base by drift and lake water-a true measure of the original displacement on its fissure plane, thus exemplifying a general principle sometimes applicable in cases of the kind here considered; namely, when two adjacent, low-standing prisms of slightly different displacement have both lost their unconformable cover of stratified rocks, they will be separated by a fault-line scarp equal in height to the original fault scarp. A second sequent of Motala River flows eastward along the scarp base from the narrowed end of Lake Roxen a short distance, and then cuts a northward course across the Örn lowland.

Like the Roxen prism, the Örn prism shows the ancient peneplain as a lowland gently inclined to the north-west, and somewhat modified by erosion and deposition since its cover was removed; this lowland is peculiar, however, in being limited on two sides, west and north, by resequent fault-line scarps, between which it enters in an acute angle; and it is in that angle that little Lake Örn, only three kilometres across, is ensconced. Next north-east is the Glan prism, the lowland surface of which, partly submerged in Lake Glan, 71 metres altitude, is even more peculiar in outline than the Örn. Its north-western side is bounded by the fault-line scarps of three high-standing prisms, all of which advance in salients to the south-east. The first of these prisms rises north of the Örn prism and west of the Glan; the second advances upon the Glan lowland and lake in a sharp-angled promontory; and the third, narrower than the other two, forms the northern limit of the Glan prism. It is probable that the Glan prism is really composite, being subdivided by the continuation of the fissures which separate its higher north-western neighbours; but the displacement of the fissures must here be small any scarps that they determine near the northern side of the composite prism are hidden by the lake waters, and none is recognisable in the stripped area of the ancient peneplain south of the lake. The last-named high-standing prism, which bounds Lake Glan on the north-east, soon terminates its narrow upland eastward in a rounded salient, beyond which a long and well-defined south-facing fault-line scarp, offset to the north of the northernmost scarp of Glan, is continued much farther eastward the stripped lowland which slants gently to this scarp from the south is submerged as it approaches the scarp base in an arm of the Baltic Sea, known as Braviken; its western end is entered by the final segment of Motala River, coming from Lake Glan.

About 50 kilometres north-west of the nearly dry lowland of the Örn prism is the wholly dry lowland of the Tjellmo prism; it is therefore not essential that, when the ancient peneplain of a low-standing prism is largely or entirely stripped of its limestones, it should be occupied by a lake in its lowest part. Some distance farther north, several lowlands and lakes of good size are found, next to which the limiting scarp rises on the south and faces north, thus indicating that the surfaces of the lake-bearing prisms were there tilted southward, and not northward. like the prisms of the Roxen district.

As far as present information goes, the Timiskaming basin does not

appear to be complicated by criss-cross fissuring; but a detail not previously noted must now be introduced: a certain part of the uplands west of the lake-limiting scarp is occupied by limestone patches near the north end of the lake. I have not been able to secure information about the altitude of these patches, but suspect that they represent the residual cover of a prism of less uplift than the rest of the western area, so that the surviving limestones there escaped complete removal during the production of the second peneplain by reason of their lying below its level. It may be also noted that the outline of Lake Kippawa, as shown farther south-east on the Standard Topographic Map, seems to be guided by several intersecting structural trends; and the closer study of that district may yet discover detailed features, such as inter-prism trenches and sharp-angled promontories, like those of the Swedish lake district here considered.

Two general questions remain to be considered. First, is it worth while for geographers, whose studies are primarily concerned with the existing features of the earth's surface, to enter so far into the past geological history of lake basins or of any other land forms as has been done above? My own feeling is that it is well worth while to do so, because of the better understanding of many existing features that is thereby gained. There is, indeed, a certain mental satisfaction that is felt when the reasons for the occurrence of any land form, and for the peculiar association of its various elements, come to be understood; no one who has experienced that sort of satisfaction is likely to discard the method of study by which it has been gained. It is, of course, somewhat helpful to describe a lake and its surroundings simply in terms of directly observable facts; but in these modern days, three-quarters of a century after Darwin began to teach the grand philosophy of evolution to a then unwilling but now grateful scientific world, directly observable facts are not enough. The meaning, the origin, the evolution of the facts, and the intimate correlations into which they all enter, must be sought for. In the case of lesser Roxen and larger Timiskaming, as well as in many other cases, these correlations are immediately helpful, first in aiding an initiated observer in his search for the facts, and later in aiding him to prepare a description that will lead an initiated reader better to visualise the landscape which the facts make up. It seems probable that, whether the explanatory scheme here set forth to account for the Timiskaming basin is correct or not, fuller descriptions of the basin would have been prepared by the observers who have studied it, if the scheme had been familiar to them, and its correctness or its error would then have been determined on the ground.

Some geographers are, however, still rather timid about this "trespassing," as it is sometimes called, upon the domain of the neighbouring science of geology in the search for a better understanding of their own problems; but it may be very confidently urged that so long as the object of the trespass is not merely to learn the history of the past, but better to appreciate the present, the trespass is well justified. It is,

indeed, quite as justifiable for a geographer to utilise geological methods of investigation in the study of a lake basin or of any other physiographic feature, as for a mineralogist to utilise chemical methods of investigation in determining the composition of a mineral, or for an astronomer to utilise mathematical methods of investigation in determining the elements of a comet's orbit. Here, truly, the end justifies the means. Of course, geology may be useful to geography; all sciences are more or less useful to one another. And when geology is thus found to be useful to geography, some return is made for the enormous service that geography has rendered to geology; for let us not forget that all that large part of the superb structure of geological science which deals with the unobservable physical conditions and processes of the past, and can consequently be learned only by a process of mental speculation, is based, and very solidly based, on existing geographical facts which, being visible, can be reached by direct observation.

A second question arises with regard to terminology. If it be agreed that an explanatory description of a certain well-defined kind of lake basin is geographically helpful, it is yet to be determined whether that kind of basin should be given a special name, at least for use by proficients if not by beginners also: and if this question be answered in the affirmative, a name will then have to be chosen. A name is, in my opinion, needed because it is practically helpful to have a definite name for a definite kind of thing; that is, because after the explanation for a special kind of lake basin has once been learned, it is so much easier to refer to it by a brief one-word name than in terms of the more or less complicated succession of conditions and processes that the explanation involves. To be sure, it has been said that the introduction

of a new term adds to the burden of science; but that is also true with regard to the introduction of a new fact or new explanation of a fact, yet we still go on seeking new facts and new explanations. The burden of science is, evidently enough, a little increased by the explanatory scheme set forth in the present article; but that burden will be lightened if the kind of lake basin to which the explanatory scheme is addressed be given a name by which it can be easily brought to mind.

Timiskaming is too long for a handy name. Roxen is more satisfactory, both because it is shorter and because the explanation above presented was applied to it a quarter-century before it was to Timiskaming. Hence, if it be understood that the type example, Roxen, occupies a basin that has been excavated in a second cycle of erosion after post-faulting peneplanation had been reached in a first cycle, and that the basin is limited along one side by a fault-line scarp, then in order to indicate the general nature of Timiskaming, it suffices to say that that water body appears to be a lake of the Roxen type; or, in short, that it is a Roxen lake in Canada. If its orientation and its dimensions are then added, a good acquaintance with it may be acquired.

THE HUMAN GEOGRAPHY OF LEWIS.1

By ALEX. STEVENS, M.A., B.Sc., Lecturer in Geography at the
University of Glasgow.

(With Sketch-Map and Illustrations.)

THERE are certain features of the geography of Lewis which give it a special interest, and even a considerable value as a type. Since it lies to a very great extent below the level of 500 feet, and culminates below 1000 feet, its meteorological and climatic conditions are to all intents and purposes sea-level conditions. Composed as it is almost exclusively of archæan rocks, modified by the processes of the same history as the mass of the Highlands, it presents a simplification of Highland geography: the conditions of soil and climate without the complications of altitude and position which, from the human point of view, divide up the desert of the Highlands into scattered oases of population, the conditions and characters of which vary with their circumstances of level and aspect.

In Lewis we find also that inertia due to the economic isolation has prevented its human geography being merged in that of the mainland; while the most Highland of mainland communities are mere outliers of mainland conditions. Isolation has made Lewis unique, especially if we except the town of Stornoway, which, in a sense, is more foreign to the landward islanders than Glasgow. The density of population in Lewis is 45 persons per square mile. Taking only the rural population, that is, excluding the population of the town of Stornoway, the density according to the latest census is still over 38. For comparison we may note that the adjacent counties of Sutherland and Inverness (the latter being also largely insular) have densities of 9 and 20 respectively, while Ross, to which Lewis belongs, has a mean density as a whole of 23, but excluding Lewis of 17; in other words, Lewis contains 40 per cent. of the population of Ross on 20 per cent. of the area. In density of population Lewis compares with such counties as Shetland (46), Peebles (44), and Caithness (41), all of which are more favourably circumstanced climatically or geologically or both. The percentage of burghal population is 14, while for the whole county it is 16 1 (compare Shetland 19, Sutherland 43, the lowest in Scotland). Stornoway (pop. 4079), moreover, is larger than any town on the mainland north of the Great Glen except Wick (4296), whereas the burghal population of Caithness is 39 per cent. of the whole. There is no doubt that these facts are largely due to inertia, which in its turn is due to isolation. This order is changing, for the rural population of Lewis showed a small decrease in the last intercensal period (38 per cent. for the whole population), but the population of Stornoway increased in the same period by 273, or 7.2 per cent.

A lecture delivered before the Society in Edinburgh on February 5, 1925.

Note. The figures after the symbol for shielings indicate the relative importance of each group, but not the actual number of huts.