not find inflammable material elsewhere, and once more spread across Europe and the world.

Of the three other disturbed areas already named, the conditions in two cannot, at the moment, be profitably discussed. At first sight it may seem that the third problem, that of the political destiny of Alsace-Lorraine, is so simple as hardly to merit detailed consideration. It is, however, in point of fact one of considerable complexity, with bearings upon many others which press for solution. On this account, therefore, it appears justifiable to set forth the elements of the situation in, so far as may be, a compact and convenient form, from the geographical side. More than this will not be attempted here, for, though the object is to attempt to lay the basis for a political settlement, that settlement is in itself outside the sphere marked out.

THE FRONTIERS OF FRANCE.

It is not possible to appreciate fully the significance of the AlsaceLorraine problem without a brief reference to the frontiers of France generally. The essential points are clearly brought out on the accompanying sketch-map (Fig. 1), which shows that, with one notable exception, the boundaries of the country are for the most part natural: natural because, save in the north-east, they are made by steep-sided mountains of recent origin, or by the sea. In the tract which extends from the vicinity of the town of Bâle to the North Sea, however, high land is either absent, or, where present, is not composed of young folded mountains.

The contrast between the two types of high land is important, and to realise its significance it is necessary to glance for a moment at the general character of the Pyrenees, the Alps proper, and the Jura, which all play so important a part in demarcating France.

To such bounding chains we may justly apply the term "natural," and that for two reasons whose relative value has changed in historical time. The first depends upon a fact of human geography, the second upon strategical needs. Pyrenees, Alps, Jura, alike, whatever their relative height, resemble one another in the steepness of their slopes, and therefore of their stream valleys; in the rapidity with which erosion takes place, and therefore in the tendency to display bare rock surfaces and masses of coarse débris; in the rapid change of slope in the valleys, and in the frequency of gorges. Generally, they show all the characters of areas which have undergone marked and geologically recent elevation. In consequence they tend, on the whole, to repel settlement, to constitute areas with a low density of population. Further, where settlement does take place, the peopled areas are isolated, and thus combination is difficult, patriotism is local rather than national, and complex communities are not likely to arise. Such areas, therefore, are naturally marginal, not central, and seem to be set apart by nature as boundary zones.

At a later stage in human history, when complex organised States have arisen in the plains which such mountains separate, these acquire a new importance, depending upon military needs. Lines of communication across them are necessarily few, and they can be artificially blocked with greater or less ease, for they are commanded by neighbouring heights. It is to the interest of the States on the near and hither side to strive to keep an invading army away from the accumulated wealth of the plains, and the physical features of the mountains facilitate this.

FIG. 1. Structural Map of France showing boundaries in 1870 and in 1914. The young folded mountains (Pyrenees, Alps) are indicated by one type of shading, the plateau regions, built up of old, hard rocks, by another. The unshaded areas are regions of unfolded, younger rocks, whose dip is for the most part slight. The area annexed by Germany from France in 1871 is dotted lightly. (In part after Barrès.)

Both of these are, of course, elementary points, but it is well to emphasise the close connection between the geological structure and history of such mountain belts, and their significance as natural frontier zones. The fact that, as the map suggests, they do not, as is sometimes loosely supposed, consist of a single chain of crests, but of a series of folds (indicated by black lines) which may exercise marked influence on the course of the rivers, is of much importance in connection with their frontier function. Their great elevation has also, of course, a marked effect on climate, and, in temperate latitudes, the mountain climate is unfavourable to settlement.

We have next to consider the character of the uplands which occur, in two separate masses, on the north-eastern frontier of France. But, before doing this, it is important to note that, as the map shows, these share certain characters with other highlands, found well within the boundaries of France. The upland of Brittany, the irregularly shaped mass known to geographers as the Central Plateau, alike resemble the marginal highlands of the north-east in their geological structure and in their history. All are composed of old, hard rocks, sometimes mingled with volcanic beds belonging to a much later period. In all the surface has been worn down by long-continued erosion, so that, except where recent faulting has taken place at their margins, their slopes are gentle, their river valleys far less anomalous in character than those of the folded mountains. There are no true mountain peaks, and exposed rock surfaces do not occur over extensive areas. On the other hand, the soil is often infertile or shallow, sometimes impermeable, and the areas tend to be covered with dense forest, or with upland bogs, or with upland pastures, or with a combination of all three, long after the lower grounds have been cleared for tillage.

It is obvious then that, though to a less extent, they share the character of the true mountain belts in tending to repel settlement, and to be, whatever their position as regards the country as a whole, marginal in relation to its densely peopled lands. In all, therefore, we find human groups which have stood more or less aloof from the historical development of the lower grounds. The Breton language in Brittany, the persistence of German on the eastern slopes of the Vosges and in part of Lorraine, the non-Catholic communities of parts of the Central Plateau, are all indications of this fact.

Thus, where such uplands occur within the country proper, their absorption in the general community is a slow and partial process; where they are placed at its edges they may be expected to show, at least for a long period, divided allegiance. At the same time, however, the facts that communication between them and the surrounding plains is much easier than in the mountains proper, and that settlement can, with advancing civilisation, be pushed steadily upwards from the plains towards their summit levels, means that their aloofness is of a far less permanent character than in the case of the mountains. Again, for reasons which also depend upon their topographical features, they are of less value as strategic boundaries, even when continuous, than the Alps or Pyrenees.

Their absorption in the general community has, however, been, within the last hundred years or so, facilitated by a geological fact which, though not shown on the map, can be deduced from it. It has been prolonged to include the southern part of England and Wales, in order to suggest the general similarity of structure between Northern

France and the former. Thus we see on it the greater part of the English plain, with the marginal uplands of Wales and of Devon and Cornwall.

Now we all know that Cornwall and Wales, no less than Brittany or the Central Plateau of France, were for a long period regions which led a life more or less apart from that of the plain. But, in the case of Wales, once an independent political entity, amalgamation in the general community has been within recent times greatly accelerated by the fact that on the margin of the old worn-down upland coal-beds crop out, just as they also crop out in the midst of the English plain in isolated masses, forming the Midland coalfields. The industrial type of life has grown up, alike at the margin of the upland, and in the heart of the plain, on the basis of the coal-beds.

Without at the moment laying stress on the point, it may be noted that in France and its borderlands also coal-beds sometimes occur at the margin of the uplands, and have brought a new density of population, a new type of life, a new complication also, to the once isolated communities of the uplands. Coal occurs especially at the north-western and south-eastern margins of that upland mass whose western extremity is named Ardennes on the map.

Let us look next in a little detail at the bordering highlands of France. The smaller, roughly triangular in shape, is named Vosges on the map. But it does not correspond to the whole of the Vosges Mountains of the atlas map, for, as we shall see later (Fig. 2), the actual nucleus of the Vosges, as represented here, is prolonged northwards by rocks of a different type, younger in age, different in character, and forming land lying at a lower elevation. The patch shown here, the High Vosges of French geographers, indicates land which over a considerable area rises considerably above 3000 feet. The highest peak, the Ballon de Guebwiller, is comparable in height to Ben Nevis, for it rises to 4670 feet, or some 200 feet higher than the great Ben.

The second patch of highland is the left bank part of the Rhine Highlands, and is itself divided into three parts. To the south-east of the Moselle river is the Hunsrück; then comes an area in which volcanic rocks are intermingled with the old sedimentary beds, forming the desolate Eifel region. Separated from the Eifel by two small river valleys, not shown here, is the region of the Ardennes. The heights here are throughout considerably lower than in the High Vosges. The area which attains 2000 feet is small, and no peak reaches the 3000-foot limit.

A point of very considerable importance, well shown on the map, is that the younger rocks, and therefore the lowlands, interpenetrate the massif to the south-west in a great bay, lying on the left bank of the Moselle. This bay is called the Gulf of Luxemburg by the geologists (Fig. 2); politically it constitutes the nucleus of the Duchy of Luxemburg. From the standpoint of human geography it forms an area in which a fairly dense population is possible in the midst of a forested, infertile upland.

The fact that the Vosges and the Rhine Highlands are discontinuous means that three gaps, of very unequal size and character, mark the north-eastern frontier of France. Of these the southernmost, the Gate of Burgundy, or gap between the Jura and the south end of the Vosges, is the narrowest. It is also commanded by the strong place of Belfort, which resisted gallantly in the crash of 1870, and has not been seriously threatened during the present war. We need hardly, therefore, regard it as a weak spot in the frontier zone. It is worth note in passing, however, that in that it permits of easy intercommunication between the Rhine valley, or rather, the Ill valley, and France proper, it has a considerable bearing upon the problem of Alsace.

The second gap, with the qualification already made, extends from the northern end of the High Vosges to the Hunsrück. Politically it includes a part of Alsace, and the High Plain or Plateau of Lorraine. Here it was that Germany, in 1871, thrust herself across what on any scheme of frontier-drawing must be regarded as the natural boundary of France, and, against the will of the inhabitants, established herself at Metz, and even beyond the Moselle. It is the problem of the political destiny of the whole belt, stretching from the Gate of Burgundy to the point where the Moselle enters the Highlands, which constitutes the Alsace-Lorraine problem proper.

The third gap is that which extends from the western end of the Ardennes to the sea, and here the actual conditions are almost as simple as they appear on this map. In other words, no natural feature, important either from the point of human geography or of strategy, separates the French and Belgian sectors of the coastal plain. Here, therefore, France is open to the foe, or at least was open according to all the older conception of strategists; we do not yet know how far the experiences of the present war may modify these conceptions. The establishment of the neutral state of Belgium was, from one point of view, an attempt to draw a moral barrier where no material one existed. In August 1914 the world saw the regard which the military Super-state of Prussia paid to a moral obligation.

THE RAMPARTS OF PARIS.

To understand the temptation which, given the German code of morality, justified the violation of a solemn pledge, it is necessary to look for a moment at the Basin of Paris.

In Fig. 1 the surface features of that basin have been omitted, though to some extent they can be deduced from the courses of the rivers as shown. Paris is the true heart of France, in a far more precise sense than London is the heart of England. Upon it, as the map shows, converges a whole network of rivers, draining from the uplands and their continuations. Two eastern rivers, however, escape the attractive influence of the basin, and by an apparent anomaly, after flowing northwards for a time, break through the marginal highlands to reach the one the Rhine direct, the other the Rhine delta. These two rivers are of course the Moselle and the Meuse, and the nature of the course of each shows that there must be topographical features in the Paris Basin of which this map tells us nothing. What these features