument, we must begin at the be- ginning, the passage of the Rhone. Pie claims to be able to fix this defmitely at Fourques, close to Aries and just above the point where the river finally separates into two branches before reach- ing the sea. He thereby brings the passage within 48 kilometres (about 30 miles) from the sea, and 30 miles below Roquemaure, the spot commonly selected by commentators. This is the corner-stone of his theory. It is on the strength of this — I must not caU it novel, for it had been anticipated by a local antiquarian of Aries, M. J. Gilles,^ but let us say resuscitated — hypothesis that he fixes the sites of all the sub- sequent incidents of the Alpine march. Now we find that this firm belief of Commandant CoHn's in the passage at Fourques is mainly based on one particular calculation. He alleges that the distance from the site of Emporion, Hannibal's starting-point south of the Pyrenees, and the passage of the Rhone, as given by Polybius, ' about 1600 stades,' corre- sponds exactly, or very closely, with the mileage both on the Via Domitia and by modern roads between that place and Fourques, but cannot be made to correspond with the mileage between Emporion and Roquemaure. I must confess at once that my mind is so dully constituted that, despite aU the ingenuity of the ^ ^I. Gilles, however, fixed the point of crossing a few miles liigher up at Beaiicaire, near Tarascon, where traces of the mooring quays of a Roman bridge of boats have been discovered, and then led the Cartha- ginians by the Drome to the Durance and Mont Genevre. See Hannibal et Publius Cornelius Scipion : Passage du Rhone (Paris, Thorin, 1872), followed by Mr. T. A. Cook, Old Provence, vol. i. p. 479, MONT CENIS AND ITS VARIANTS 63 critics, I still abide by the opinion I expressed many years ago,^ that in this Hannibalian con- troversy supposed exact correspondences in dis- tance may be more of a snare than a guide. But in this particular case I am prepared to argue with Commandant Colin on his own ground. For he may quite fairly plead that it is an exceptional case, since on Polybius's own statement this distance had been accurately measured along a Roman road before he A\Tote his history. Commandant Colin proceeds to point out that if Polybius in calculating the distance allowed 8J stades for each Roman mile, which is the accurate proportion of the normal Greek stade to the mile, ' about 1600 stades ' will roughly correspond to 284 kilometres or 192 Roman miles. He finds the itineraries give a distance equivalent to 268 kilometres to Fourques along the Roman road from the point — the modern toA\ii of Figuieres — at which a traveller settmg out from Emporion would strike it, and to this he adds 20 kilometres for the cross-comitry march from the camp near Emporion to the highroad at Figuieres. He thus ascertains the distance from Emporion to Fourques by the itineraries as 288 kilometres, while along modern roads it comes out at as approximately 289 to 294 kilometres. These correspondences in distance appear admirable. On the other hand, the dis- tance from Figuieres to Roquemaure proves to be at least 300 kilometres, and (with 20 kilometres ^ Alpine Journal, vol. xi. p. 271. I stated that in my opinion and in that of the late Sir E. Bunbury, ' exactness in distances has been made too much of in this discussion. An accurate measurement in a mountain country is a comparatively late product of civilisation, and old and modern roads are apt to differ in length.' Commandant Colin himself tells us that from Axles to Turin was 375 kilometres according to the Roman itineraries, but is 395 to 400 by modem roads (p. 327). 64 HANNIBAL ONCE MORE added for the cross-country bit) from Emporion 320. Thirty-six kilometres or more, insists Commandant CoHn, is too great a discrepancy to be tolerated. But let us look a little closer. Polybius bases his statement of the distance on the Roman mile- stones, and these he specifically states were at intervals of eight stades. At a later date the Greek stade was, for convenience sake, often accommodated to this extent to the Roman mile. Is it not plausible to suppose that Polybius (who Commandant Colin elsewhere, pp. 223 and 252, assures us had a natural disposition to ' simplify ') ^ anticipated this accommodation and reckoned here eight stades to the mile ? If he did this, the 1600 stades will amount to 296 kilometres. But can we be sure that Polybius made any allowance for the first bit, the cross-country march from the camp to the high-road, other than by his ' cr^eSoi/,' about ? Was he not simply calculating from the measured milestones on the highroad ? If this is a fair inter- pretation, we may add 20 kilometres to the 296 and so get a distance which fits with Roquemaure as well as Commandant Colin' s does with Fourques. I shall be told that this is all mere guesswork. So it is ; but when such guesses are plausible, it seems to me that one guess is as good as another. As I have said already, all these assumed corre- spondences in distance furnish very risky ground to build on. And the whole of Commandant 1 Polybius gives the distances from the Pillars of Hercules to the Pjrrenees as 8000 stades, from Carthagena to the Plains of the Po as 9000 stades. But according to his more detailed statements they amount to 7200 and 8400 stades respectively. Polybius, writes Mr. Strachan-Davidson {Appian, p. 133), ' shows his weakest side when he tries to calculate distances.' MONT CENIS AND ITS VARIANTS 65 Colin's and Professor Wilkinson's theory as to the march of Hannibal — not only the Passage, but the Confluence, the Island, and the battlefields — hangs in the first place on the figures quoted above ! The distance from Emporion, however, is far from being the only statement to be taken into account in fixing on the position of the passage. Polybius states that the passage was nearly four days' march from Scipio's camp on the sea. This has been located by general consent at Fos (Fossa Maritima) close to the eastern mouth of the Rhone. Polybius further says that it took Hannibal four days to march from the passage to the Confluence below the Island, a spot that has hitherto been commonly identified with the meeting of the Rhone and the Isere. Roque- maure, the point where most commentators have located the crossing-place, is about 60 miles distant from Fos ; while from the meeting of the Rhone and Isere it is distant 70 miles. These distances accord sufficiently well with the number of marches given. Recently, however, several writers besides Com- mandant Colin have followed the lead of M. Gilles in asserting that the passage must have been much lower down, somewhere between Aries and Tarascon. The arguments alleged in favour of this view may be briefly summarised as follows : (1) It meets the expression ' the single stream.' (2) The Rhone above Avignon is both too rapid to admit of an army crossing in the way described, and not wide enough to suit the details of the description. E 66 HANNIBAL ONCE MORE (3) No boats would have been found so far from the sea. (4) The Roman road from Spain to Italy sub- sequently crossed the river at this point. (5) The Roman cavalry could not have covered the ground between the battlefield and its own camp in the time stated had the Carthaginians been as far away as Roquemaure. To these arguments it may be replied : (1) That the expression ' the single stream ' in the case of a river often divided by islands and flowing in more than one channel need not necessarily be taken as used in opposition to a bifurcation at the top of its delta. (2) That between its junctions with the Ardeche and the Durance, the fall of the Rhone is less than half a metre ('49 metre) in a kilometre, or less than 1 in 2000,^ and that the width of the Rhone at Roquemaure is given by Commandant Colin as 200 to 250 metres (650 to 820 ft.), while the com- bined width of the rafts was, we are told, 300 feet, less than 100 metres. (3) Why should no boats have been forthcom- ing ? It was not seagoing craft that Hannibal wanted. The Rhone in classical and even recent times, before good roads and railroads, was one of the great routes of inland traffic with which, as Strabo pointed out, Providence had favoured Gaul. (4) Polybius speaks of the passage used by the Carthaginians in these terms, ' the passage of the Rhone for those going up the stream towards its sources.' It is not to be supposed that there was ^ See Cook, Old Provence, vol. ii. p. 230. MONT CENIS AND ITS VARIANTS G7 oaly one passage of the Rhone between the con- fluence of the Isere and the sea. (5) This rests, I beheve, on a misunderstanding of the texts, which I read thus : 1st day. — Passage of elephants, skirmish, reconnaissance and com- mencement of return of Roman cavalry. 2nd day. — Assemblage and speeches of Gallic delegates, arrival at Scipio's camp of Roman cavalry. 3rd day. — Hannibal starts north and Scipio in hot pursuit (o-7revSwv, Polybius says, translated by Mommsen ' in the utmost haste.' See also Livy). 4th, 5th, 6th days. — Scipio marches to the passage, arriving on the third day after Hamiibal had left it. I note that neither of the texts states the number of days Scipio actually spent in his hurried march : we infer them from Polybius's general statement of the distance. Let us now consider the objections to the Lower Passage. In my judgment, they are obvious and fatal. It is (by Peutinger's Tables) only 33 Roman miles (30 English miles) from Fos to Aries. ^ This distance camiot reasonably be reckoned as nearh^ four days' march for an army hurrying without its baggage (the baggage had been shipped) to over- take the foe. Commandant Colin, in his desperate attempt to make the distance from Fos to Hannibal's camp at the crossing ' perfectly accord ' with the text of Polybius, converts the Greek author's measure of distance, an average march, into a particular statement as to Scipio's rate of progress. ^ I see no reason to assume, as some writers do, in order to make out a more plausible distance for the four daj's' march, that the Roman army was obUged to make a wide detour to the east in order to get to Aries. The Fourques passage is two or three miles above Aries. 68 HANNIBAL ONCE MORE Commandant Colin seeking some supporting evidence for Scipio's sluggish progress tells us that Hannibal marched up the flat Rhone valley at only nine miles a day. But he omits to warn his readers that this estimate depends entirely on his view as to the place of the passage being adopted. He may be reminded of his OAvn maxim, ' C'est un procede bien vicieux que d'asseoir une hypothese sur une autre.' According to the more generally accepted belief Hannibal marched 18 English miles a day, the ordinary rate of a Roman army, up the flat valley from Roquemaure to Valence ; and Commandant Colin has himself allowed Hannibal, when he gets into the mountains beyond Grenoble, to do 13 to 16 miles a day, and when he approaches the pass, the pace quickens to 17 miles a day (pp. 389-424). The Lower Passage is about 100 miles from the confluence of the Rhone and the Isere, a distance too great for Hannibal's host to have accomplished in four days.^ This objection has necessarily forced itself upon the upholders of the Lower Passage, and they have consequently had recourse in order to get over it to an expedient which does credit to their courage. Taking advantage of the fact that the name of the second stream both in Polybius and Livy is a doubtful reading, they boldly suggest that it was nothing else than the Sorgues, the Fountain of Vaucluse. In the name of the pleasant coimtry to^vn of L'Isle-sur-Sorgues, so-called because the canals of the Sorgues flow round its waUs, they make bold to recognise a survival of the name of ' the Island.' ^ The passage at Tarascon is about ten miles north of that at Fourques. MONT CENIS AND ITS VARIANTS C9 This transfer of the Confluence appears to me to be m direct contradiction to every detail given by the classical authors, devoid of all geographical probability, and unworthy, therefore, of serious consideration. Nor can I understand how Poly- bius's four days' march from the Passage to the Confluence is worked in ; for from the Lower Passage to the neighbourhood of Bedarrides, which Commandant Colin selects for the Con- fluence, is only about 32 miles. In my mind there is no room for doubt that the Island of Polybius and Livy is the triangle lying between the Rhone, the Isere, and the Mont du Chat. I do not forget that Professor Wilkinson, following Commandant Colin, writes as follows : ' The hypothesis that the Island was the Rhone valley above the Isere, is inconsistent with most of the statements made by Polybius about the Island. This land was especially fertile, which is not the quality of the land north of the Isere. The mountains do not, even in appearance, form the base of a triangle of which the two rivers are the two other sides. Polybius distinctly says that the mountains forming the base of the triangle are impassable.' Let us see exactly what Polybius does say. The Island, he reports, ' is populous and corn-bearing,' the mountains at its base are ' difficult of approach and access, and almost, so to speak, impassable.' I have traversed this region four times by road and rail. I have also twice crossed the mountains on its east. It is still populous and corn-bearing, and it would be difficult to improve on Polybius's description of the abrupt range above the Lake of 70 HANNIBAL ONCE MORE Bourget. It was not till a.d. 1670 that Charles Emmanuel of Savoy made a road for wheels across it. Professor Spenser Wilkinson may naturally be disposed to set aside as prejudiced my personal im- pression. He may find some support for his own in a number of citations from a recent popular work, Ardouin Dumazet's Voyage en France. They show that here as elsewhere fertile and relatively poor and unproductive soils are mixed, but that there is abundance of weU-cultivated land. But I am confident that the general impression made on the mind of most modern tourists, more particularly on those coming from the brown hills of Provence, will agree with the description of Polybius. The following extract from the first edition of Murray's Handbook for Switzerland (1838, p. 322) may possibly help to convince readers that my impressions are likely to be those of the average traveller. ' Yenne existed in the time of the Romans. . . . It lies on the ancient route from France to Italy by the Mont du Chat. The whole neighbourhood is very rich in cultivation, corn, wine and fruit trees abound.' . . . ' The extreme richness of the country cannot fail to draw the attention of the traveller ... its excessive fertility is its striking feature. . . . The scene on looking back towards France is one of the most fertile in the world, studded with villages and towns.' ^ * Commandant Colin and his English follower both make a strenu- ous attempt to prove that the district north of the Sorgues and east of the Rhone is more worthy of the epithets ' populous and corn-bear- ing,' than that north of the Isere. At the present date the portions under irrigation are undoubtedly well cultivated and fruitful. But the page from Lentheric Commandant Colin himself quotes informs us that in the ninth century the plain was drowned by the stagnant waters MONT CENIS AND ITS VARIANTS 71 It will, I hope, be agreed by most critics that the ancient identification of the Island with the triangle between the Rhone, Isere, and Mont du Chat remains unshaken. Commandant Colin widens the field of dis- cussion by entering into somewhat elaborate essays on various kindred subjects, such as the physical characteristics of Southern Gaul, the local distribution of the mountain tribes, and the literary methods of Polybius and Livy. Into these often valuable and useful digressions I cannot afford space to follow him in any detail, and as regards Polybius, I in a great measure agree with him. But I may venture to point out, as examples, a few instances among many in these subsidiary chapters where he seems to me to let faU assertions prejudicial to his main argument. There is at first sight a good deal of plausi- bility in the suggestion that if Hannibal crossed the Rhone north of the Durance, the passage of the latter river would have been found a difficulty, or at least a cause of delay, both to the cavalry raid and to the subsequent advance of the Roman army from Fos. It is urged that there is evidence to show that in early times and up to the twelfth century the Durance was deeper and navigable.^ of the Sorgues, which formed everywhere pestilential marshes, and that L'lsle-sur-Sorgues 'n'6tait jusqu'au ix* siecle qu'une agglomeration informe de mis6rables cabanes de pecheurs construite sur pelotes dane la plaine noyee.' ^ The fact that the boatmen of the Durance are mentioned in Roman inscriptions allotting them seats in the Amphitheatre at Aries is brought forward as proof. But there were also boatmen, nautae — raftsmen, probably, who employed their ' utriculares ' in time of flood — on the smaller streams the Ardeche and the Ouveze, just as there are still on the Tarn. So that no certain inference can be drawn that the Durance had a greater volume in early times. See Cook, Old Provence, vol. i. p. 58; vol. ii. p. 248. 72 HANNIBAL ONCE MORE Again, when the supposed slowness of Hannibal's march from the passage to the Island — Com- mandant Colin's Island (Bedarrides) — has to be accounted for, we are told ' Annibal dut faire de tres petites journees en franchissant les divers bras de la Durance.' But when there is no occasion for special plead- ing the facts take a different complexion. We read (p. 32) of the lower Durance, where it meets the Rhone : ' La Durance s'y etale et son lit a parfois deux kilom.etres de largeur. Effrayante au moment des crues, elle est a peine un obstacle lors des basses eaux. Tous les bras secondaires sont alors a sec, le seul qui subsiste est large de cinquante a cent metres et sa profondeur est insignifiante. II est souvent gueable.' On page 292 : ' Dans la saison des basses eaux la Durance ne constitue qu'un obstacle tout-a-fait insigni- fiant.' There are therefore no adequate grounds for assuming that the passage of the different branches of the Durance, more divided then than now, was likely to have been at the season in question, when as a rule the waters are at their lowest, any serious impediment to an armed force. ^ I turn to the comments on the distribution of the tribes. On pp. 105-6 we read : ' Quand les Romains entrent en Gaule dans I'an 154, puis en 125-121 avant J.C, des qu'ils ont franchi la Durance ils rencontrent les Allobroges.' On sub- sequent pages we meet with such expressions as ' La suprematie guerriere des Allobroges dans la * Its waters are bt>Iieved to have flowed in several widely separate channels into the Rhone in Roman times. See Colin, pp. 33-4. MONT CENIS AND ITS VARIANTS 73 Vallee du bas Rhone ' (p. 106) ; a quotation from Strabo, ' Les Allobroges rangerent sous leur com- mandement les petits peuples voisins' (p. 132), ' En I'an 63 les Allobroges decident a secouer le joug.' ' Les Allobroges ravageant la Gaule Nar- bonnaise,' etc. (p. 133). ' LesRomains ont rencontre les Allobroges sur I'Eygues en 124-2 ' ^ (p. 140). We learn also (p. 80) on the authority of M. Lentheric that before the Roman conquest the Allobroges had a mint, and that their money was current in Southern Gaul. Nor does Commandant Colin pause to question, or explain, the further statement contained in the passage he cites that ' the Allobroges of Savoy and the Lake of Geneva ' and ' the Allobroges of Dauphine ' had a different coinage. Yet it is a striking proof of the extent and importance of the race — and if M. Lentheric's assertion is true, it is difficult to account for Com- mandant Colin's emphatic repudiation of Livy's statement with regard to the Allobroges. I would ask my readers to compare the fore- going references with the boundaries assigned to the Allobroges in Commandant Colin's map, and to the limits to which he would confine them in ^ Desjardins, La Gaule Romaine, vol. ii. p. 216 : ' Nous cherchons a retrouver I'etendue des domaines de ces peuples tels qu'ils 6taient a I'arrivee des Remains. La difficulte d'un pareil travail resulte d'abord de la p^nurie ou nous sommes de documents gdographiques pour cette ^poque, et ensuite de ces liens de patronage et de clientele si repandus dans la Gaule, liens qui empechent d'isoler ces peuples et parfois mcme de les distinguer entre eux. Ces hens, plus ou moins durables, unis- saient par une dependance plus ou moins 6troite les petits peuples aux grands, et cr^aient quelquefois meme une 6chelle a deux ou a plusieurs degres comme pour les fiefs du moyen age, t6moin les Allobroges, qui, lors de leur lutte supreme contre les Remains, rangdrent sous leur eommandement les petits peuples voisins,' etc. The concluding sentence illu trates and confirms my argument. 74 HANNIBAL ONCE MORE telling the story of Hannibal, when he is indignant at the suggestion that a tribe bearing this name could be found north of the Isere, in the posi- tion which at a later date was the centre of the race. Again, Caesar, we are told (p. 99), marched 27 kilometres, nearly 20 Roman miles a day, for many days, and that not in the Rhone valley, but over very difficult ground and the Col du Lautaret ! But when we come to Hannibal, we are assured that he only averaged 9 Roman miles a day while in the Rhone Valley, and this is enforced by the statement that armies have seldom marched more than 50 kilometres (35 Roman miles) in four days, unless before modern battles (p. 337). Yet Commandant Colin quotes from Vegetius a statement (p. 325) that the average day's march for a Roman army was in the plain 24 Roman miles, in the mountains 20 Roman miles, and his follower, Professor Wilkinson, adds that Napoleon counted 24 leagues (72 Roman miles) as four days' march. With respect to Polybius, I agree in the main with Commandant Colin' s remarks. But as to Livy, I find it somewhat difficult to reconcile his various criticisms. On one page we learn that Livy's accounts are generally drawn from good sources and are not to be thrown aside a priori ; that they are more detailed than those of Polybius ; that it is only in their arrangement that they are defective and may lead a careless reader to false conclusions (p. 143). But a hundred pages later he tells us that in ' the text of Livy, two blunders, two enormous blunders, spoil the whole picture, MONT CENIS AND ITS VARIANTS 75 the introduction of the Allobroges in or near the Island, and their suppression in the region of the first fight.' He goes on: * Polybius, we have re- peated it, and we repeat it again, had been in the country, had seen the Island and the Allobroges, visited the battlefield and the town pillaged by Hannibal. Livy, who remained at Rome and who, had he gone to the Alps would have foimd no exact tradition as to Hannibal, has no authority at aU in a matter of this kind.' ^ The above are not unfair specimens of the way in which our author is at times carried away by his argument. We must now follow Hannibal's march through the mountains as indicated by Commandant Colin. It is, as I have shown, on the strength of the 1600 stades from Emporion to the river that Com- mandant Colin fixes the passage of the Rhone at Fourques ; and it is on this identification that he asserts the Island to be the district opposite Avignon and immediately north of the junction of the Rhone and the Durance ; that he indicates, as the scene of the first battle, the acute angle in the Isere Valley, a few miles below Grenoble, and as the scene of the second battle, a narrowing in the Arc Valley near St. Michel. He thus stands com- mitted to an Arc — Dora Pass — but has still to ^ The coupling of the Island and the Allobroges in this sentence of Commandant Colin's is open to misunderstanding. It may therefore be expedient to remind readers as to what Polybius says. Polybius tells us that Hannibal's march through the Allobroges of the low country was smoothed by the Island chief, but that he was attacked by the Allobroges when he got into the mountains. It is a reasonable inference that Polybius did not look on the Island tribe as Allobroges, but considered the territory of that tribe as extending for several da\ s' march to the south or east. Polybius visited the Alps some seventy years after Hannibal's Passage. 76 HANNIBAL ONCE MORE make liis choice among several. To do this gives him no difficulty. For he has his second crucial test ready ; he has persuaded himself that the famous speech attributed to Hannibal is to be taken neither as the artifice of an historian, nor as a summary of what a great general might be ex- pected to say on reaching the crest of the Alps and seeing the downward slope and the streams running another way. He treats it as practically a verbatim report (Commandant Colin insists on every turn of phrase, and sometimes in his translation improves the turn) taken, we must suppose, by some Greek secretary on Hannibal's staff, perhaps Silenus, whose work is said to have been one of the principal authorities used by later writers. From this premiss, or assumption, he proceeds to the logical conclusion that the pass to be recognised as Hannibal's mu.st be one which commands a view of the plain of Piedmont, and also a glimpse of the Ligurian Apennine. For, in order fully to satisfy his demand for literal accuracy in the speech, there must be something in the view to hint the road to Etruria and Rome. From the Col du Clapier he believes himself to have enjoyed such a view, and he feels assured — as far as I know, rightly — that no other pass in the Western Alps, with a top broad enough for an army to camp on, enjoys the required prospect, which is one of his ' crucial requirements.' I assume, as does Mommsen, that the Cartha- ginian army in place of marching up the left bank of the Isere crossed into the Island. Professor Spenser Wilkinson takes a different view and believes that Hannibal marched up the left bank MONT CENIS AND ITS VARIANTS 77 and was forced into the hiUs of the Vercors at the bend of the stream below Grenoble. This sup- position hangs together with the Lower Passage of the Rhone. In order to fit the distance given by Polybius from the Passage to 'the Ascent of the Alps leading to Italy,' it is essential for Professor Wilkinson to find an ' Ascent ' some- where below Grenoble. To me it does not com- mend itself, for I find it hard to believe that Hannibal's guides would have involved him at the beginning of his march in needless diffi- culties. To render his plunging into the hills of the Vercors probable, we must assume that the passage of the Isere was a serious undertaking. I do not believe this. In a country so far civilised as this region must have been in order to furnish an army like Hannibal's with supplies not only of food, but of clothing, bridges of boats or ferries were not likely to be wanting over the minor rivers. The way in which Livy emphasises the difficulties in the passage of the upper Durance indicates that they were exceptional. After taking Grenoble, identified with the chief town of the AUobroges, Hannibal is described as marching up the valley of the Isere in three and a half days to Aiguebelle, or thereabouts (some 40 miles), and in two and a half more to the Rocher de la Porte below St. Michel (32 miles), where Professor Wilkinson lays the scene of the second battle. Thus Hannibal is brought up the Maurienne to the entrance of the glen that leads both to the Petit Mont Cenis and the pass now known as the Col du Clapier. I wiU do my best with the help of a map to put before my 78 HANNIBAL ONCE MORE readers the somewhat comphcated local facts as to the latter Col, and its relations to the better-known and lower passes. After an ascent from Bramans in the Arc Valley to the Planais basin, a second and steeper zigzag leads to the much higher level of the Savines glen.^ Here the traveller leaves on his left a broad grassy gap, crowned by some pastoral huts, which forms the Col of the Petit Mont Cenis. It leads on its farther side by an easy descent to the lake and the road of the great pass. But the way to the Clapier lies due east up a narrow glen of rock-strewn pasture which rises gently some 1200 feet in four miles. Commandant Colin calls it in the same sentence an ' immense plaine ' and a ' couloir.' The latter term would seem to me to be the more appropriate. But the glen does widen considerably towards its head, which is barred by a very low ridge somewhat under 1000 Awards in length. This forms the Alpine watershed and connects the base of Mont Ciusalet with the Mont d' Ambin : in the space between them lies an Alpine tarn, the Lac de Savines. The ridge rises in its centre into several knolls, while at the north end there is a shallow and narrow gap, Brockedon's Col de Clairee, the modern Col du Clapier. The wayfarer standing in this depression has no view whatever of the plain of Piedmont ; to see this he must climb one of the higher knolls, some five to ten minutes' scramble. Thence the eyes sweep ^ The frequent references made to the gentleness of the ascent from the Maurienne to the Little Mont Cenis are misleading. It is only gentle when the true Mont Cenis has been climbed first, that is, for travellers who start from Lanslebourg, like the Vaudois in the Rentr6e Glorieuse. |)Ldu CLAPIER AND fENIS PASSES ^^'^Ji THE COLduCLAPIER AND '■^^4 MONT CENIS PASSES Scale of Miles i^:-"-l 1/ y MONT CENIS AND ITS VARIANTS 79 the valley of the Dora, and see beyond it, enclosed between the crag of San Michele and the opposite range, a narrow strip of the plain with on the horizon the hills of Montferrat crowned by the Superga above Turin. This, at least, is all I could see on a fairly clear day. Commandant Colin describes a wider view, ' toute la plaine de Piemont, la Lombardie et k droite I'Appenin Ligure.' * Certes,' he exclaims, ' Annibal pouvait montrer a ses soldats non seulement la plaine du Po mais la route de Rome jusqu'en Etrurie.' If Commandant Colin will test his recollection by a large-scale government map on which the altitudes are given, he will, I believe, find that the high range south of the Dora Valley effectually blocks out the Ligurian Apennine and the road to Etruria. But this is, I admit, an irrelevant detail, except for those who deny to Hannibal imagination and to Polybius picturesque invention. The gap in question serves, as I have said, as a pass, a short cut from Bramans to Susa : owing however to the very precipitous character of the descent into Italy, it has never come into common use. From the pasture below the Col a rude staircase path zigzags down among cliffs and waterfalls for fully 4000 feet before it reaches the relatively level meadows in which the streams that form the Clairee unite. It is not long before the water falls again into a ravine, through which it finds its way to join the Dora. The modern pedestrian avoids this second plunge by following a terrace path, which guide-books call giddy, beside the canal that supplies the large village of Jaillon (Giaglione) on the slope immediately above Susa. 80 HANNIBAL ONCE MORE The track here described, a smugglers' and peasants' pass, is that now known as the Col du Clapier. It seems occasionally to have been con- fused in military literature with the much more serviceable Petit Mont Cenis. First drawn out of obscurity more than seventy years ago by the chance that Brockedon, one of the earliest English Alpine explorers, and the colleague of the John ^lurray of that date in supplying the material for the first edition of the famous Handbook for Travellers in Switzerland and the Alps of Savoy and Piedmo7it (1838), crossed it, this bypass has re- ceived a good deal of attention. Mr. Coolidge has examined and catalogued the literary refer- ences to it with his customary perseverance and accuracy.^ Professor Wilkinson has visited it and quoted at length from the Handbook and Blackwood'' s Magazine (vol. xxxix., May 1836) ^ Brockedon's descriptions of it. ' The pass,' Brockedon wrote, ' is on the Italian side, the steepest I have ever traversed ! ' The knees of most modern tourists who follow him down the 4000-feet staircase will give them cause to agree in this verdict ! M. Ferrand, who has thoroughly explored the neighbourhood, tells us that this staircase is too steep for cattle, and that the herdsmen on the Alp on the Piedmontese side bring their cows up by another way. To M. Ferrand we also owe some fresh facts whicli ^ Eivista del Club Alp. Ital., vol. xxx. No. 6, 1911. 2 W. Brockedon travelled widely in the Alps, and besides his prin- cipal work, The Passes of the Alps, 1828-9, published in 1833, Jownals of Excursions in the Alps, the Pennine, Graian, Cottian, Rhetiaji, Lepon- lian, and Bernese, and in 1836 a series of articles in Blackwood's Maga- zine. Brockedon's portrait by himself is hung among the portraits of painters in the Uffizi Gallery at Florence. MONT CENIS AND ITS VARIANTS 81 confirm and illustrate part of the literary material collected by Mr. Coolidge. He points out, that from the lake at the head of the Savines glen, two separate tracks, divided by the promontory on which Commandant Colin and Professor Wilkinson are fain to plant Hannibal, cross the watershed at points some 800 to 1000 metres apart. That to the north, leading through the narrow gap called by Brockedon the Col de Clairee to the staircase path, is, as I have shown, the modern Clapier ; ^ while that to the south after a short descent crosses level ground draining towards the Dora and containing some little tarns, the Lacs du Clapier, overhung by a glacier. M. Ferrand makes the important statement that he is unable to confirm Commandant Colin' s reported discovery of the remains of a Roman road on both sides of the watershed. This is a matter that should be further investigated. I was imaware of the alleged dis- covery, when on the spot. But to recognise and distinguish Roman from later roadwork demands more of an expert than the present writer. In his preface, I must repeat. Commandant Colin throws doubt on his own recognition, for he writes of ' une ancienne chaussee pavee, peut-etre une voie romaine.' ^ ^ It is also sometimes termed very appropriately ' the Col de I'Escalier.' * I quote the two passages : ' Le chemin qui passe a droit de I'^cran- chure du Clapier est bien celui qu'Annibal a suivi. Une voie romaine est encore la pour prouver que les anciens employaient ce passage. Cette voie pav^e de larges dalles avec des murs de soutenement bien construit se voit sur un longueur de deux kUometres du cote de I'ltalie. EUe disparait a Fapproche du fameux rocher qui a arrete les ^Mphants pendant trois jours. Sur le versant franQais la voie romaine ne suit pas la rive droite du Lac Savine dans les prairies ; elle longe le pied des escarpements de la rive gauche et les 6bouhs I'ont recouverte en grande F 82 HANNIBAL ONCE MORE After traversing the level strewn with small tarns this track climbs over rough screes to the brow of a precipitous spur (locally known as La Coscia) that overhangs the Clairee glen. M. Ferrand describes it as, like most cow-paths, free from any danger in good weather to ordinary hill-climbers, but passing in places close to precipices where a fall off it might have serious consequences. A conspicuous stone-man now marks the summit of the ascent. Beyond the screes the path, skirting the hillsides, reaches a gap in a lower spur, and then descends upon Exilles in the upper valley of the Dora, some eight miles above Susa.^ This path owes its name — the Col de Touilles — to some chalets close to the last ridge traversed and a long way below the Coscia spur. It is, in fact, not a pass, but part of a high-level route, and was in 1689 used as such in its entirety by a band of Vaudois in the famous ' Rentree Joyeuse.' Start- ing from Lanslebourg they, in order to avoid their enemies, crossed the Mont Cenis (6893 ft.), the Little Mont Cenis (7166 ft.), the gap (about 8250 ft.), south of the modern Col du Clapier (three passages of the Alpine watershed), and then climbed to the top of the cliff (the Coscia, 8807 ft.) before descending to Touilles and Exilles.^ partie. A peine apergoit-on ca et la un morceau de la chauss6e entre les blocs amoncel6s.' — Colin, p. 397. ' M. Colin dit d' avoir vu sur la rive droite du Lac de Savine des traces de voie romaine. Je n'en ai point vu, mais il est vrai que les ^boulis recouvrent tout. II dit aussi d'avoir vu deux kilometres sur le plateau des Lacs Clapier. Je n'ai rien vu, et la ce serait la prairie qui aurait pu recouvrir.' — Ferrand, Revue des Eludes Anciennes, vol. x. No. 1. ^ A bypath leads back to the lower glen of the Clair6e (or Clarea) by a circuitous route. * The name ' Clapier ' has apparently been given to the northern gap only of late years. I suggest that it has been transferred from MONT CENIS AND ITS VARIANTS 83 It was also used at a later date (1745) by a de- tachment of Franco-Spanish troops in an attempt to assail the fortress of Exilles from the rear. These adventures of small bodies of troops, lightly equipped and compelled for special pur- poses to seek out obscure bypaths, offer little or no analogy to the passage of an African army, numbering at this time, at the lowest calculation, thirty thousand men, with their baggage train. No general provided with competent guides — and we have no grounds for doubting the com- petency or loyalty of Hannibal's guides — would dream of selecting for such an army, and that in late autumn, this circuitous high-level route along the Alpine watershed as a short cut to the Dora Valley and the Piedmontese plain. The only plausible explanation of Hannibal taking it would be that his guides were traitors, and deliberately planned his destruction. I make a present of the suggestion (for what it is worth) to the advocates of this so-called Clapier — in fact the Touilles — route. Commandant Colin quotes from Honorius, a the southern track. Clapier = scree, and the characteristic of the latter track is that it traverses screes. There are none on the direct pass to the Clairee, which ought, in my opinion, to retain the name used by Brockedon. Mr. Coolidge is, I think, mistaken in identifjang ' La Coche ' with the modern Clapier. It is obviously a patois term equivalent to the Itahan ' Costa ' or ' Coscia,' a name given on the Italian maps to the spur of the Mont d'Ambin the Touilles track crosses. This is the highest point in the ' high-level ' route, being 761 feet above the gap in the watershed. There are thus really three passes, besides the Little Mont Cenis, on the high-level route, the true Clapier, the Col de la Coche, and the Col de Touilles of the Italian maps. The Col des Rochers P6nibles of the French map, and the * Rocher Ptnible' of the 1.50000 Italian Survey of 1871-4, would seem to be identical with the Col de la Coche. Its name, at any rate, is a paraphrase in good French of Clapier. 84 HANNIBAL ONCE MORE fifth and sixth century geographer, an obscure hst of passes in the Cottian Alps : ' Matrone, Cottidie, Marciane, Jule, Emingallo ' (sic), and identifies four of them as follows : Mont Genevre, Col de FEchelle, Col de la Roue (Oulx being ' Ad Martis '), and Col du Clapier. Any identification of the last he prudently gives up — though Mon Galese might have served him as a rough guess. ^ The first two names may either both refer to the Mont Genevre, or to that and some neighbouring pass ; the third, I agree, may be the Col de la Roue, the ' Jule ' is far more likely to be the Col de Frejus or the Mont Cenis than the Clapier. I have argued else- where that the Mont Cenis was probably Caesar's pass. I should add that Commandant Colin (p. 260) accepts Osiander's hypothetical identification of Modane with a Forum Julii. It would suit my theory that Juhus Caesar passed through the Maurienne. But the evidence for the name seems to me too slight to be relied on. Let us now see how Commandant Colin will deal with the difficulties encountered in the descent. He will have nothing to say to the staircase of the modern Col du Clapier. ' II ne parait pas vraisemblable que jamais une armee ait passe par le sentier vertigineux du Qapier, ou la largeur manque en meme temps que la raideur est ex- cessive.' 2 Hannibal, he says, from the Lac de ^ See Coolidge, Simler et les Origines de V Alpinisme, as to the use of ' Mon Gales ' in old maps, etc. 2 The quotations from Colonel Perrin given by Commandant Colin in support of this view indicate that the Colonel did not appreciate the exact meaning of ' Clapier.' It is not a chff, but a slope of broken rocks, a scree, the Cumberland ' clatter.' MONT CENIS AND ITS VARIANTS 85 Savines, took the southern, the right hand, gap and followed the traverse until where, near its highest point (the Coscia or Col de la Coche), the track was found to be broken. The Carthaginians first tried to turn the obstacle by climbing on to the ' veritables champs de glace et de neige ' of the Glacier de I'Agneau which hangs on the side of the Mont d'Ambin. But finding the foothold on the fresh snow lying on the surface of the frozen stuff very bad, they gave up the attempt and returned to the broken track. This they repaired by blasting the rock and bringing up large timber. How it came that at this height (8000 ft.) they found the ' huge trees ' of Livy ready to their hands Commandant Colin omits to tell us. The elephants could not have gone down to fetch them, for they were unable to cross the broken path. Here we leave Commandant Colin, for the question of what is a reasonable identification of the district 'of the Taurini who live close to the mountain skirts ' has been already gone into at sufficient length. I would, however, venture to urge again on the military critics one matter for their careful consideration. There is, as I have shown, no evidence forthcoming that the chief town of the Taurini was on the site of the modern Turin. But even assuming that it was, a tribe strong enough to decline to negotiate with, and to defy, the leader of a host of some thirty thousand men — with elephants — cannot have been restricted to the neighbourhood of any single town. It must have occupied a very consider- able territory, such a territory as is assigned 86 HANNIBAL ONCE MORE to the Taurini by Ptolemy and the ancient geo- graphers. One more criticism I must make before parting with this ingenious and suggestive author. Com- mandant CoHn unluckily supplies no index, but I have searched his pages assiduously and in vain for any reference to the verdict of Varro, or to the ex- istence of the military route of the Vars-Argentiere from the Durance into Italy. Yet the Comman- dant has read and refers (p. 405) to Mr. Marindin's articles in The Classical Review, and he must, there- fore, be fully aware that these are matters worthy of some mention, if not consideration, in any new contribution to the Hannibalian controversy. I have still to consider certain points in Professor Wilkinson's recent addition to the literature of our subject which are not fully covered by my comments on Commandant Colin's book. Pro- fessor Wilkinson's essay, first delivered, he teUs us, as a lecture before the University of Oxford, is in the main an able abridgement, as far as the famous march is concerned, of the French work, to which he frequently refers. His object in printing it has been to put prominently before an English audience the novel and somewhat startling theory that has found favour of late years among French miUtary critics, together with the gromids, or assumptions, on which this theory is supported. I will endeavour as far as possible not to weary my reader — assuming any reader to have got so far — by a repetition of objections that have already been more or less fully taken. Professor Wilkinson as an antagonist has the merit of putting his points concisely and clearly and without any MONT CENIS AND ITS VARIANTS 87 qualification. Thus he writes : ' The district of the Taurini is represented by the modern city of Turin. We may, therefore, take the neighbour- hood of Turin as a point fixed by the agreement of Livy and Polybius and may work from that fixed point.' Again he tells us : ' The nature of the view from the Col du Clapier is the crucial evidence that this was Hannibal's Pass.' To these * crucial points' I need not return. We must agree to differ. When we come to the passage of the Rhone, I find the English critic improving on the argument of his French predecessor. He does allude to Commandant Colin' s aU-important distance from Emporion to Fourques, but he also calculates backwards 2600 stades from Avigliana, which he takes as the end of the Alps, to the Passage, and so by a reverse process gets safely to the same spot — Fourques. No critic that I know of has yet failed to make the distances on this part of the march fit in with his theory. Proceeding to details, we find Professor Wilkin- son disagreeing with Commandant Colin as to the breadth of the river above Avignon. He beheves it to be too narrow at Roquemaure to meet the classical historians' statements as to rafts. But CoMn gives its breadth as 650 to 820 feet (p. 292) and the rafts only covered 300 feet.^ The Professor next tells us : ' Livy distinctly states that Scipio marched in order of battle ' from Fos to the passage. ' Quadrato agmine ' appears to me applicable only to the last day of the march. He concludes that ' the distance from * See ante, p. 66. The French survey maps confirm Colin's figures. 88 HANNIBAL ONCE MORE Fos to a point on the river a mile or two north of Fourques perfectly accords with the text (of Polybius) with which in all that concerns the passage of the Rhone that of Livy precisely concurs.' Tliirty-four miles is therefore, in Pro- fessor Wilkinson's view, the perfect and precise equivalent to ' about a four days' march ' ! And let me again note that the passage in which Polybius calls it a four days' march is one where he is giving the time normally occupied without any particular reference to Scipio's particular rate of march. This Polybius, contradicting his modern commentators, explicitly says was hurried. Professor Wilkinson continues : ' Livy distinctly states that Hannibal marched in order of battle, a proceeding which is inconsistent with rapidity of movement. Hannibal's average rate of march up the valley of the Rhone was less than nine miles a day. Scipio marching with his army drawn out for battle and carefully fortifying his camp each night would be very unlikely to move faster than this.' But I do not find in Livy any such state- ment as to Hannibal's average rate of march up the Rhone valley, or that Scipio fortified his camp each night, or that he marched all the way in battle array. The first statement appears to me an ille- gitimate inference (see p. 68), and the two latter unwarranted expansions of Livy's actual words. ]\Ioreover, Commandant Colin takes quite a different view (p. 290), ' Les Legions devraient marcher d'une allure plus vive que I'armee-car- avane d'Annibal.' W^hen such military authori- ties disagree the ordinary man perhaps may be excused for holding to his own opinion. MONT CENTS AND ITS VARIANTS 89 Next comes the Island. Polybius's statement that it has at its base ' momitains difficult of approach and access, and almost, so to speak, impassable,' is converted into ' mountains hard to approach and to enter, and therefore impassable ' (the italics are mine), which is a very different thing. Professor Wilkinson adopts Colin' s view that the Island is the district opposite Avignon under Mont Ventoux and that the second river is the fountain of Vaucluse — the Sorgues — though he is apparently ready, if pressed, to substitute for it the Durance. For the rest of the march up to the pass, for the two battles, our authors are at one. To Professor Wilkinson the description of the descent suggests a glacier. As far as my Alpine experi- ence goes, and Mr. Coohdge and I believe all English mountaineers agree with me, the classical narratives do not indicate the passage of a glacier. Glacier ice — particularly near the foot of a glacier — does not yield to trampling ; the difficulties described are exactly those which are met with ordinarily on avalanche beds, or where there are two layers of snow from successive faUs. Professor Wilkinson accepts the doubtful emend- ation of Livy which is taken to involve a vertical precipice 1000 feet deep — where Polybius mentions a break of the same length in the path. Professor Wilkinson concludes his chapter thus : * The explanation suggested by several visits to the pass, and repeated study of the text of both writers, is that Hannibal made use both of the descent into the "amphitheatre" (the staircase) and of the '' cor niche'''' (the Coche-Touilles traverse). 90 HANNIBAL ONCE MORE The infantry would start down the ordinary path, and in spite of the accidents many of them would go that way. Hannibal would think it too dangerous for his horses, mules, and elephants, and would start with them along the corniche. Then he would try the glacier, and having given that up would repair the broken corniche for the animals.' Of where he found the big trees that supplied bonfires and helped to mend the road, we are told nothing. The spot indicated is far above the tree-level. Here it will be seen Professor Wilkinson im- proves on his predecessor by bringing ' the stair- case ' into some use. He next proceeds ' to give reasons why the passes other than the Clapier must, in his opinion, be unconditionally rejected.' As to the Little St. Bernard, our reasons are the same ; the Mont Genevre he disregards because it has ' no view of Italy, no glacier.' A glacier, it seems, has also become a crucial requirement. From Professor Wilkinson's version of the march, we turn to see how he treats the argu- ments of earlier writers. He does not deal, nor can any writer on the subject be expected to deal, with them in any detail. But it seems to me a fair matter for regret that in his lecture and booklet, when referring to Dr. Arnold's history. Professor Wilkinson did not notice the edition of his chapters on The Second Punic War, brought out in 1886 by his own friend and Dr. Arnold's grandson, the late W. T. Arnold, a scholar and a critic of sound judgment. There, in an appendix of eleven pages, may be found a valuable comment on the MONT CENIS AND ITS VARIANTS 91 classical texts written with a full knowledge of the literature of the subject up to that date, together with a clear and judicial summary of the whole controversy. There also readers may learn that the Clapier hypothesis was even a quarter of a century ago no novelty to EngHsh writers. It is true they will not discover any support for that hypothesis. W. T. Arnold wrote, ' So far as I can discover, the only serious argument (for the Col du Clapier) is that it commands an ex- ceptionally good view of Italy. This point, of the view of Italy, is not one of those on which it is advisable to lay stress,' and again, Arnold's last words are, ' I should be disposed to say that IVIr. Freshfield — for I do not regard Colonel Perrin, with his Col du Clapier, as a serious rival — holds the field.' ^ I now pass on to Professor Wilkinson's remarks on my argument con tamed in the Alpine Journal (vol. xi. p. 267). He does not seem to have been acquainted, when writing, with my subsequent notes on the subject {Alpine Journal^ vol. xiii. p. 28, and Proceedings R.G.S., new series, vol. viii. p. 638). In these, while modifying in some important respects my previous argument, I commented, not only on Colonel Perrin' s Clapier hypothesis, but also on the analysis of recent Hamiibahan literature contributed in 1884 to the Berliner ^ Professor Wilkinson has subsequently explained {Geo. Journal, vol. xxxvii., No. 6, p. 670) that in his behef his friend and former colleague was led astray by an exaggerated respect for my Alpine experience, and that he cannot therefore be regarded as an independent judge. This seems to me — to say the least — a strained view (see ante, p. 12). But even if my modesty allowed me to accept it as far as my adoption of the Col de I'Argentiere is concerned, it can have httle or no bearing on W. T. Arnold's summary dismissal of the Col du Gapier. 92 HANNIBAL ONCE MORE Philologische Wochenschrift (vol. iv. pp. 705, 737, 769), by Professor Schiller, in which that learned writer concluded in favour of one of the Durance passes, a view also expressed more recently by Mr. Coolidge in the Alpine Guided Professor Wilkinson's objections have for the most part been already met in my comment on Commandant Colin' s volume. He writes : ' Livy says that from the Durance onwards Hannibal marched across a plain to the Alps ' ; and he adds, ' There is no plain whatever in the upper valley of the Durance, or between it in the Alps.' But Livy's own words are : ' Campestri maxime itinere.' The paraphrase given seems to me forced ; ' for the most part over open (or level) gromid ' would be nearer the Latin, and in any case the sentence must be read in connection with the immediate context, a description of the rugged character of the surrounding scenery. In fact, the road from Gap through Embrun to Guillestre is either over relatively open ground or in a valley, and there is no considerable ascent. As Mr. W. T. Arnold, and M. Chappuis before him, have pointed out, Livy's description of the landscape of the upper Durance vaUey is, despite all denials, singularly faithful to its present aspect. Professor Wilkinson emphatically asserts that the pass of the Barricades in Val di Stura cannot, in the absence of opposition from an enemy, offer any difficulty whatever to the passage of an army. ^ Alpine Guide, vol. i. p. 106 : ' The Col du Clapier is mentioned by French topographers in the eighteenth century, while Hannibal, according to a very improbable theory, and certainly the Wal- denses, traversed it.' Mr. CooUdge himself inclines to the Mont Genevre. MONT CENIS AND ITS VARIANTS 93 He is also persuaded that ' on the descent from the Col de Larche (de I'Argentiere) there is no possi- bility of men or mules falling over a precipice.' The best answer will be supplied by the quotations from Paulus Jovius and the Marquis de St. Simon ■contained in these pages (pp. 45 and 103). Professor Wilkinson comments somewhat severely on assumptions made in support of hypo- theses other than his own. The problem of Han- nibal's Pass is one which cannot be attacked without assumptions ; but, where they deal not with opinions but with ascertainable facts, it is a reasonable requirement that they should be such as a contact with facts, local or historical, does not dispose of. Do Professor Wilkinson's assump- tions invariably support this test ? For example, he assumes his passage of the Rhone ; he assumes that the district of the Taurini was confhied to ' the neighbourhood of Turin.' On these two assumptions his route is largely based. Again, he would fain persuade his readers that before carriage roads were made the modem high-road- passes were not more convenient than others for armies on the march. This, surely, is to con- tradict history ! From Hannibal's time onwards the passes of the Western Alps which have been traversed by armies — I do not speak here of small detachments, or bands of fugitives — have been the passes over which the carriage roads of the eighteenth and nineteenth centuries were subse- quently constructed. It is across them that Roman and mediaeval and modern armies have marched, princes ridden, and pilgrims plodded. The Roman passes — the Comiche Road, the Col de 94 HANNIBAL ONCE MORE I'Argentiere, the Mont Genevre, the Mont Cenis, and the Little St. Bernard, reappear as the Passes of Brockedon's volumes. In selecting and improv- ing the gaps in the chain that are both the lowest and the most accessible from the neighbouring valleys the Romans showed their usual practical capacity, and modern roadmakers have had suf- ficient common-sense to follow in their footsteps. Before parting with Professor Wilkmson, I would again put to him this final inquiry. He agrees that Varro's route, ' Juxta mare per Ligures,' is identical with the coast-road — the Corniche. Can he suggest any new representative of the Graian pass in place of the Little St. Bernard ? Does he agree with the Master of Balliol in accepting the Mont Genevre as Pompey's Pass ? If, as must seem to most critics who know the Alps, these identifications of three out of the five passes in the catalogue quoted from Varro are indisputable, can he supply any plausible ground for rejecting the surely obvious and almost inevitable inference that the two intermediate passes are referred to in their proper topographical order, going from south to north ? This is a point on which no adequate explanation has hitherto been offered by any of the writers who, miable to discredit the passage in question, are yet loth to give to it its due weight. I trust I have made it clear that Commandant Colin' s and Professor Wilkinson's case rests, by their own frank avowal, on their identification of Fourques, near Aries, as the passage. ' Pour nous,' writes the Commandant, ' tout le probleme est la,' and he adds ; ' c'est sur ce point que les MONT CENIS AND ITS VARIANTS 95 erreurs les plus graves ont ete commises et avec line unanimite presque complete.' ^ I am afraid I must continue to walk in the paths of error with the blind but enormous majority. For I lay down this last essay to solve the ancient riddle wholly unconvinced. The speeches put in the mouths of great men by classical writers seem to me to be at best but unstable keystones to build round ; the shifting of the Island strikes me as a desperate expedient ; while the march of Francis i. across the Alps from Grenoble to Cuneo in A.D. 1515 offers in too many respects analogies to that of Hannibal to be hghtly disregarded. But while insisting on the weaknesses, I appreci- ate, I hope, fairly, the strong points, both positive and negative, of those who are partisans of the Clapier and opponents of the Mont Genevre or Vars- Argentiere routes. They are, I think, entitled to urge that Polybius uses expressions which may be taken to suggest, if they do not positively assert, that he himself had crossed the actual pass used by Hannibal, and to ask if in this case he would have quoted Hannibal as pointing out the plains of the Po, when he must have known that they were in- visible ? I dare not go as far as ]\lr. Marindin, when he writes, ' No reasonable man will trouble himself about the view from the top.' I can conceive a reasonable man falling into a mood so devoid of picturesque imagination, or of the appreciation of it in others, as to make him join in the demand for such a view. My opponents may, I think, legitimately ^ It must be borne in mind that Colonel Perrin, Lieutenant Azan, and M. Ferrand, (Commandant Colin'a comrades in upholding the claims of the Col du Clapier, do not agree with his theory as to the Passage of the Rhone. 96 HANNIBAL ONCE MORE go on to suggest that the snow conditions found at or near the broken road indicate that the incidents connected with them occurred at a level higher than that of the Barricades. They may urge the stronger claims of some other and loftier defile. They may make the most of the dis- crepancies and the difficulties that undoubtedly occur in dealing with Livy's narrative. They have a case which, boldly urged, is, I admit, super- ficially plausible, so long as the other side is not heard. But the question is : have they a case which can prevail in the court of any competent and unbiased judge? I do not for a moment expect to convince critics who have already com- mitted themselves to an opinion. But it seems to me worth while to point out to the public and to students who are not Alpine travellers, or as yet sworn to any particular theory, the weakness of many of the buttresses which have been piled up to support what, to one who knows the Western Alps, seems an incredible hjrpothesis. The Col du Clapier has, it is true, made of late some distinguished converts. The view of the three French officers Perrin, Azan, and Colin has the support of M. Ferrand in their own country, and of Professor Spenser Wilkinson in ours. But I must continue to direct my efforts to ' under- mine the credibility ' of a belief according to which the best guides that the mountain tribes could provide deliberately selected for Hannibal's army one of the loftiest and, on the Italian side, the steepest and least practicable of the bypasses of the Western Alps, and that at a time of year when winter was already settling on the heights, and MONT CENTS AND ITS VARIANTS 97 despite the fact that two passes respectively 1000 and 1200 feet lower, and far easier, lay close at hand. That the African host was further in- duced to make an excursion on one of the few glaciers that can be found near the watershed between the Maritime Alps and the Mont Cenis adds, to my mind, a finishing touch to this strange paradox. I agree with Mr. Coolidge and Mr. W. T. Arnold that the claims of the Col du Clapier ' cannot be considered as serious.' Converts to a definite theory on the pass of Hannibal are, as any one who has studied the literature of the subject must realise, apt soon to become bigots. Let me pause, therefore, and define my own point of view lest I too fall under the same condemnation. The Little St. Bernard, I hold, may be put out of the question ; the Durance passes — the Mont Genevre and the Col de I'Argentiere — attract my sympathy. But the final discovery has not yet been made and verified. M. Ferrand suggests that some mi- employed millionaire, or society of antiquaries, should take to spade-work and dig about the Cottian Passes,^ or drain the lakes on their sum- mits. The suggestion is a fascinating one to the Treasurer of our Roman Society ! For until a suit of Punic armour or an elephant's skeleton has been brought to light the Alpine adventures of Hannibal must, I fear, continue, as in Juvenal's days, to interest boys and perplex schoolmasters. Mark Twain's summary of the ancient argument has ^ Any enthusiasts who are prepared to undertake the task may be glad to know that there is a ' Club hut ' below the glacier on the Touilles route. As the keys are kept at Susa, the passing traveller coming from Savoy does not easily profit by it. G 98 HANNIBAL ONCE MORE lost none of its pungency. ' The researches of many antiquarians have already thrown much darkness on the subject, and it is probable, if they continue, that we shall soon know nothing at all.' After carefully considering all these fresh con- tributions, I still hold that as between the Mont Genevre and the Col de I'Argentiere, the evidence, setting aside Pompey and Varro, is very closely balanced, but that the scale inclines in favour of the southern pass. If we admit the passage from Varro, and assume that his catalogue gives the passes in geographical order, the question is, of course, solved in favour of a Durance pass, other than the Mont Genevre. Yet, while I protest with a most positive conviction against all the northern passes, I cannot work myself into an absolute belief in the Argentiere. My mind is still open to consider any plausible substitute south of the Mont Genevre. Before laying down my pen, let me indicate in a few final sentences the probable outline of the story of the Durance passes as it has presented itself to me in the course of this inquiry. The Keltic hordes poured down through Gaul till they came to lands already occupied and ill- suited for pasture.^ The fame of Italy led many of their tribes up the valley of the Durance. When they reached its angle, they held straight on over the broad downs of the Col du Vars and up to the gentle crest of the Argentiere. For centuries this remamed — it remains to this day — a natural and frequented highway for the people ^ See Livy, lib. v. c. 34. MONT CENIS AND ITS VARIANTS 99 on either side of the chain. Then the Romans, masters of the valley of the Po, approached the eastern base of the Alps, and, no longer content with sailing oversea, sought a highway to the Province and Spain. The broad trench from which the Dora Riparia issues — next to Monte Viso the most conspicuous physical fact in the view from Turin — led their generals on to the dis- covery of the Mont Genevre and, when they wanted to go north — to Gallia, and not to Hispania or the Provincia — of the Mont Cenis. It is curious how many writers on the ancient passes seem to overlook the fact that a traveller, or an army, aiming at Lyons, and one making for Aries, will naturally take different routes through the Alps. It was the annexation of Gaul that gave a reason for the Mont Cenis. It was the subjugation of the Salassi that displaced it in favour of the Little St. Bernard. The Argentiere lost much of its importance after the opening of the Mont Genevre, as the Monte Moro and Gries and the Great St. Bernard lost their importance by the construction of the Simplon road. In modern history it flashed out again into temporary notice, and then its very existence was ignored by the dryasdusts who construct orography in their cupboards. Yet military engineers are still alive to its im- portance, and every year hundreds of sturdy Piedmontese, descendants of the old Ligurian stock, pour backwards and forwards over it on their way to and from their winter work in the cities of Provence. APPENDIX I Francis the First's Passage of the Col de l'Argentiere The view here put forward that Hannibal crossed the Col de l'Argentiere may be illustrated by reference to the story of a later invader of Italy, who met on this pass with adven- tures afiEording in many respects a singularly exact parallel to those of the Carthaginian commander. In the absence of the original authorities, the general reader may obtain from Murray's Handbook, or Brockedon's Passes of the Alps, the main facts as regards the passage of the Alps by Francis the First in a.d. 1515 before the battle of Marignano. He may seek further details from Sismondi or Michelet. But he will do so at his peril. Sismondi, as Brockedon has pointed out, goes hopelessly astray in his local geography, while Michelet indulges his talent for the picturesque vsdth even more than his usual freedom. The following passage may be taken as a specimen. This is the introduction to his description of Francis the First's passage of the Alps : ' Le vieux Trivulce se mit k courir les Alpes, et trouva cet affreux passage entre les glaces et les abimes — sauvages gorges ou nul marchand, nul colporteur, nul contrebandier n'avait imprime ses pas. La virginite de leurs neiges n'etait effleuree, depuis la creation, que par I'enfant de la montagne, le craintif et ruse chamois, et parfois aussi, peut-etre, par I'intrepide folic du chasseur que la passion entraine apres lui aux corniches etroites des gouSres.' Some virriters assume Trivulzio to have led the body of cavalry that crossed the Col dell' Agnello. This, however, is not borne out by the contemporary authorities I have con- sulted. He was not with the advance guard that captured Prosper Colonna. See Paulus Jovius. 100 APPENDIX 101 From such of these authorities as are accessible to the student I make some selections here. The first will be a letter written from the village of St. Paul, below the Col de Vars, by Francis the First to his mother — Louise de Savoie. It is printed in Rochas d'Aiglun's Les Bourcei et leur role dans les guerres alpines (privately printed, Paris, 1895). I reprint it here from a copy kindly furnished me by M. Ferrand : ' Madame sens (ce) quy me gardera de vous fere plus longue lettre sera que je monte a set eure a cheval pour la grant trete qu'avouns a fere pour parfere le demeurant de la joumee, vous asseurant que nous sommes dedans le plus estrange paj^s ou James feut home de sete conpagnye. Mes demayn espere destre en la plene de pyemont avesque la bande que je mene ; (ce) quy nous fera grand plesyr ; car il nous fache fort de porter les armes parmy ses montagnes, pour ce que, la plus part du tamps, nous faut estre appie et mener nos sevos (chevaux) par la bryde, de quoy le grand mestre est byen marri, Et croy quy n'oret veu seus (ce) que voyons seroyt impossyble de croyre mener gens de seval (cheval) et grosse artyllerye come fesons, Et croyez, Madame, que ce n'est sans pene, car ey ne fouse aryve, noustre artyllerye grousse foust demouree ; mes, Diou mercy, je la mene avesque moy ou bientoust apres. Je ne vous mande ryen de la defete qu'a fete le mar^chal de Chabannes et de la prise de Prosper Coulone, ny ossy I'evenement du syege de Sylles, remettant tout sur le grand mestre, vous avysant que fesons bon guet, quar ne sommes qu'a synq ou sy yeus (six heues) petites des Souyses (Suisses) ; et seur ce poynt vous va dyre bonso}^. — Votre tres-humble et obeyssant fils Franco ys.' Aymar de Rivail, who wrote in 1535, confirms this state- ment : ' Per collem Argentariam, Rupem Scissam (The Barricades), Brezieram (Bersezio), Intermontium (Demonte), et Cuneum in Itaham Franciscus cum equitatiis mole et tormentis pervenit.' * I cannot find what is the authority for the statement in Rochas d'Aiglun's text that Francis sent his heavy artillery * De Allobrogihue, I. ix. (Lyons, 1844), octavo. See Les Sources dc VHistoire de France, X VI. Siecle, par M. Hauser. Paris, H. Picard et Fils. 1906. 102 HANNIBAL ONCE MORE (or any part of it) over the much higher Col dell' Agnello. But I presume it is on the strength of this passage that Professor Wilkinson has questioned my statement as to the pass over which the large cannons were taken, Paulus Jovius on this point seems to support the natural conclusion I draw from the King's letter, that he had his guns with him. He describes the bypasses in the neighbourhood as ' pedestria concisaque itinera, quae nisi magno labore atque periculo patefieri non possent.' And as to the Col dell' Agnello and its passage, he writes : ' unum expeditum agmen per Agnelli collem, impru- dente vel conjuvente Joachimo, angustiarum custode, supra Cunium oppidum processerat.' Note that it was an expedi- tum agmen — a Hghtly equipped detachment — and it is referred to later on as cavalry.^ Further details as to the incidents of the march may be gleaned from one Pasquier le Moyne, a rhyming chronicler, who expatiates on the ruggedness of the country and the narrowness of the paths : ' Les lieux ^talent si etroits Qu'on n'y passait ni deux ui trois A la fois, mais un seulement Apres I'autre tout bellement.' ^ He tells us that while waiting for the Pas de la Reyssole to be made practicable Francis gave a great supper to his captains at St. Paul. Then we have Paulus Jovius. In the fifteenth chapter of the Historia sui Temporis, pubHshed in 1533, he gives a brief sketch of the Western Alps : ' Coctiis Argentariae Alpes junguntur apud antiquos parum celebratae quod propter angustias et frequentes viarum asperitates sarcinariis jumentis, ne dum curribus, vix pateant. Argentariis vulgo in Pro- vinciam transmittunt qui per Avennionem atque Narbonam in Hispanias proficiscuntur .^ (The itahcs are mine.) He goes on to give the following account of Francis the First's passage of the Col de I'Argentiere : In August 1515 the King set out from Grenoble to pass the Alps, with an ' incredible number ' of carriages, powder- ^ Geographical Journal, vol. xxxvii. No. 6, p. 672. * Pasquier le Moyne, Le couronnement du Roi Frangoia : Voyage et Conqueste de la Duche de Millan. Paris, 1520. APPENDIX 103 waggons, and 72 large cannons and 300 small pieces, drawn by 5000 horses. The issues of the Mont Cenis and the Mont Genevre, the ' Cluse ' or defile of the Dora below Susa, the country about Turin, were in the hands of the enemy. An- other route had to be found. That up the valleys of the Drac and Durance was selected. The army reached Embrun without difficulty of any sort. There the generals reposed and re victualled their forces. Then they set out with five days' food for the passage of the chain. Their route led them across a ford of the Durance to Guillestre, where they camped, and over the Col de Vars into the upper valley of the Ubaye. On the descent they encountered the Pas de la Reysolle. Having, by means of blasting, made it passable for cannon, they, on the the next day, crossed the Col de I'Argentiere. In the Pas de la Reysolle and again in the gorge of the Barricades the greatest difficulties had to be overcome — guns were s\vung from rock to rock with ropes ; but on the third day from the Barcelonnette valley the whole force successfully traversed the latter defile and issued into the open valley of the Stura.^ Of the state of the western passes of the Alps at the begin- ning of the sixteenth century we have a further account in a *■ I had better give the original text of Jovius : ' Ita summo regis studio quum nihil laboris atque periculi milites recusarent, a Gratianopoli Vigiliam atque inde junctis omnibus copiia Muram, et recto itinere Eburodunum exercitvis ductus. Eburudurio profecti Borbonius atque Trivultius, qui primo agmini prseerant, quinque dierum commeatu proviso, ad S. Clementis atque Crispini Alpinos vicos contenderunt. Inde sub Itevam relicto Genebra monte, Druentiam flunien exercitus vado trajecit : positisque ad Gilestram castris et mox Avaltio monte superato ad S. Pauli rupem magno la- bore perventum. Eam, quod abrupta inaccessibilisque erat, incredibili celeritate ferro pandunt, tormentaque traducunt. Sequent! die in Barcelloniam vallem descensum. Ea ingentibus saxis et iniquissimis collibus interpositis impedita magnam rerum desperationem afferebat. Nam ligonibus dolabrisque proscindere saxeos coiles, exiequare crepi- dines, et, quum nullus per derupta equorum usus foret, subjectis mili- tum humeris tormenta transvehere necesse erat. Interdum ea magnis funibus ad scopulos et stipites arborum circumductis suspondebantur, et versatilibus machinis ergatarum et trochlearum artihcio de rupe ad rupem, intercedentibus profundissimis vallibus cum summa admira- tione totius exercitus, trahebantur. Nonnullis etiam in locis nudarum rupium latera, ubi via deerat, suppositis tibicinibus interjectisque lon- guriis muniebant, et insuper injectis stratisque virgultorum fascibus ces- pitibus ac glebis pensiles vias transeuntibus curribus parabant. Ita mira fabrorum industria et singular! militum labore in Argentariam vallem cuncta exercitus impedimenta traduxerunt. Sequent! die a Larchia Ebergiaque oppidis in Asturiae vallem universum agmen se demisit. 104 HANNIBAL ONCE MORE rare tract by one Jacques Signot, first printed in a.d. 1507, and several times reissued up to 1518. A copy of the 1518 edition is in the British Museum Library.^ It may be worth while to analyse briefly his text, which has some bearing on our argument, both in its mention of the Argentiere and in its omissions. Signot deals only with the passes leading from Gaul, or France, to Italy, that is, from the frontier of Allemagne to the river Var, ' Et tous les passaiges dessus ditz commencent des la frontiere d' Allemagne et durent jusque a la mer Ligustique . . . et n'y plus d'aultres passaiges,' he writes. Signot, it is clear, draws the frontier of Allemagne close to the Great St. Bernard, that is, at the western hmits of the bishopric of Sion. It is near the linguistic frontier. We may compare Leonardo da Vinci's statement that the Monboso he climbed, now generally identified with one of the southern spurs of Monte Rosa, above Val Sesia, still known as Monte Bo or Boso, was on the borders of Allemagne. From Savoy Signot mentions three passes. The first is the Great St. Bernard ; the second, the Little St. Bernard. He describes the defile of Fort Bard, which he believes Hannibal opened, the inscription long supposed to record his passage,^ and the Roman bridge at Pont St. Martin. Signot' s third pass is the Mont Cenis (Mont Senys), over which he tells us Charles the Great went on his way to Rome in 773. (See as to Charles and the Mont Cenis Mr. Coohdge's essay, English Historical Review, July 1896.) The remaining seven routes, he tells us, all traverse Dauphine, the Saluzzese, or Provence. The Mont Genevre, he says, 'est le meilleur et le plus aise passaige mesmement ubi Pedeporcium montem, qui mediam vallem intersecabat et iter reddebat iniquissimum, pari munitorum artificio fractis atque revulsis inimanibus petris domuerunt et exsequarunt. Ab Pedeporcio Avennam atque inde Sambucum et mox ad Italiae fauces tribus tantum diebus eo itinere consumptis incolumis exercitus pervenit.' * Siguot Jacques : La totale et vraye description de tous les passaiges . . . des parties de Gaules es parties d'Ytalie et signamment par oii pas- serent Hannibal . . . Charles viii., Louis xii., rran9ois, etc. . . . Paris, 1507 and 1515. See Les Sources de VHistoire de France, XVI. Siecle, par H. Havard. Paris, A. Picard et Fils, 1906. Reprinted Ann. de la SociM des Touristes du Dauphind, vol. x. 1884. ^ See as to the inscription, Vaccarone, Le Vie delle Alpi Occidentali, p. 60. Turin, 1884. APPENDIX 105 pour la conduite de Tartillerie. Et de tous les aultres n'y a celui par lequel on la peust faire passer fors cellement par icelluy passaige.' Signot mentions three ways to Brian9on : (a) by the Col de la Croix Haute, where one passes by the ' Mont de la Croix Haute.' This is the famous ' Mont Inaccessible,' first cUmbed in A.D. 1492 (see the Alpine Guide, vol. i. p. 178) ; (6) by La Mure, Corps, and St. Bonnet, and over the Col de Manse to La Batie Neuve, or another parallel pass he calls the Col de Chauvet (Col Bayard ?) to Gap ; (c) the third is the Col du Lautaret, but is ' fort difficile.' Signot repeats the popular saying (' les bonnes gens de pais desine ') that the Dora and Durance both spring from fountains on the top of the Mont Genevre. The descent from the Mont Grenevre to Susa is described, and also the alternative way into Italy by the ' Val de Peroux,' that is over the Col de Sestrieres. This is counted as the fifth pass. The two passages of Charles vin. are mentioned at some length. In both the Susa route was preferred. I may point out a noteworthy point in connection with our present contro- versy. Signot mentions a number of variants in the approaches to the Mont Genevre ; he gives various routes across the Alps from Val de Queyras. But when deahng with the Mont Cenis we find no mention whatever of the Little Mont Cenis, or the Clapier, which its advocates tell us was a frequented track at this date. For the sixth pass Signot starts from Guillestre and goes up the Val de QuejTas, through the gorge ' where there are fifteen bridges in five leagues.' At Chateau QuejTas two tracks part — one goes by St. Veran and over the Col dell' Agnello to Chateau Dauphin, and so by the Val Varaita, or Val Maira, into the Saluzzese and Italy. It would seem that the mention of St. Veran is a sHp, as that glen does not lead to the Col dell' Agnello, but to the Col Blanchet and Col de St. Veran. {Alpine Guide, vol. i. pp. 50-1.) Signot alludes to, but does not reckon among his ten l^asses, the Col de la Croix from Abries to Val PelUce. His seventh pass, which hes north of the fifth and sixth, is the famous Pertuis du Viso, or Col de la Traversette, the tunnel 106 HANNIBAL ONCE MORE through which was pierced in 1480 by the Marquis of Saluzzo, with the pecuniary aid of Louis xi. {Alpine Guide, vol. i. p. 68, and Signor Vaccarone's Pertuis du Viso, Turin, 1881.) The eighth pass is the Vars-Argentiere route, by which Demonte is reached, and an entrance to Piedmont. Here travellers bound for the Riviera of Genoa turn to the right by Cuneo, Ceva, and Savona, that is, over the Col d'Altare. The ninth pass is the coast road from Nice. Signot asserts that at Turbia Julius (sic) Caesar built not only the Trophy, but a triumphal arch on his return from the conquest of Great Britain ! The tenth pass is the Col di Tenda. Signot's work with the map found in some of the editions is reproduced in the Annuuire de la Societe des Touristes du Dauphine, vol. x. Grenoble, 1884. In an article in the monthly magazine of the ItaUan Alpine Club 1 Dr. F. Mader informs us that the first mention in post- classical hterature of the Col de I'Argentiere is found in a Provengal poem, ' La vida de St. Honorat,' by Raymond Ferant, completed in its present form about a.d. 1300, but revised from a text about a century older. It relates how St. Honorat, oddly described as the son of a pagan king of Hungary (he was, in fact, bom in a noble family of Lorraine in the fifth century), was converted, and from VerceUi fled over our pass and past the Lac de la Madeleine to the Isles de Lerins. The approach to the lake is thus described : ' Per un cendier estrech E reguardan al puez a som d'una niontayna El Mont de I'Argentiera en la forest estrayna.' The3^ chmbed ' by a narrow path and gazed on a lake on the summit of a mountain, the Mountain of Argentiera, in the wild forest.' Gioffredo "^ relates that Charles Emmanuel i. of Savoy in 1590 described the lake in a letter to his wife as one of the most charming, and the mountain as one of the most beauti- ful, he had seen : a very surprising sentiment for the date ! ^ ' Appunti Storici su alcuni passi delle Alpi Marittime,' Dr. F. Mader. Rivista M ensile del CI. Alp. Ital., vol. xxxii. No. 7. * Storia delle Alpi Marittime, Turin, 1839. APPENDIX II M. CHAPPtns' Theory One more theory remains to be noticed, that of M. Chappuis. This writer is original only in the choice of his pass, following for the rest the hne of the Marquis de St. Simon and a ^vrite^ generally referred to as ' The Cambridge Anonjonous of 1830.' M. Chappuis set forth his views at length in a report addressed to the Minister of Pubhc Instruction, and pubhshed in the Revue des Societes Savantes des Departements, 2nd series, vol. iv. (1860).^ The text from Varro, which M. Chappuis treats almost as his own discovery, forms the basis of the argument. WTiile adopting the same general direction for Hannibal's passage through the Alps, he differs from me both as to the hne of march and the point of passage of the watershed. His pamphlet is long, and I can refer only to a few saUent points. The supposed difficult}^ about the Tricastini is got over by making Hannibal turn to the left into their country before reach- ing the Confluence. From Grenoble the Carthaginian army is led by a more westerly road to the Col de la Croix Haute, reach- ing the First Alps at Lalley. This is a suggestion possibly de- serving consideration, which, however, to be thorough, should be made with exact knowledge of the traces of ancient roads in this district. In his further progress we find M. Chappuis taking strange hberties. He takes Hannibal not over the Col de Vars, but up the lower gorges of the Ubaye. In order to reconcile Livy's statement that the Carthaginian army met -with no difficulties until after it had crossed the Durance, he postpones the first battle, placing it at Le Lauzet. To make this theory plausible, he is forced to treat the fifteen days assigned by Polybius for the passage of the Alps as covering ^ M. Chappuis has also written other works : Etude Archeologique et Geographique sur la Valli'e de Barcelonnette a Vepoque Celtique (Besan^on 1862) ; Examen Critique de V opinion de CcbHus Antipater sur le passage d'Annibal dans les Alpes, Paris, 1864 ; and Sentences de M. Terentius Varro, 1856. 107 108 HANNIBAL ONCE MORE only a portion of the 1200 stadia, which, according to the same authority, was the length of the passage, ignoring the fact that Polybius has fixed the number of days each section of the march took as precisely as he has the distances, and that there is no room for an interpolation between the ten days of the 800 stades and the fifteen of the 1200. Having got Hannibal on so far to begin with, M. Chappuis is naturally embarrassed to find employment for him in the days that follow. He reduces his marching pace to 11 m.p. a day. But even at this rate he gets too quickly to the watershed. So the poor elephants have to tramp on past the valley of the Ubayette and the low gap of the Col de I'Argentiere in search of the sources of the Ubaye and an obscure pass of over 9000 feet, apparently now impracticable for mules, known as the Col de Roure. Having crossed this, they are made to reach Val Vraita by a second high pass. Like most Itahan valleys, Val Vraita has its gorge, and M. Chappuis discovers a path by which Hannibal may try to avoid it. Why the great general neglected a low and easy pass for a high and difficult one, how his army managed late in autumn to spend three nights at such a height, are considerations which do not seem to have troubled our author any more than they did the Marquis de St. Simon. To adopt a homely proverb, M. Chappuis seems to me to ' have got hold of the right end of the stick,' but not to have known how to handle it. He fails to make allowance either for the formidable aspect the Alps wore two thousand years ago to a host of Africans, compared to that which they wear to us to- day, or for the further enhancement of the actual terrors of the passage in the course of literary description. Instead of being satisfied by a broad general resemblance between the narrative and the features of the country, he tries to reaUse and localise each sentence of Livy. Consequently, in his search for adequate difficulties, he selects a pass which he him- self tells us seems impossible unless allowance is made for changes in the ground having made it less practicable.* ^ ' Je dois dire que ce passage du Col de Roure au Vallon du Lautaret est un passage difificile, et qu'on se refusera tout d'abord a admettre que Tarmee d'Annibal ait pu le franchir, si Ton n'a pas v6cu au milieu des montagnes et etudi6 les prodigieux changements que les habitants eignalent de toutes parts.' APPENDIX 109 Add to this that he seems to be drawn by local interest to endeavour to concentrate as many incidents of the campaign as possible, or even more than is possible, in the valley of the Ubaye. At the same time M. Chappuis deserves credit for having been the first French writer to recognise the full weight of Varro's statement, and for having brought out forcibly some interesting illustrations of the classical texts. He shows, for instance, how the roads from Grenoble to the first pass (which- ever it was) he, not through an arid wilderness, but a rich tract of vineyards, and how the scenery suddenly becomes stem and sterile as the basin of the Durance is approached, and he states with local knowledge that Livy's description is exactly apphcable to the upper Durance, a fact which my own experi- ence fully corroborates. APPENDIX III The Roman Pass from Susa to the Maueienne In order not to interrupt the main argument, I have reserved for a separate note my reasons for citing the Mont Cenis among classical passes, and for identifying it with the fourth pass of Varro, assumptions for which, though the former had previ- ously been made by so respectable an authority as Cluverius, I have, doubtless, been taken to task in the minds of some readers. Many recent authors have held that the Mont Cenis was unknown to the ancients. The grounds assigned for this behef are the absence of any explicit reference to the pass by name in classical authors, the fact of its not being noted in such of the classical itineraries as have survived to our days, and the lack of Roman milestones in the Maurienne. The pass, it has been asserted, first appears in history in a.d. 754, as the Mons Cinesius of Pepin. Yet, if there is a want of direct proof of a Roman road on the pass itself, there is, in my opinion, a more than sufiicient amount of circumstantial evidence that the valleys of the Dora Riparia and the Arc were already in Roman times connected by a frequented route. If this fact can be estab- hshed, it seems to me highly improbable that the Roman and mediaeval tracks crossed the chain at widely different points. But should any one take a different view, I am not disposed to argue a detail which is for my purpose imma- terial. If the existence of a Roman pass from Italy into the Maurienne is granted, I shall be content to leave partisans of ' Cols ' varying from 500 to 4000 feet higher than the Mont Cenis to prove their case as best they can. As I have already admitted, although Roman remains have recently been found at St. Jean de Maurienne, no conclusive no APPENDIX 111 piece of local evidence, such as a milestone, has yet been produced for the Mont Cenis, Mr. Ellis,^ it is true, quotes Grillet, who, in his Didionnaire des Dipartements du Mont Blanc et Leman, speaks of a paved Roman road as still trace- able on the pass in 1807. Grillet's evidence is so far good that the fact he reported went against his own prepossessions, founded on the silence of the itineraries. But on the other hand, his statement remains, to the present day, isolated and unverified by Mr. EUis or others. The presumption therefore must be held to be against it, until, as is very desirable, some antiquarian possessed of the common sense ^ and technical knowledge necessary to distinguish between Roman and mediaeval road-work has examined carefully the remains of the old paved tracks in this region. Even, however, if such an examination does not lead to any discovery, there will be no ground for concluding that the pass was not in use in Roman times. For of other Roman roads in the Alps no local record, or only such a record as might easily have perished, remains. Etymologists offer us their aid in a series of derivations on which it is difficult, perhaps, for any one but an etymologist to lean with any confidence. They derive Venaus from Venatio ; from the Latin verb ferre, the name of the village Ferriera, where at the foot of the steep ascent from Italy burdens were transferred from carriages to men's or mules' backs. The name occurs frequently elsewhere in the Alps, e.g. on the route of the Col de I'Argentiere near the Barricades, and might more naturally be coimected with ferrum and iron mines. The roadside inns on the top of the pass are known as Le Tavernette (Tabemae). In Lanslebourg we are told to recognise Lancium, in Thermignon Interamnes, in Modane Ad Mutationem. ^ On the Ancient Routes between Italy and Gaul. Robert Ellis, 1867. 2 The mention of this qualification will not seem superfluous to those versed in Hannibalian literature. Witness the following extract. I purposely suppress the author's name : ' Les traces du camp y sont visibles ; on y retrouve meme le lieu oil furent places les elephants, la disposition singuliere et la dimension des levees de terre ne per- mettant que cette explication ; les silos pour le grain a la mode arabe, les abris creuses dans le sol pour se defendre de la bise, enfin le fameux rocher qu'Annibal fit eclater par le feu et deblayer avec Vaceto, le pic piemontain,' etc., etc. One is at a loss whether most to admire the ingenuity of a discoverer who outdoes Schliemann in his identifica- tions, or the acutenesa by which the vinegar is turned into a pickaxe ! 112 HANNIBAL ONCE MORE I turn from local to literary records. We have Caesar's own statement ^ that on the occasion of his sudden visit to Italy to fetch reinforcements against the Helvetii, he recrossed the Alps with his legions, 'qua proximum iter in GaUiam Ulteriorem erat,' by the shortest road — not, observe, to the Province, which he mentions in the next sentence, but to Further Gaul. Again he tells us that he touched the terri- tory of the Vocontii — so that he must have passed near Grenoble — and crossed that of the Allobroges, probably by the road which avoided Vienne, mentioned by Tacitus {Hist., ii. 66). His object was to overtake the Helvetii on their west- ward march before they could cross the Saone ; he made accordingly for the country of the Segusiani (' in quorum agro colonia Lugdunum,' Pliny, iv. 18), probably for the best bridge of the Rhone above its confluence with the Saone. We have Cicero's authority that the entire distance from Rome to the district of the Segusiani was 700 m.p. What are the distances by the various passes leading by the Vocontian frontier to Lugdunum ? Mont Genevre and Gap .... 758 m.p. Mont Genevre and Col du Lautaret . 704 m.p. MontCenis 730 m.p. Mont Cenis and Col de la Coche . . 705 m.p. We have no proof that either of these short cuts, the Col du Lautaret or the Col de la Coche, was known to the Romans at this date. In later times the Lautaret was crossed by a Roman road. On the Col de la Coche we are assured by local writers there are also traces of a Roman road, and M. Pilot (in his Histoire de Grenoble, p. 7) tells us that popular tradition associates it with Caesar.^ The appeal to distance can hardly be held satisfactory as between the last three routes. But it is at least conclusive ^ Caesar's Commentaries, book i. ch. x. Compare Napoleon i.'s La Route la plus courte. On this ground alone, declining to consider any literary evidence, Napoleon pronounced oflf-hand that Hannibal must have crossed the Mont Cenis. Much more would he have assigned this pass to Caesar, whose very words he unconsciously translated. * See J^tude de I Ancienne Voie Romaine de VOisans et de ses Annexes, par J. H. R. Grenoble, 1878. The author, a medical man, shows intimate local knowledge, and discriminates, apparently with authority, between Roman and mediaeval work. x^PPENDIX 113 against the first, that by the Mont Genevre and Gap, which Mr. Froude adopted with the remarkable expression, ' Caesar returned by the shortest route, over the Mont Genevre.' * For though the Mont Genevre may have been ' opportunius ' for Pompey when he was on his way to Spain, it does not in the least follow, as some have hastily assumed, that it was so for Caesar, marching in an opposite direction. We must look at the further indications of his hne of march given us by Caesar in the names of the tribes he encountered or through whose territory he passed. Among the mountains he was harassed by the Caturiges, Graioceh, and Centrones. Now, though the best kno\\Ti branch of the Caturiges Hved near the modern to-mi of Chorges, on the Durance, another portion of the tribe, according to Strabo, was found between the Saiassi and the Centrones. The Graioceli Uved in the Maurienne, the Centrones in the Tarentaise,^ and no mention has ever been found of portions of either tribe farther south. On the Mont Genevre and Gap route Caesar would have met the southern Caturiges ; on the Mont Cenis, the Graioceh and Centrones, and probably the northern Caturiges. Therefore, so far as the tribes serve us, the balance incUnes strongly to the Mont Cenis. For I can see no sufi&cient ground for the assumption of some critics that there could not be Caturiges in two places, and that therefore the ancient geographers who mention Northern Caturiges are ^^^:ong. There is good reason to beheve that in the Alps tribes bearing the same name, and probably branches of an original stock, were found widely separated. It is difficult to see on what grounds writers have argued as if in Gaul members of the same tribe never parted company, or at any rate never did so without assuming a new name. The Alps in Roman times were very much in the condition the Caucasus is in now ; and in the Caucasus frag- ments of the same tribe bearing the same name are found ^ Map-makers are very pliant ! Mr. Froude's map-maker got rid of the inconvenient Centrones and Graioceli, erased the Arc, and piled a mountain range across its bed. But he almost clears himself of any suspicion of having done this with an object, for he heis made the hills behind Gap the main chain of the Alps ! See Life of Julius Caesar. * The Maurienne was known as ' Garocelia ' (Oehlmann, Die Alpen- pdsse im Mittelalter) ; St. Jean, as Sanctus Joannes Garocellius (Grillet). H 114 HANNIBAL ONCE MORE at some distance from one another. But without going so far, our own country supphes a good example of scattered tribes. There were Belgae, Parisii, and Atrebates in Britain, as well as in Gaul ; Damnonii in Scotland, as well as Devon, Cornabii and Cornavii, Cantae and Cantii, in England and Scotland, Brigantes both in Ireland and England. For references and a full discussion of this march I must refer readers to Ellis, On the Ancient Routes between Italy and Gaul. Though I cannot accept all IVIr. EUis's arguments, I agree with him that Caesar crossed the Mont Cenis. I come to this conclusion on the following grounds : first, because the Mont Cenis is the shortest and best route from Italy to GaUia Ulterior, and Caesar was a hkely man to discover it ; secondly, because the indications of his route given by Caesar fit the Mont Cenis better than the Mont Genevre.^ Critics have rehed on a passage in Ammianus Marcelhnus ^ as evidence on the one hand that the Mont Cenis was an ancient pass, on the other that there was no pass across the Alps from Susa except the Mont Genevre. Gibbon found in the text a ' clear, livety, and accurate ' description of the Mont Cenis ; Law is equaUj' positive that the description apphes admirably to the Mont Genevre. To me it seems obvious that either the author or his text has fallen into confusion. The passage reads Uke a description of two passes mixed up together. Possibly some sentences have been lost before ' a summitate.' \Miatever maj'^ have been the real intention of the "waiter om' other information renders it impossible to infer from this passage that the Mont Genevre was the only pass known to the Romans in this district. We are clearly told that Cottius opened several routes.^ ^ Polyaenus, a writer of the second century, in his Stratagems of War furnislies some picturesque details as to Caesar's encounter with tlie mountain tribes. Noticing that the mists generally lay along the valleys in the early part of the day, Caesar profited by them to make unobserved a flank march, which brought his legions to a position above and commanding that of the enemy. The tribes finding their retreat threatened, fled incontinently, like Afghans, to their homes. 2 Book XIII. ch. X. ^ The inscribed vases discovered in the lake of Bracciano sliow that a second pass to the sources of the Durance was in use under the empire. Desjardins, Gaule Romaine, vol. i. p. 94. Desjardins' identi- fication of the second pass is contested by a recent writer, M. Vallentin, Les Alpea Cottiennes et Grates : Oeographie Gallo-Romaine. APPENDIX 115 The argument from the silence of the Itineraries may be broken dowTi by the instance of the Col de I'Argentiere and other passes, the roads over which, though unnoticed in these documents, are proved to have existed by local monuments. Many of the pages devoted to elaborate discussions of the road-map named after Peutinger have been wasted. For the writers assume as a basis for their argument that it is an exhaustive table, whereas — in the words of the late Sir E. Bunbury — it is a valuable but incomplete document. When we find many Alpine roads and passes used by the Romans in the Central Alps omitted in it, it becomes mere loss of time to argue on the supposition that in the Western Alps such omissions are impossible, or even improbable. There seems to be no evidence to show that Constantine, A.D. 312, crossed the Mont Cenis, as Gibbon beheved, rather than the Mont Genevre. But we have the clearest proof possible of the importance of the Maurienne, and of its close and intimate connection with the Cisalpine province, and particularly with Susa. We find that, from the earliest times of the Christian Empire, the Mauriemie was not only ecclesi- astically, but pohtically, attached to Turin. ^ It was not till the sixth century (562) that a new diocese of St. Jean de Maurienne was created, and in a.d. 596 the Pope complains on behalf of the Bishop of Turin that the latter has thereby been defrauded of his dues.^ The remonstrance seems to have been \\ithout result, and Susa also fell under the control of the new See. We find the clergy of St. Jean de Maurienne and Susa meeting under the authority of Abbo, the governor, to Neither author has been led by the name Turio, found in the list of stations, to identify the pass with the neighbouring Col des Tures, described at the beginning of the century as practicable for light artillery. 1 • La Geographie ecclesiastique d'un pays est, a tres peu de chose pres, la geographie de ce memo pays a I'epoque romaine. Le tableau des eveches, archeveches, est celui des civitates antiques selon leurs liens de subordination.' Renan, Marc-Aurele, p. 412 ; see also Freeman's Historical Geography. The Maurienne in a.d. 72G was a pagus or division of the district of Susa. The privilegium of that date, by which Abbo, Governor of Susa, gave the Alpe^ in Cenisio to the monastery of Novalesa, has been preserved. Vaccarone, Vie delle Alpi, etc. - See Gallia Christiana, vol. xv. The authors give reasons for abandoning the claims to a higher antiquity for the diocese, founded on the appearance of a Maurian or Marian bishop at Rome in a.d. 337. 116 HANNIBAL ONCE MORE found the great monastery of Novalesa in a.d. 726. And yet we have been asked to beheve that these prelates and their flocks had not discovered the Mont Cenis or any other direct route between their homes ! Proof is here afforded that the Mont Cenis was in common use from the introduction of Christianity. Can there be any- one who will still maintain that from the time of Caesar the Mont Cenis alone of all the great GalUc passes was unknown to the Romans ; that the citizens of Susa never employed their eyes to discover this conspicuous gap in the alpine chain ; that it was never brought under their notice by cowherds or native travellers ; that for generation after generation the legions garrisoned and adorned the town without ascertaining that one of the easiest passes across the alpine watershed lay within a morning's walk of their quarters ? Surely a proposi- tion so opposed to all we know of Roman character and enter- prise ought only to require to be stated in order to be dismissed. In Carlovingian times, the importance of the Mont Cenis was fully recognised. Pepin crossed it four times ; Charles the Great at least once, and probably more often ; Charles the Bald died in the Maurienne. We learn from one charter that in a.d. 825 a hospice belong- ing to the monks of Novalesa had already been estabUshed on the top of the pass. In another charter, the copy of which in evidence is dated (for 774 ?) 874, Charles the Great is made to confirm and enlarge the grants of his predecessors to the monastery, bestowing on it ' totam Novahciensem vallem et stratum.' This document, however, has been suspected on other grounds besides the date to be a forgery ; at any rate it is an early forgery, and the expression I have quoted from it deserves notice. I may add that in the Middle Ages the Mont Cenis was known as the Strata Romana. That it should have acquired this name — as Simler suggests — because it was the principal pilgrim's road to Rome is, of course, possible. But the use of the word strata, the origin of our own ' streets,' and gener- ally a token of a paved Roman way, points rather to the deri- vation of this distinguishing title from an old Roman road, the stratum of Carlovingian charters. To sum up the preceding argument, I maintain that, in the APPENDIX 117 absence of direct evidence to the contrary, there is a strong presumption that a pass so obvious as the Mont Cenis was used by the Romans from the time they held the country on both sides of it. I have further endeavoured to show that Caesar's summary of his forced march is best reconcilable with his having crossed the Mont Cenis ; and I have proved that the use of the pass can be carried back with certainty bej^ond the sixth century, since the close connection already existing between Susa and the Maurienne at that date is only explicable on the supposition that they had long before been united by a direct and frequented high-road. APPENDIX IV Mr. Maeindin on Fuchs and Osiander I MAY be thought to have unduly neglected works to which much importance has been attached by some critics abroad, Fuchs's HannihaVs Alpeniihergang, Vienna,, 1897, and Osiander's Der Hannibalweg, Berhn, 1890. They do not appear to me to add materially to the argu- ments with respect to the Mont Genevre and Mont Cenis. I may cite in support of this view some sentences from the discussion of these treatises by Mr. Marindin in The Classical Review. Mr. Marindin, who writes with authority on matters of scholarship, sweeps aside as untenable many of the nice verbal distinctions insisted on by the German writers. Of Osiander he writes : ' This is a treatise full of industry and abundant in citation of authorities, but it adds nothing of any importance to the controversy, and it is to be feared only tends to obscure the issue. There is in the whole book a want of logical argument and a failure to appreciate the real value of whatever evidence is brought forward at each step.' ' Professor Osiander deserves all credit for the industry and research of his treatise, and it is all the more to be regretted that he leaves the controversy just where he found it. He has neither strengthened the case of the Mont Cenis nor weakened that of the Mont Genevre.' While holding that 'Dr. Fuchs (in accepting the Mont Genevre) comes to a right conclusion in the end on the main question, and that many of his arguments are true and valu- able,' Mr. Marindin strongly condemns his verbal criticisms. Dr. Fuchs, he holds, strains the Greek language, unduly magnifies the significance of certain Greek words, and mis- translates others. APPENDIX V A Short Bibliography of Recent Publications Arnold, W. T., The Second Punic War. Macmillan & Co. : London, 1886. This work is thus referred to by Mr. E. S. Shuckburgh in his translation of Polybius (Macmillan, 1889) : ' I have no intention of re- discussing the famous question of the Pass by which Hannibal crossed the Alps. The reader will find an admirably clear statement of the various views entertained, and the latest arguments advanced in favour of each, in the Notes to W. T. Arnold's edition of Dr. Arnold's History of the Second Punic War.' Azan, P., Annibal dans les Al]}es. Paris, 1902. Brockedon, W., The Passes of the Alps., vol. ii. Murray : London 1828-9. Capes, Livy, Books xxi., xxii. Macmillan : London, 1883. Chappuis, C, Bajiport au Ministre de I' Instruction PuUique sur le passage d' Annibal dans les Alpes. Paris, 1860. Annihal dans les Alpes. Fallier : Grenoble, 1897. Colin, J., Annibal en Gaule. Paris, 1904. Coolidge, W, A. B., Rivista del Club Alp. Italiano, xxx. 6. Dodge, T. A. Hannibal, 2 vols. Houghton, Mifflin & Co. : New York (no date). Diibi, Die Bomerstrassen in den Alpen. Jahrbuch des Schweizer Alpenklub, vols. xix. and xx. Dalp : Bern, 1884-5. Ferrand, H., Recherches du Col des Alpes franchi par Annibal, Casale, 1912. Revue des Etudes Anciennes, vol. x. No. 1 ; vol. xiii. No. 3. Freshfield, D. W., ' The Pass of Hannibal.' Aljnne Journal, Nos. 82 and 93. Longmans, 1883 and 1886. Proceedings R.G.S., n.s., vol. viii. ; Geographical Journal, vol. xxxvii. Nos. 4 and 6. 1911. Fuchs, J., Hannibal's Alpeniibergang. Vienna, 1897. Hennebert, Histoire d'Hannibal. Imprimerie Nationale, Paris, 1870 and 1878. Heselmeyer, E., Hannibal's Alpeniibergang irn Lichte der neueren Kriegsgeschichte, 1906. Marindin, G. E., The Classical Revieiu, vol. xiii. No. 5 ; vol. xiv. No. 5. 119 120 HANNIBAL ONCE MORE Neumann, Das Zeitalter der Punischen Kriege. Kobner : Breslau, 1883. Oehlmann, Die AlpeniMsse im Mil t el alter, Jahrbuch fur Schweizerische Qeschichte, vols, iii. and iv. Zurich, 1878 and 1879. Osiander, W., Der Hannibal weg. Berlin, 1900. Perrin, Marche d'Annibal des Pyrenees au Po. Paris, 1883. Schiller, Ueber den Stand der Frage, welchen Alpenpass Hannibal beniltzt habe. Berliner Philologische fVochenschrift, vol. iv., 1884. Strachan-Davidson, J. L., Afpian, Book I. The Clarendon Press, Oxford, 1902. Wilkinson, S., Hannibal's March through the Alps. The Clarendon Press, Oxford, 1911, and Geographical Journal, vol. xxxvii.. No. 6. Vaccarone, Le Vie delle Alpi Occidentali. Candeletti : Turin, 1880. For classical references see Cluverius, Italia Antiqua, c. xxxiii. For fuller lists of older works, see Fortia d' Urban, Dissertation sur le passage du Rhone et des Alpes. Paris, 1821. (Brockedon's copy is in the E.G-.S.'s library.) For a list of more recent works, M. Azan's book quoted above. Professor Wilkinson also recommends ' a full account of the literature given by M. Camille JuUian in the first volume of his Histoire de la Gaide.' Printed by T. and A. Constable, Printers to His Majesty ftt till' Edinburgh University Press Telegrams : ^i and 43 Maddox Street, " Scholarly, Reg. London." g^^^ 5^^^^^^ ^^^^^^^ ^ XT '^oc^^'ilf V- February, 1914. No. 1883 Mayfair. J-' v t Mr. Edward Arnold's SPRING ANNOUNCEMENTS, 1914. KULU AND LAHOUL. an account of m^ Xatest Climbina ^ourne^s in tbe fbimala^a. By LIEUT.-COL. THE HON. C. G. BRUCE, M.V.O. 6th Gurkha Rifles; Author of "Twenty Years in the Himalaya." With Numerous Illustyations from the Author's Photographs. One Volume. Demy Svo. 12s. 6d. net. It has been said that no one can ever hope to know more than a few corners of the Himalaya. Yet an exception must surely be made in favour of Colonel Bruce. Outside Nepal there is no considerable section of the range, from Sikkim to Chitral and Hunza, with which he has not some personal acquaintance, while of Nepal itself he knows as much as is possible for a European. This has been shown in his valuable earlier work, " Twenty Years in the Himalaya," in which he describes his varied experi- ences as a mountaineer and an officer of the Gurkhas, including ascents in the company of Conway, Mummery, Longstaff, and other well-known climbers. In his latest expedition Colonel Bruce filled up his most con- siderable outstanding gap by a thorough exploration of the mountains of Kulu and Lahoul, which have hitherto been neg- lected by climbers. The country visited lies nearly due north of Simla, in the centre of the group of small states which intervene between Garhwal and Kashmir, and a very large amount of most interesting ground was covered in the six months spent there. The record is not a mere tale of mountaineering achievement, for the author treats his subject lightly, and at the same times brings his wide knowledge of native life into play in describing the history and folklore of the two districts. Mrs, Bruce contributes a chapter on " A Lady's Point of View." LONDON : EDWARD ARNOLD, 41 & ^ MADDOX STREET, W. 2 My. Edward Arnold's Spring Announcements. A CLIMBER IN NEW ZEALAND. By MALCOLM ROSS, A.C., F.R.G.S. With numerous beautiful I llust rations. Demy 8vo. Probable price, 15s. net. A work dealing with climbs and exploration in the New Zealand Alps. As one of the founders and Vice-President of the New Zealand Alpine Club, and editor of its journal, the author is well qualified to write on the subject of climbing and mountain exploration in his native land. The charm and adventures attendant upon exploration in the Alps of a new country are most graphically described. Indeed, there is not a dull page in the whole book, while some of the adventures were of quite a thrilling nature. The book is some- thing more than a mere record of ascents. From beginning to end it has quite a literary flavour, and there is, also, running through its pages a quaint vein of humour and philosophy. Such chapters as " Above the Plains" and " An Interlude " will appeal to a wide circle of readers. Mrs. Malcolm Ross, a well-known New Zealand authoress, comes into the book a good deal, and her amusing descriptions of camp cookery and of the Kea, or mountain parrot, add to the interest of the work of her daring and talented husband. The author has illustrated his book with some very excellent and artistic pictures from his own photographs. HANNIBAL ONCE MORE. By DOUGLAS W. FRESHFIELD, M.A., Vice-President of the Royal Geographical Society ; Treasurer of the Hellenic AND Roman Societies; formerly President of the Alpine Club; Author of "The Exploration of the Caucasus," "Round Kinchinjdnga," etc. Illustrated. Svo. 5s. net. In this little volume Mr. Freshfield has put into final shape the results of his study of the famous and still-debated question : " By which Pass did Hannibal cross the Alps ? " The literature which has grown up round this intricate subject is surprisingly extensive, and various solutions have been propounded and upheld, with remarkable warmth and tenacity, by a host of scholars, historians, geographers, military men, and mountaineers. Mr. Freshfield has a solution of his own, which, however, he puts forward in no dogmatic spirit, but in such a fashion that his book is practically a lucid review of the whole matter in each of its many aspects. To an extensive acquaintance with ancient and Mr. Edward Arnold's Spring Announcements. 3 modern geographical literature he unites a wide and varied ex- perience as an alpine climber and a traveller, and a minute topographical knowledge of the regions under discussion ; and these qualifications- — in which many of his predecessors in the same field of inquiry have been conspicuously lacking— enable him to throw much new light on a perennially fascinating problem. THE LIFE OF ADMIRAL SIR HARRY RAWSON, G.C.B., G.C.M.G. By LIEUT. GEOFFREY RAWSON, R.I.M. With Illustrations. Demy 8vo. 12s. 6d. net. This book, which has been written by a relative of the late Admiral Sir Harry Rawson, gives a deeply interesting account of the Admiral's long and varied career in China, the Near East, Africa and Australia. Entering the Navy as long ago as 1857, when the sailing ship still held the seas, Sir Harry, as a young cadet, thirteen years of age, took part in the China War of 1858-60, being present in the Calcutta's launch at the capture of the Taku forts. In 1863, when only nineteen years of age, he held Ning-po for three months against the rebels, with 1,300 Chinese troops under his command. He served for two commissions as Commander of the battleship Hercules, and acted as Flag-captain in both the Channel and Mediterranean Squadrons. During the Egyptian Campaign of 1882, he acted as principal Transport officer, and after his promotion to Flag rank, was appointed Commander-in-Chief on the Cape Station, 1895-98. Whilst holding this important position, he landed the Naval Brigade which captured Mweli, the stronghold of a rebel Arab chief; part of his squadron bombarded the palace at Zanzibar; and, in 1897, ^^ commanded the famous expedition for the capture of Benin, which was successfully accomplished. For more than two years he commanded the Channel Squadron, and the work contains a chapter devoted to the naval manoeuvres of 1900. The book terminates with a well-told account of Sir Harry's Governorship of New South Wales, and a chapter on his short- lived but no less famous brother. Commander Wyatt Rawson, who led the historic night march over the desert to Tel-el-Kebir. 4 Mr. Edward Arnold's Spring Announcements. RICHARD CORFIELD OF SOMALILAND. By H. F. PREVOST-BATTERSBY, War Correspondent of the "Morning Post" in South Africa and Somalii.axd. Illustrated, Demy Svo. It was the news of his death that made the world acquainted with Corfield's name, but everyone Avho had served with him knew him as a man of extraordinary charm, with some secret source of power which had an amazing influence over savage peoples, and regarded a notable future as assured to him. He had gone as a mere boy to the South African War, serving afterwards in the South African Constabulary. He had just returned home on leave from his five years' service when he was appointed to the new Tribal Militia being raised in Somaliland. He remained there till the country was evacuated, and the militia disbanded, in 1910, acquiring a remarkable reputation for hand- ling the most " difficult " natives in Africa. He was at once transferred to Northern Nigeria, and after a year's work, signalized by striking successes with the pagan tribes, was specially requisitioned to raise and command a force of camelry, to which the governance of Somaliland was to be entrusted. In nine months he performed miracles in a country literally reeking with blood, 100,000 of the inhabitants having been massacred since the country was evacuated, and restored in that short time the confidence of the natives in the honour of the British Government. He lost his life in a gallant attempt to check a Dervish raid, which was spreading fresh ruin over the country ; but those who read his life will discover that his action touched a much deeper issue than the succour of a starving people. He died for England's old ways of honour — for the keeping of faith to unfor- tunate dependents, for his country's good name. TEACHING FOR LADS. 3For use in Bible Glasses ant) Confirmation Classes. By the REV. PETER GREEN, M.A., Rector of St. Philip's, Salford, and Canon of Manchester; Adthor of " How to Deal with Lads," " How to Deal with Men." Crown 8vo. 2s. 6d. net. This book is intended primarily for the use of those who have to give instruction to boys in Sunday Schools, Church Lads' Mr. Edward Arnold's Spring Announcements. 5 Brigades' Bible Classes, and similar institutions, and has been written in response to many and constant inquiries from people desirous of help in such teaching. The course of teaching out- lined is based upon what the author has himself used with his own classes. Canon Green has the exceptional faculty of com- bining a knowledge of philosophy with a wide experience in actual practice. He is a theorist who has really put his theories to the test. In his two previous works there has been found a freshness about his whole treatment of the subject that is most stimulating, and it may be confidently asserted that readers of the new volume will derive equal pleasure and profit from its perusal. THE ORIGIN OF ATTIC COMEDY. By F. M. CORNFORD, Fkllow and Lecturer of Trinity Collkge, Cambridge; Author of "From Religion to Philosophy," "Thucydides Mythistoricus," etc. One Volume. Demy ^vo. 8s. 6d. net. This book tries to show that the peculiar structure and features of Aristophanic Comedy can be explained by an hypothesis like that by which Professor Gilbert Murray explains the form of Greek Tragedy. A detailed examination of the extant plays brings to light, under all their variety, a plot-formula composed of a regular sequence of incidents : Agon, Sacrifice, Feast, Marriage, Komos. The theory is that this canonical sequence presents the stereotyped action of a folk-drama (like the English Mummer's Play), which itself preserves the outline of a well- known seasonal ritual. This ritual turns out to be substantially identical with that which, according to Professor Murray, lies behind Tragedy. The stock masks in the Old Comedy are shown to be the essential characters in the supposed folk-drama. On this view, various features of Aristophanic Comedy, such as the Agon and Parabasis, about which widely different views are now held, can be satisfactorily explained. It follows also that Aris- tophanic Comedy is not a patchwork of heterogeneous elements, but a coherent whole, and in all probability a native growth of Attic soil. Athenian Comedy and Tragedy have a common source in ritual of the same fundamental type. The causes of their differentiation — a problem of great interest in the history of drama — are treated in the last chapter. The book is intended, not only for classical scholars, but for all readers interested in the origin and growth of the drama. Greek has been confined to the footnotes, and a synopsis of the plots of Aristophanes' plays has been added, to enable the ordinary reader to follow the argument. 6 Mr. Edward Arnold's Spring Announcements THE REVIVAL OF THE RELIGIOUS LIFE. By FATHER PAUL BULL, Of the House of the Resurrection, Mirfield. Cvown 8vo. ^s. 6d. net. The purpose of the author of this volume is to give an outline sketch of the ideals of the Religious Life, and the attempts to realize that life in history, and to trace the various manifestations and interpretations of the ascetic and mystic spirit of the Gospel. Father Bull has spent twenty-two years of his life in trying to realize these ideals, and feels that both the infinite joy of the attempts and also his consciousness of abysmal failure may have qualified him to offer an opinion on the glory and the perils of the Religious Life — the word " Religious " being here used in its original technical sense for a corporate life under rule. In urging the need for a fuller revival of the Religious Life for men in the English Church, the author deprecates mere imitation of what has been done in past times and in other lands, and considers it im- portant to accumulate as soon as possible a body of experience, to discuss principles and see what adaptations of these principles are necessary to meet modern conditions. The book should prove of great interest and utility to all who are concerned with the relisfious tendencies of our times. ESSAYS ON FAITH AND IMMORTALITY. By GEORGE TYRRELL. Arranged, with Introduction, by M. D. PETRE, his Biographer. One Volutne. Crown 8vo. 5s. net. This volume comprises, for the most part, matter hitherto un- published which existed in the form of notes and essays amongst the MSS. which Father Tyrrell left behind him. To this have been added a few articles already published, either in England or abroad, in various periodicals, but which are not now easily obtainable, and which possess a certain importance. One of these latter essays, in particular, has attained considerable celebrity, and is yet almost unprocurable — namely, the one entitled " A My. Edward Ar)wld's Spring Announcements. 7 Perverted Devotion," which played so large a part in the drama of its author's life. The essays have been divided into two parts, of which the first may be said roughly to deal with problems of faith, the second with the ever-recurring problem and mystery of personal immor- tality. This second part is specially characteristic of George Tyrrell's mind, with its spiritual tact and delicate power of intuition. Some of the many to whom this question is the ques- tion of all questions will find in these essays, not a scientific answer to their difficulties, but illuminating hints and suggestions to help them in their own search. The essays, in fact, as a whole, are not definite treatises, but the musings and gropings of a deeply spiritual mind in its quest of truth. NEW NOVEL BARBARA LYNN. By EMILY JENKINSON, Author of " Silverwool," "The Soul of Unrest," etc. Crown d>vo. 6s. In " Barbara Lynn " the author returns to the English Lake District — the scene of her novel " Silverwool " ; and from the first page, where the reader is taken to the " Lonely Steading in the Dale," right on to the end, he is not only surrounded by the mighty hills and vast solitudes of Nature, but made to feel the pulse of romance and tragedy. " Barbara Lynn " shows a distinct advance in the work of this gifted young writer. The style, as in her earlier books, is powerful, individual, lucid. There is the same love of wild and majestic scenery, and power to convey impressions of it to others ; the same knowledge of human character and of sympathy with country folk, and capacity to enter into their lives and experiences. But with these gifts, which have already given distinction to Miss Emily Jenkinson's work, we hear a deeper chord struck in " Barbara Lynn." It comes out especially in the presentation of the character of the girl whose name gives the title to the story. Barbara Lynn is a splendid creature ; and the picture of her spiritual victory, her mysticism and self-sacrifice, is something unique in current works of fiction. In our judgment this is a book, not merely to delight readers for a season, but to remain a notable contribution to our literature for a long time to come. 8 Mr. Edward Arnold's Spring Announcements. Jmportant Books IRcccntl^ 356uet). LORD LYONS : A Record of British Diplomacy By LORD NEWTON. 2 vols. 30s. net. "A great biography, a classical record of the career of one of the greatest Ambassadors. " — Saturday Review. Life and Letters of George Villiers, FOURTH EARL OF CLARENDON. By Sir HERBERT MAXWELL, Bart. 2 vols. 30S. net. " Not since Lord Morley's ' Life of Gladstone ' was published have we had so i tribution as this book to the history of the Victorian era." — Daily Chronicle. PRIMATE ALEXANDER: Archbishop of Ar Edited by ELEANOR ALEXANDER. 12s. 6d. net. Third 1. " A book that is most likeable, besides having a great charm that the reader wJ appreciate, even when he cannot exactly define its quality.' — Spectator. WILLIAM AUGUSTUS, DUKE OF CI LAND : His Early Life and Times, 1721-. By the Hon. EVAN CHARTERIS. i2S. 6d. net. " Mr. Charteris has the gift of making his characters live ; he is at once vivid ai His analysis of Cumberland's military campaigns are especially lucid." — Daily Mai A CIVIL SERVANT IN BURMA. By Sir HERBERT T. WHITE, K.C.I. E. 12s. 6d. net. "It contains some of the best descriptions of Burma and the Burmese people met." — Outlook. SPLENDID FAILURES. By HARRY GRAHAM. los. 6d. net. " 'We should imagine that few, even among well-read men and women, cc volume without learning something they did not know before." — Times. THIRTY YEARS IN KASHMIR. By ARTHUR NEVE, F.R.C.S. With Illustrations. 12s. "Another of the splendid contributions to geographical and ethnological scie alone missionary endeavour would be more than justified." — Outlook. SHAKESPEARE'S STORIES. By CONSTANCE MAUD, Author of " Wagners Heroes," MAUD. Illustrated. 5s. net. New and Revised Edition. PAINTING IN THE FAR EAST. By LAURENCE BINYON. With 40 Plates. Crown 4to. LONDON : EDWARD ARNOLD, 41 & 43 MADDOX STR '5iT/ OF CAl-IF PNi' N'l^ELES University of California Libraiy Los Angeles f^j£liisbook^ DUE gndie last date stamped below. HB-R^rfiewals 310/82^-9188 RtC'U LD-URl 81' ■^c^ 3 1158 00415 4828 tOS A u! 3H SCHOO^