Chapter VI: Preface: To the Fourth Edition (4)

First, the map is on too small a scale to show a certain number of features, which, though unimportant in the general landscape, are essential to the traveller on foot. This is true of rocks, for instance; open rock, extending over a considerable surface, will always be marked, but hidden ledges, especially small ones, are more often not marked, and this may lead to disaster if one trusts the map too exactly. For instance, in the sheet numbered xi. 37, a range will be seen rising to the left of the main road, which bisects the map from north to south: I mean the range running from the Spanish frontier to the Pic-du-Ger. This ridge is intersected by two profound valleys, and the whole of it is a mass of greater or smaller limestone ledges, more or less masked in the density of the forests. Yet it is impossible to indicate these on such a scale, save here and there by sharp hatching. These limestone ledges are in this particular case such, that unless one knows the paths extremely well, it is impossible to cross the ridge at all, but one would have no idea of that from merely consulting the map. On the other hand, every rivulet, however small, is distinctly marked, and that is something of a guide when one has tried to ascertain one’s position in a valley. This map has a further advantage of marking in the clearest way the paths by which the various ports are approached, and after a considerable use of it in many places, I can say that when you have lost the path, the indication afforded you by the 1/100,000 map is invariably right—upon the French side. However unreasonably the line seems to acting upon the map, if it lies to the left of a stream, or beneath a particularly clearly marked rock, then it is to the left of that stream, or beneath that rock that you must cast about if you want to find it, and if you find another path in another direction, you may be certain it is but a random track, which will mislead you, however clearly it may appear for the moment. When, in first using these maps, my companions and I neglected such information, it invariably led to trouble. For instance, in the lower crossings of the Sousquéou, the map gives the path everywhere on the north, or right bank of the stream. There is a spot just before the first rocky “gate” of this ravine where all indication of further travel upon the right bank disappears, and on the contrary a fine-made path crosses over by a strong bridge to the further or left bank. We thought the map must be in error, and crossed by the bridge, with the result that we spent a whole day cut off by a bad spate from the further side, and were for some hours in peril; for the bridge once crossed, this false path disappeared within half a mile. If we had pinned ourselves to the map, kept to the north bank, and cast about in circles, we should have found the path again but a hundred yards or so further on, running precisely as it was indicated on the survey. The importance of the 1/100,000 map in thus giving all tracks accurately will hardly appear to the reader unused to the Pyrenees, but it will be seen clearly enough when we come later to speak of travel upon foot in the mountains.

It is a defect of the 1/100,000 map that heights, though accurately marked, cannot always be as accurately referred to the exact spots standing near the figures. This is because the heights are marked in pale blue ink, and the ambiguity is accentuated by that absence of contour lines which is the chief fault of the series. The method of marking is to point a small blue point close to the figures, and this dot marks the exact spot to which the figures refer. Where the figures are printed in a white space, and where there are no other features to interfere with them, this small blue spot is plain enough, but where they come upon woodland or steep shading, or other print, it is almost impossible to discover the dot. Thus, for instance, in the xi. 37 sheet to which allusion has just been made, a little lake will be found right upon latitude 42° 50′, just before its intersection with longitude 2° 40′. The height of this lake is given as 2170 metres, and the small blue point to which that altitude exactly refers is unmistakably marked at the southern extremity of the lake; but immediately to the right of those very figures, one of the highest peaks of the Pyrenees, the Bat Lactouse, marked 3146 metres, presents no point of which one can be certain. The frontier happens to cross this peak, and the little blue spot has got lost in the chain of black dots marking the frontier and in the print of the name of the mountain.

As a general rule, however, if you are in doubt as to what a figure may refer to, you are pretty safe in referring it to a peak, rather than to a pass or a group of houses in the neighbourhood. I have said that the accuracy of the map is undoubted for the French side; it is less certain upon the Spanish, where indeed its accuracy is not guaranteed. It is the best map to use upon the Spanish side (save for that restricted district over which Schrader’s contour map applies), but do not, upon the Spanish side, take the map against the evidence of your senses, as you will be wise always to do upon the French side. The map is notably wrong upon the Spanish side where unfinished works are concerned; it is not revised with the same frequency and care as upon the French side; for instance, the big new road from Sallent up to the French frontier goes in long winding zigzags, which make the total distance between eight and nine miles. The 1/100,000 map marks it in dots as though it were not finished, makes it far straighter than it is, and thus reduces the distance by nearly half.

Finally, the 1/200,000 map gives the best bird’s-eye view of the whole district, and is the only one showing contours, and penetrating further upon the Spanish side than any other. It will be my advice to those who desire to take a walking tour of some length in various parts of the range, to equip themselves with the whole set of the 1/200,000 maps (5 sheets), with the whole of the 1/100,000 map, but only with such of the 1/800,000 (the uncoloured map of the Ministry of War) as cover small districts of the nature of which one is in doubt. Those, on the other hand, who purpose spending their time in one or two valleys only, should, without fail, purchase the sheets of the 1/100,000 survey covering that district, and would do very well to add to these all the corresponding sheets of the 1/80,000 survey.

With these remarks, most that can be usefully told to my readers with regard to the maps of the Pyrenees has been told them, but perhaps a few final notes will not be without their use, thus: The English traveller must always remember that none of these maps comes up to the English one-inch Ordnance for accuracy and detail—the scale forbids this. Next, let him remember that the dates of revision of each map will differ, as do the dates of revision of ordnance maps in every country. For instance, I have before me, as I write, the 1/200,000 of Luz, purchased in this year (1908); no date of revision is attached to it, but the new road (which is at present an excellent carriage road, one of the best in Europe, up the Gallego to the French frontier) is marked, at first as a lane, afterwards as a mule track. On the 1/100,000 (Laruns sheet), purchasable this year, the new road is marked as existing for traffic, but not fully completed beyond a point about three miles from the frontier, and its true form is not given but merely indicated. It is evident that these sheets were revised at different times (the Laruns sheet bears a date six years old), and that we must always take the later of any two impressions, if we can obtain it. The highways of the Pyrenees upon the French side especially, both by road and by rail, are being extended with such rapidity that every year makes a difference to the accuracy of the information conveyed.

It remains to enumerate with their titles the maps covering the district: in England they may be most easily obtained from Messrs. Sifton & Praed, The Map House, St. James’s Street. This firm provides the 1/200,000 for the whole chain of the Pyrenees range mounted on canvas, the most useful map perhaps for motoring and cycling. Any sheet of the 1/100,000 can also be obtained from them, as all are kept in stock, but by far the most convenient form in which to carry them is to have them folded in the stiff cover issued by the French Government: to get them in this form, a few days’ notice in London will be needed. From the same firm the military maps can be procured in a similar manner, but I do not know whether all are kept in stock as a regular thing.

In ordering the sheets of the 1/200,000 (if one does not purchase them as a whole), reference is made not to numbers, but to names. There are five sheets, “Bayonne,” “Tarbes,” “Luz,” “Foix,” and “Perpignan,” the price of which in England is 10_s._; the whole series can also be purchased mounted.

The sheets of the 1/100,000 map may be referred to either by the names of their central towns, or by the index number of the series in which they are printed. It is difficult to say what numbers of these maps exactly cover the range, unless one knows how far from the watershed towards the plain the traveller intends to go. The smallest number sufficient to cover the actual watershed and the highest peaks is 16, or, for the whole frontier, 17. These sheets are by name (going from the Atlantic to the Mediterranean, from west to east), St. Jean-de-Luz, Bayonne, St. Jean Pied-de-Port, Mauléon, Ste. Engrace, Laruns, Luz, Gavarnie, Bagnères-de-Luchon, Val-d’Arouge, St. Girons, Mont Rouch, Perles, Ax-les-Thermes, Saillagouse, Ceret, and Banyuls. Referring to their numbers in the series upon the index map, they are respectively viii. 35, ix. 35, ix. 36, x. 36, x. 37, xi. 37, xii. 37, xii. 38, xiii. 37, xiii. 38, xiv. 37, xiv. 38, xv. 38, xvi. 38, xvi. 39, xvii. 39, and xviii. 39. It will be observed that in the index map of the 1/100,000 series, the divisions running from north to south are marked in Roman numerals, those from east to west in Arabic numerals, and that the gradual increase in Arabic numerals from 35 to 39, corresponds to the gradual trend southward of the Pyrenean chain from the Atlantic to the Mediterranean.

Very few of my readers will be concerned with the main crest of the range alone; it will therefore be necessary to add to that list northward of the frontier (the lower Arabic numerals) the further sheet according to the district each may have chosen to travel in. A certain number of extra sheets are necessary to those who travel in the main chain only, for instance, “Perles” (xv. 38) includes within the limits of its sheet the frontier upon either side, but this frontier so nearly approaches the northern limits in one spot, that it will be quite impossible to travel in this part until we also add the sheet “Foix” (xv. 37), to the north of it. Even the little lake of Garbet, which is not three miles from the crest of the range, is half out of the map and half in.

Those who desire a complete collection of all the sheets of the 1/100,000 survey, extending from the furthest mountain over the Spanish side up on the foothills into the French plain, may remark the following lists: in series viii. 35; in series ix. 35 and 36; in series x. 35, 36, and 37; in series xi. 35, 36, and 37; in series xii. 36, 37, and 38; in series xiii. 36, 37, and 38; in series xiv. 37 and 38; in series xv. 37 and 38; in series xvi. 37, 38, and 39; in series xvii. 38 and 39; in series xviii. 39; in all twenty-five sheets will cover the mountainous region in this survey, and anyone who desires a complete map of the French Pyrenees, with as much of the Spanish side as the survey includes, should possess them all.

Schrader’s map is in six sheets upon the scale of 1/100,000 and with contours. It is essentially a climber’s map. Detailed maps of special districts of course exist in many shapes, but they must be sought for in the periodical reviews, and in monographs in which they have appeared. Finally, it may interest the reader to know that in the Casino of Bagnères-de-Luchon he may inspect a fully detailed relief map of the whole range on a scale somewhat larger than one inch to a mile, though the inspection of it rather satisfies curiosity than affords any guide to travel.

Schrader’s map is of the greatest value for one particular piece of touring, which I shall describe later in these pages. Meanwhile it may be as well to add a further note upon it here. It is by far the best, so far as it goes, of all the Pyrenean maps; it is due to private enterprise, and if the whole range had been done in the same way there would be no need to discuss any other type, it would amply suffice for all purposes. Unfortunately, whereas the range, within the limits laid down in this book, stretches in length from a degree east of Paris to nearly four degrees west of that meridian, covering, that is, four or five degrees of longitude, and stretches in latitude from 43° 25′ to at least 42°, Schrader’s survey covers only 1½° in longitude (namely, from 1° 10′ west of Paris to 2° 40′), and in latitude extends over no more than half a degree, namely, from 42° 20′ to 42° 50′.

As the reader may see by comparing these bearings with a general map, Schrader’s map is intended to include no more than the very high Pyrenean peaks: it is the result of many years of careful individual survey, begun before the war of 1870 and carried on to quite the last few years.

Like the French Home Office map, it is in the scale of 1/100,000, and, like it, it is printed in colours, but unlike the Home Office map, it shows the invaluable feature _of contours_. You have an exact plan of the country before you, and in clear weather, with the aid of this map, you can fall into no error in connexion with the relief of the land. The contours are at some distance, at 100 metres or 328 feet apart, but this in such country is an advantage; indeed, the cramping of the closer contours on the official 1/200,000 map, greatly detracts from their usefulness. Not only are contours marked, but all rocky places are given with the greatest care, and the impression of relief is helped by shading as well as contour lines. The only drawback of the map, apart from its restricted area, lies in the absence of any indication of woods. As to the steepness, to which woods are often a guide, his contours amply make up for the deficiency, but for camping it affords you no indication. On the other hand, all cabanes and all paths are very clearly marked.

All heights and distances with which you will have to do in these hills upon either side are marked in metres, save in the popular talk, which measures distances by the time taken to traverse them. With this I shall deal in a moment. Let me first deal with what is a constant source of trouble to Englishmen on the Continent, the turning of the metrical system of measure into its English equivalent.

There are two ways of doing this. One is the application of quite easy and rough rules of thumb, the other is the more complicated process which aims at a fairly high degree of accuracy. It is the first of these of course which most people will want to know, and there are two simple rules, one for heights and one for distances.

The rule for heights is, divide by 3, shift the decimal point one place to the right, and you have the height in English feet, _within a certain limit of error_, which I shall presently detail.

The rule of thumb as applied to measures of distance is to take the number of kilometres (a kilometre is 1000 metres, and is, as one may say, the French mile), divide by 8, and multiply by 5, and you have the corresponding number of English miles _within a certain limit of error_, which I shall describe presently.

For all ordinary purposes these two rules are sufficient, though in both cases they somewhat exaggerate. They make a French distance measured in English miles a little too far, and a French height, measured in English feet, a trifle too high.

The exact constant of error is, in the case of the heights, 1.6 feet in every 100. Thus if your rough calculation gave you a height of 10,160 feet, the exact height ought to be just 10,000; you see upon the map in the blue figures referring to metres, “3048” (which happens by the way to be within two steps of the height of the Bac Lactous). You divide by 3, add a 0, and get 10,160, and you know by the constant of error that the true height is just exactly 10,000 feet.

The knowledge of this constant gives us a rough-and-ready method of getting a height within a very small degree of accuracy, and for any purposes where such accuracy is required, I recommend it. It consists in cutting off the last three figures, multiplying what is left by 4, and then again by 4, and subtracting that from your first rough calculation. It sounds complicated, but it does not take half a minute, and you will be well within two feet of any height; for most heights you are likely to calculate, you will be right within a few inches.

For instance, you see 2403, in blue figures upon the map dividing by 3 and shifting your decimal point, you at once get 8010; there is your rough calculation, which you know to be a trifle in excess of the truth. Cut off the last three figures and you have left 8, multiply 8 by 4, and then again by 4, and you have 128 as the amount of your error. The peak is by this calculation 7882 feet high, and rough as the rule is you are within 20 inches of the truth: the exact height of such a peak in English feet is 7883.7624....

However, if you want absolute accuracy, multiply the French measure by 3.2808992, and you will be sufficiently near the truth to save your soul.

As to distances, the exact proportion of error, when you turn miles into kilometres by dividing by eight and multiplying by 5, is 2 inches or so short of 50 feet too much in every mile; when, therefore, you are dealing with a hundred miles, you are very nearly, but not quite a mile out in this form of calculation. The error is, within a very small fraction, 1%.

If therefore you want an easy rule for turning your rough calculation into an accurate one, you cut off the last two figures and subtract from your total the figures thus left. For instance, 244 kilometres divided by 8 gives 30½, and that multiplied by 5 is 152.5; cut off the 52, leaving “1” on the left, subtract that 1 (making 151.5), and you are within a few yards of accuracy. As questions of distance count nothing in mountains compared with questions of height, I will make no mention of decimals, but proceed to a very different matter, which is the way of counting that the _mountaineers_ have, and this you will do well to heed blindly.

When you are tired and distracted and wondering perhaps whether you can push on, if you have the good luck to find a shepherd, he will tell you your distance to such and such a place in _hours_. The Spanish, the Gascon, the Béarnais, and the Catalan dialects all use the same words, so far as sound goes, for this kind of measure, and the Basque will never speak to you in Basque: it is part of the Basque tenacity never to do this. So if you find yourself in any part of the high hills where a man can talk to you of distances, you always hear the same sounds “Dos Oras,” “Quart’ Ora,” “Mi’ Ora,” and the rest. This habit, as every reader knows, is universal throughout the world wherever true peasants exist; but in mountains, whether they be Welsh or African, it is not only universal, but it withstands all the invasion of the modern world.

What I would particularly impress upon anyone going into the Pyrenees is this, that such a method of counting is exceedingly accurate, and is moreover the only accurate method. Nothing is more fatal to a civilized man of the plains than to take his little measuring stick and measure upon the map by the scale the distance between two points, saying, “It will take me so many hours.” There was a Basque at Ste. Engrace who very well expressed to me the contempt which mountaineers have for that method of the plains. A deputy of the French Parliament had stopped in his inn, had thus measured the distance from the village to the pass, and would not believe that it could take three hours. It always takes exactly three hours. I have done it in four by careful dawdling, and the dawdling, when I came to reckon it up, had taken exactly one hour out of the four. Now, measured upon the map, that distance, as the crow flies, is precisely three miles, but it takes three hours none the less. You will not do it in less, and what is odder, you can hardly do it in more, for if you deliberately go too slowly, you are done for in no time, and if you halt, you will find that your halt fits in exactly to make the walking time three hours. Similarly, over the Pourtalet, from the last Spanish hamlet to the first French one, is six hours; part of the way you may choose between a good road and a mule track, but whichever you choose it is six hours; and there is nothing more astonishing in Pyrenean travel than the accuracy of this rough method. As I said just now, you must heed it blindly; it is by far your best guide.

The use of maps has one last thing to be said about it, which applies particularly to the Schrader map and to the 1/100,000, and this is that where you think you see a short cut, and the map gives you no track, there the short cut is to be avoided. I say it applies particularly to the Schrader and 1/100,000, because these two maps are so particular in detail that you think their information must be enough without the further aid of a path. Moreover, the path sometimes takes such apparently needless turns that you are for escaping it by an easier cut.

You will never succeed. You may indeed succeed in a bit of exceptionally hard climbing, you may not lose your life, but you will most certainly wish that you had never attempted the unmarked crossing of the ridge you have attacked. It is obvious that the exception to this doctrine would be found in a piece of genuine experiment. If you say to yourself for instance, “I can get over the shoulder of the Pic d’Anie into the valley of the rivulet beyond, which has no name, but which runs into the Tarn of Uterdineta,” you will probably do it, but it will not be a short cut from the Val d’Aspe into the valley of Isaba, though it is the shortest way. These temptations for cutting across the hills come very often in one’s first experiments in the Pyrenees: they get less frequent as one knows more of them. These mountains are full of vengeance, and hate to be disturbed.

NOTE.—A convenient map for viewing the whole range is the 1/400,000
which is sold by Messrs. Sifton & Praed, mounted in two sheets,
and in a case. It is especially of use in showing a large belt
of the Spanish side. Motorists in particular should see it.

IV

THE ROAD SYSTEM OF THE PYRENEES

There are two kinds of platforms for travel in the Pyrenees—mule tracks and great, highly engineered, modern roads. No others exist. When, therefore, one is describing travel in the Pyrenees, one must separately describe the opportunities of wheeled travel open to all vehicles, however elaborate, and of travel on foot or with a mule. As the last will take up the greater part of my space, I will speak of wheeled travel first.

To understand what are the opportunities of this, one may take as one’s standard the roads which can be traversed by a motor car. Those passages which a motor car cannot use cannot be used by a bicycle or a carriage, for the roads of the Pyrenees are, as I have said, either very good broad roads, well graded, and with a hard surface, or they do not exist; the change is always abrupt throughout the chain from an excellent highway, carefully engineered, to a mule track.

The scheme of Pyrenean roads, as it exists now, is, briefly, _first_: a couple of great lateral roads on the French side, which may be called the upper and the lower road; _next_, four roads traversing the chain (six if you count the roads along the sea-coast at either end, which I omit—the one goes by St. Jean de Luz, the other by the Pass of Lacleuse or La Perthuis); _thirdly_, a series of roads, numerous on the French side, rare on the Spanish, which penetrate the valleys but do not cross the chain, and end at a greater or less distance from the watershed.

The main lateral road from the Atlantic to the Mediterranean, along the base of the Pyrenees, links up all the towns upon the plains; it joins Bayonne to Pau, Pau to Tarbes, Tarbes to St. Gaudens, and so on through St. Girons, Foix, and Quillan to Perpignan: this may be called _the Lower Road_. The upper road has been but recently completed. It is made up of sections, some of which are old highways, some links quite newly built, and the characteristic of the whole is that it skirts as nearly as possible the crest of the main chain, crossing at some places very high passes over the lateral ridges, and everywhere keeping right up against the high summits of the range. The whole line runs from Perpignan over the Col de la Perche up the Val Carol and over the Puymorens to Ax, Tarascon, and St. Girons. At St. Girons, it is compelled by the conformation of the country to touch the lower road, but it leaves it at once to pass from Fronsac to Luchon; thence through Arreau, Luz, Argelès, Laruns, Oloron, and Mauléon—all the high mountain towns—to St. Jean Pied-de-Port, and thence back again to Bayonne.

The four roads over the ridge into Spain lie all of them on the western side of the hills. They are, first, the road through the Baztan valley, which connects Bayonne with Pamplona; secondly, the Roman road over Roncesvalles, 12 or 15 miles to the east of this, which used to be the high road between Bayonne and Pamplona before the Baztan road was built, and which was during all history the westernmost road of invasion and communication between Gaul and Spain; thirdly, the road which goes over the Somport, which was also a Roman road and the chief one, uniting Saragossa with the French plains; fourthly, a road parallel to this and not 10 miles east of it, running over the Pourtalet Pass and joining the Saragossa road lower down. No other roads cross the range from France into Spain until one reaches the Mediterranean, and all these four lie within the first westernmost third of the Pyrenees.

It would be quite easy to open other roads which should unite the last of the Spanish highways with the first of the French, notably over the easy pass of Bonaigo, where 20 miles of work would be enough, and through the Cerdagne, where there are no engineering difficulties. One such road is now in process of completion between Esterri and St. Girons over the pass of Salau. Another, which was begun from the valley of the Ariège into Andorra, was abruptly stopped, and it will probably never be completed. There are some half-dozen other places, where a road could cross and where the French are building their side of it: but the Spaniards are reluctant to meet them.

Of the roads of the third kind, roads running up the valleys but not attempting to cross the mountains, one may say that on the French side every valley has one or more good roads, the one drawback to the use of which in a motor is that you are compelled, unless you can take a cross road from one high valley to another high valley, to go back by the way you came into the plain.

Not only has every valley its highway leading to the very foot of the main range, but often the bifurcations of the valley will have roads as well. Thus along the valley of the Nive you can go in a motor not only to St. Jean Pied-de-Port, but also right up the eastern valley to a country-side called the “Baigorry” as far as Urepel; along the next Basque valley to the east, you can go from Mauléon in the plains right up into the hills as far as Larrau, but you cannot go to Ste. Engrace, where the valley splits, because the track thither, though a good one, will not take wheels. You can go up the branch valley from Oloron as high as Aritte, and the main road up the Val d’Aspe (which is that leading to Jaca by the old Roman way), has lateral branches, one taking you to Lourdios, the other across the foot hills to Arudy and the Val d’Ossau. The valley of Lourdes has a road which, with the exception of the roads over the passes, goes nearest to the main watershed. I mean the road to Gavarnie; and the Val d’Aure, which comes next to the westward, has a road going as far as Aragnonette, almost as close to the last cliffs as Gavarnie is; and there is an embranchment to the east which takes one to the very foot at the Hôpital of Rivanagon in one of the loneliest parts of the hills. The road to Bagnères de Luchon is carried some miles beyond that town, as far as the Hospitalet, which stands at the foot of the pass into Spain. The road to Viella in the Val d’Oran goes on up to within a mile or two of the pass of Bonaigo. A road from St. Girons takes one up the valley of the Lez as far as Sentein, which, like Gavarnie, lies right under the main chain, while the road from the same town up the main valley of the Sallent goes up to the watershed itself, and is being constructed to cross it, and to afford (over the pass of Salau) one more badly needed passage into Spain. The valley of the Ariège has a road all along it, almost to the sources of that river. It is continued through the Cerdagne and down the valley of the Tet into the Roussillon.

There is not a main valley on the French side of the Pyrenees which has not its great carriage road, and most of the lateral valleys have now the same kind of communications. The journey up them is nearly always of the same kind, save the few which are prolonged to carry over the watershed into Spain. There is the succession of two or three enclosed plains or jasses after one has left the plains, the sharp pitch up to one flat, and then another, through short but steep rocky gorges, till we reach the little terminal mountain village, sometimes not more than a group of three or four buildings, lying under the last escarpment, and in sight of the frontier ridge above it. Of this terminal sort was Urdos until Napoleon III pushed the road out beyond it into Spain; Gabas, until the Republic did the same with the road there; and of this sort still an old Hospitalet, Sentein in the Val d’Aure, and though it is in a state of transition, for the road is now being pushed beyond it, of this sort is Gavarnie. Little places almost as old as our race, with no history and no national memories, but with immemorial traditions, rooted as deep as the mountains, were brought into the life of our time by that new activity of the French, which is to many foreigners so hateful, to many others so marvellous.

On the Spanish side there are no roads of this kind penetrating the valleys except the incomplete road to Isaba from Pamplona by way of the Val d’Anso, and the short stretch from Sandinies to Panticosa.

A road is being made up the Val d’Anéu, but it is not yet finished, and a road goes just so far up the broad Segre valley as Seu d’Urgel.

All the other valleys have mule tracks alone.

The general scheme of existing roads in the Pyrenees is roughly as upon the map on previous page, where it will be seen that much the greater length of the chain is impassable to a wheeled vehicle.

Motoring sets a standard for every other form of wheeled traffic, I will therefore first speak of this kind of travel. The best road to take with a motor, if one wishes to obtain a general idea of the Pyrenees, is the Lower Road (by Tarbes and Foix) from Bayonne to Perpignan; one may then come back again from Perpignan to Bayonne by the upper road, many parts of which are of very recent construction and which goes right through the highest part of the chain across the main lateral valleys of the Pyrenees. Such a round—about 500 miles altogether—gives one from far and from near the whole of the French Pyrenees: from the first one sees the chain as a whole before one: by the second one mixes with its deepest valleys.

The first day’s run from Bayonne had best end at Tarbes; it is a town central with regard to the chain, and it is also a very pleasant place to stop at under any conditions; not cosmopolitan like Pau, and not in a hole and corner like Foix.

The lower road from Bayonne to Tarbes runs through Orthez, Puyoo, and Pau, and if one starts early, Pau is a good halting-place for the middle of the day. This part of the road is, during the whole of its length or nearly the whole of it, a rolling road of the plains with no striking points of view save in where it tops a slight rise. It first follows but runs above and north of the valley of the Adour below it, next descends after the first 20 miles or so to cross the Adour, and so comes to Peyrehorade, the first town (and railway station) upon its course. During all this first part of the run one has sight after sight of the range which stretches out eastward before one to the south rising higher as it goes; and one sees at first before one upon the horizon, later abreast of one and due south, the pyramid of the Pic d’Anie, which is the first of the high peaks.

From Peyrehorade to Pau, between 40 and 50 miles, the road goes through Orthez along the valley of the Gave de Pau, for the most part following the river bank and allowing but few sights of the range; but at Pau itself it rises on to the high plateau of the town whence the most famous general view of the Pyrenees is spread before one.

From Pau there are two roads to Tarbes; for curiosity and for general travel it is the road round by Lourdes which is generally taken, and that is during the whole of its length a lowland road though it runs among the foot hills; but the better road on such a drive as I am describing is the direct northern road, which, after it has climbed on to the plateau of Vignan, goes up and down steep small ravines until it comes down again upon the main valley of the Adour and the plain of Tarbes.

There are on this road two points, one just after one leaves the railway line, not quite half-way to Tarbes on the climb up to Vignan, the other just before the loop and descent above Ibos, which afford fine views of the range to the south, and one begins to gather one’s general impression of these mountains, which, more than any other range, present an appearance of simplicity and the united effect of a barrier. Tarbes, less than 30 miles from Pau, may seem a short run for one day from Bayonne, but it breaks the journey exactly and conveniently.

After Tarbes (where the hotel for you is the Hotel Des Ambassadeurs) the road goes through much broken country, passing by Tournay up on the high plateau of Lannemezan to Montréjeau. It is a road full of short hills, but it is necessary to take this section in order to go eastward from Montréjeau and to proceed through St. Gaudens, taking an elbow by St. Martary and so down to St. Girons.

After St. Girons one follows the new and excellent road which runs along the valley side by side with the new railway to Foix. From Foix to Nalzen your way is to go along the main road from Foix up the Ariège Valley for some 4 miles and then turn to the left, leaving the railway and making due east. From Nalzen continue to Lavenalet; there take the right-hand road to Belesta and Belcaire; thence, when you have crossed the plateau, a very winding road takes you down hundreds of feet, on to Quillan. After Quillan you have a few miles through the very little known and wonderful gorges of Pierre Lys to St. Martin, through which gorges the railway accompanies you. Do not follow it round by Axat, but cut across by the road which goes eastward to La Pradelle. This road takes you across a low pass to the watershed of the Mediterranean. From La Pradelle to Perpignan the road is a perfectly clear one through St. Paul and Estagel. It is a straight, good road, following the valley all the way, save the last stretch, which runs across the plains between the river Agly and the Tet.

This second day will of course be far longer than the first; it is nearer 200 miles than 120. If you would break it, however, break it rather after the short run to St. Girons, than at Foix, for though Foix be nearly the half-way house, yet the accommodation is better at St. Girons, and so is the cooking.

A two days’ run of this kind from the Atlantic to the Mediterranean, following such a route, gives you the whole distant range in one general appearance, and gives it you better than you will have along any other line with which I am acquainted.

The way back by the upper road from east to west through the Pyrenees is a piece of travel quite peculiar to these mountains; nowhere else in Europe is there a lateral road driven right across the buttresses or supports of a main range. The Pyrenees possess such a road in their highest part. What the French have done here is as though the Italians had driven a road from the sources of the Dora Baltea right under Mont Rosa, and the Matterhorn to Lake Maggiore, or as though the Swiss had driven one from Faido and Fusia right over into the valley of Domo d’Ossola. From Tarascon in the valley of the Ariège to Laruns in the Val d’Ossau—that is, over all the central part of the chain and for just over half its length—a mountain road goes right up against the main heights (only once coming near the lowlands at St. Girons), crossing the high, perilous passes which lie between the upper valleys. By taking advantage of this new piece of engineering you can return from Perpignan to Bayonne through the midst of those hills which the road just described from Bayonne to Perpignan showed you in a distant general view: when you have so returned you will have seen the heart, the French Pyrenees.

I will now describe such a return journey by the upper road. From Perpignan you will do well to run the first day to Ax. The road is the great road from the Roussillon into France. You go up the valley of the Tet (which is the main river of the Roussillon) through Prades with the Canigou first right in front of you, and at last rising steeply to your left. You continue through Prades up the gorges and tortuous zigzags of the Upper River until you come to the head of the pass at Mont Louis: there the broad and easy valley of the Cerdagne opens to the south, sloping gently before you. The road runs down, almost as in a plain, to Bourg Madame, where you must turn to the right up the Val Carol to Porté. The pass above Porté (called the “Puymorens”) though long, is of an easy gradient, and once over it you run down all the 18 miles to Ax, following the valley of the Ariège.

Ax is, of course, an early stopping-place. The whole distance from Perpignan is under 140 miles, but Ax is so much more comfortable than Tarascon that it is better to make one’s halt there.

Next day go down the valley as far as Tarascon and there take the mountain road off to the left; it is not a national[1] road but it has a perfectly good surface in spite of a considerable climb. One little col comes almost immediately at Bedeillac, after that you climb steadily up the valley to the Col-du-Port (which is about 4000 feet high) then down the mountain side to Massat, which lies on the western side of the pass and about 2000 feet below it. Thence it is an ordinary valley road until you come to St. Girons again.

[1] The French metalled roads are of three main kinds, supported
by the State, the County and the Parish respectively. Of these
the first and most important are called “National Road.”

From St. Girons you continue this progress parallel to the watershed and right among the high peaks, by taking the cross road from St. Girons to the valley of the Garonne. Just before the railway station at St. Girons turn sharp to your left, taking the road which goes up the left bank of the Lez. At this starting point you are not more than 1300 or 1400 feet above the sea; at Audressein (300 feet up) turn to the right, cross the river, and begin to climb the upper valley until you reach the col of Portet-d’Aspet at about 3400 feet, that is, some 2000 feet above St. Girons, and between 15 and 20 miles from that town. From this col the road descends rapidly down the valley of the river Ger, falling in 5 miles 1500 or 1600 feet. At the end of the 5 miles you take a road that goes sharp off to the left before reaching the village of Sengouagnet, this road going off to the left crosses a low watershed, makes, at the end of another 5 miles, a great loop round the forest of Moncaup (the church of which village you leave to the left just before making the turn), and comes down into the great open plain into which the valley of the Garonne here enlarges. It is one of the finest enclosed plains in the Pyrenees, and to come down upon it by this road is perhaps the best way to approach it.

The first village in this plain is Antichan, thence several long windings take one down to Frontignan below, and thence it is a straight road through Fronsac to Chaum where there is a bridge over the river, and where the plain of which I have spoken terminates in a narrow gateway through the hills. You cross the river by this bridge, fall at once into the great national road upon the further or left bank, and a straight run of not more than 12 miles in which one only rises 300 or 400 feet up the tributary valley brings one to Bagnères-de-Luchon. Though at the end of an even shorter day than was Ax from Perpignan, Bagnères will make a convenient stopping-place after a good deal of hill climbing and roads the surface of which, especially in the early summer, is occasionally doubtful. Bagnères has, of course, everything that people motoring can want, it is the capital of the touring Pyrenees, and even if this cross journey has not proved enough for one day, the character of Bagnères makes it the right place to stop at on the second day.

Though Bagnères is right in the middle of the mountains, but a mile or two from the frontier of Spain, not 6 miles, as the crow flies, from the watershed and within ten of the highest peaks of the Pyrenees, yet the importance of the town has caused good communications to spring up around it, and there is an excellent road crossing straight over from the high valley of Bagnères into the next high valley, the Val d’Aure. It starts at the market-place just opposite the new church, crosses the col called “Port-de-Peyredsourde,” and comes down into the main road of the Val d’Aure at Avajan, which follows down the stream at an even gradient to Arreau, 7 miles further on.

Arreau is the capital of the Val d’Aure, and when you have reached it you will have come about 20 miles from Bagnères.

The next parallel valley to the Val d’Aure is that of the Gave-de-Pau: the valley which has at its mouth the town of Lourdes, and at its head, right under the Spanish frontier, the famous village and cliff of Gavarnie. There is, indeed, a small subsidiary valley in between where the Adour takes its rise, and of which Bagnères-de-Bigorre is the capital, but it is shorter and stands lower than the two main valleys upon either side. The section I am about to describe, the great new road from the Val d’Aure to the Valley of Lourdes, just touches this upper valley of the Adour but does not pursue it.

The cross road from Arreau in the Val d’Aure to Luz in the valley of Lourdes is the steepest and the most diverse in gradient, as it is also by far the finest in scenery, of all the new sections which have recently been pierced through the highest parts of the range and between them build up what I have called “The Upper Road.” The distance as the crow flies from Arreau to Luz is not 20 miles, but the long windings of the road which take it over two passes, and the northern diversion necessary to turn the great mountain mass of the Port Bieil, lengthen it to nearly double that distance.

There is no mistaking this road. It branches off at Arreau, leaving the valley road not half a mile beyond the bridge and going to the left up a little side stream, the name of which I do not know. Within 2 miles it crosses this stream and begins to take the long complicated and graded turns up the mountain. One must be careful, by the way, at the point where the road crosses the stream to turn sharp to the right and not go straight on towards Aspin, for though one can get to the main road again from Aspin, it is by roads too steep for a motor. If one so turns to the right, the road goes up to the col in great zigzags and climbs in some 6 or 7 miles the 2000 feet between Arreau and the summit, thence it falls rapidly for 3 or 4 miles to a point where the new road cuts off the corner the old road used to make. It is important to recognize this point, not only because it saves one at least 6 or 7 miles of travelling, but also because it saves one going right down into the valley of the Adour and climbing up again. I will therefore attempt to fix for the traveller the exact place where he must turn off to the left, though the description is difficult on account of the absence of any landmark.

As you come down from the Col d’Aspin, you run through a wood along the mountain side for perhaps 2 miles. The road sweeps round the curve of a gulley on emerging from this wood, crosses the rivulet of that gulley, and comes down close to the stream at the foot of the valley which is the source of the Adour. Just at this point a road will be seen coming in from the left, descending the slope of the valley beyond the stream and crossing it by a bridge. This is _not the road_ you are to take. You must continue on the same road you have been following down from the pass, until, in about half a mile, it crosses the stream to the left bank, and approaches on that bank a wood that lies above one on the hill. Immediately after this bridge there is a bifurcation; one branch goes straight on, the other goes off to the left; this last is the one you must follow. The branch going straight on is the old road which leads down the valley of the Adour, and from which one used to have to double back some miles on at an acute angle to reach Luz. The new road, which you must thus take to the left, cuts off that angle.

There are no difficulties from this point onward. The road winds a good deal round the hill-side, and almost exactly 5 miles from the point where you turned into it you come again upon the main road to Luz over a bridge that crosses a stream. Just where you join that main road it begins its long climb up to the pass called Col du Tourmalet.

This pass is the highest and steepest on the secondary or lateral passes, over which the new roads have recently been driven. It is just under 7000 feet in height, is everywhere practicable, and once it is surmounted there is a clear run down of some 10 miles and more (following the valley called locally that of the Bastan) to Vielle and to Luz in the main valley.

Of all the crossings between the high valleys of the Pyrenees this is the one best worth taking. The height of the pass, the great mass of the Port Bieil dominating one side of the road, and of the Pic-du-Midi dominating the other, give it an aspect different from any other of the secondary roads, and comparable only to the two main passes of the Somport and the Val d’Ossau.

From Luz a great national road takes one down the valley to Argelès and the railway, a distance of about 18 miles, and the end of about as fine a piece of engineering as there is in Europe. From Argelès, which is just above Lourdes and whence Lourdes can be reached at once by road or by rail, the cross road which I am describing goes on over another high pass into the Val d’Ossau.