shire. Through the courtesy of the Messrs. Newall of Dalbeattie, the writer has been enabled to identify some of these with the grey granite which is now being so extensively quarried in the neighbourhood of Criffel for ornamental purposes; and Prof. Geikie of Edinburgh has spared the time to look over and name a box of stones which could not be traced to any rock in the Lake district, thereby enabling the writer to identify many of the far-derived stones in the drift, and affording a tolerably sure clue to the course taken by the ice. The result is that the writer has been enabled to prove the existence of great quantities of the drift from the south side of the Scottish southern uplands in the Eden valley, up to the top of Stainmoor itself. The stones are smaller and are less common as we advance towards the head of the valley; but there they are, glaciated, and associated in such a way with the boulders from sources nearer at hand as to leave no doubt that these Scotch boulders too were carried up to the summit of Stainmoor by the same stream of land-ice that bore the Ennerdale syenite up the Eden valley and over Stainmoor, instead of following the course it would take under ordinary glacial conditions, and moving westwards towards the low ground of the Solway.

Line D, coinciding in great part with the foot of the Cross-Fell escarpment, approximately represents the northern boundary of the Shap-granite drift; and beyond the area in which this rock occurs, the line represents, as nearly as can be made out at present, the north-eastern limit of the Scotch drift in the Eden valley above Brampton*. A glance at the course taken by the stream of Scotch drift ice will show at once that it had precisely the direction that, on its meeting with the stream from the western side of the Dale district and the Howgill Fells, would cause the combined currents to take the direction which the striæ on the slopes of Stainmoor, and the direction of boulder-dispersal, show that the ice had there.

In considering the causes that impelled the Scotch drift up the Eden valley, we have to bear in mind that to the east of Carlisle great quantities of the same drift have gone over the watershed between the Eden and the South Tyne eastward to the North Sea, and that therefore the transporting current seems to have moved in a generally easterly direction over the low ground about Carlisle until it reached the north-west corner of the Cross-Fell escarpment, against which it split, the southern half coasting along the escarpment itself, while the other passed into the valley of the South Tyne. The general direction taken by the ice sheet at the north-west corner of the Lake district seems to have been about N. 15° W. or N. 20° W. Eastward from this part the direction becomes more northerly; and finally, if we may judge from the form of the ground about Hawes Water, the course taken by the ice there must have been about N. 25° E., a direction which the map shows was maintained from near that part up to the edge of the Dale district.

If we can trust the evidence derived from the direction of transSince this paper was written I have found far-travelled boulders some distance to the north of this line, near Melmerby.

portal of the Scotch boulders, the ice from Kirkcudbright and Dumfries must have flowed across the Solway in a nearly southeasterly direction, as thousands of boulders from the Galloway granite are scattered over the north-west of Cumberland. Such a stream of ice flowing in a south-easterly direction from Kirkcudbright, meeting with the opposing current from the north-west of the Lake district, must have resulted in the sending off of a combined current having a direction nearly magnetic east. About Carlisle the northern part of this stream must have had its direction again modified by the outflow from the local ice of the high ground between Eskdale and Teviotdale, which would cause it to take a turn still nearer to the east, while the southern part of the stream would be forced against the Cross-Fell escarpment and compelled to flow in a direction nearly parallel to it.

Once fairly in the Eden valley the slightest advance towards Stainmoor placed the upper part of the Scotch ice-stream more and more within the influence of the Lake-district ice, until, on meeting with the powerful north-easterly flow from the western side of the Dale district, the current was turned towards the comparatively low part of the escarpment near Brough and forced over on to the eastern side of the watershed.

It can hardly be doubted that the northerly line of fells ranging through High Street to the eastern side of the foot of Ullswater, backed up, as it is, by a still higher range extending from Helvellyn towards the high ground about the Caldbeck Fells, must have exercised a very considerable easterly impelling influence upon the upper parts of the Eden-valley ice. The fact that the striæ found on these fells show that the ice moved in the main along the valleys does not at all disprove the existence of higher currents flowing in other directions.

It has been remarked that one of the greatest difficulties we meet with in trying to account for the drifting of the boulders over Stainmoor is that many of the ice-markings seen in the Eden valley are plainly right across the path taken by the drift. Prof. Ramsay's theory that there were currents flowing in various directions at different levels in the ice-sheet over any given spot helps to explain not only this, but, as will be shown further on, much else connected with the drift that would be difficult to explain in any other way. As before observed, much of the ice that filled the Eden valley came from the high ground to the south; and the easterly drifting of the boulders has been inferred to have resulted from the meeting of the northerly-flowing local ice with the stream that flowed parallel to the escarpment. Therefore it is probable that at low levels the local ice would be pressing outwards far to the north of the currents which, in the higher parts of the great stream, were flowing eastwards full of boulders. In this way it is easy to explain the existence of the scratches at Gathorn near Crosby Ravensworth which are referred to by Prof. Harkness in his paper on the distribution of Shap-granite Boulders. These striæ lie at, or near, the bottom of a valley which lay exactly in the path of the ice coming from the Howgill Fells; so that it is not improbable that an

undercurrent of local ice flowed northwards along this hollow, while at higher elevations the main stream conformed to the general direction taken by the Eden-valley ice.

The influence of the local ice upon the Eden-valley stream generally is well shown near Tebay, where the map (Pl. II.) shows that the southern boundary of Shap-granite dispersal has been deflected a long way to the south, because just about there a considerable depression exists, and the local ice does not seem to have risen high enough to keep the Eden-valley stream so far from the Silurian country as it has done in the neighbourhood of higher ground. So too with the same boulders in the valley of the Bela. Here the north-easterly-flowing ice passed over a deep valley in which the lower layers of the ice were sheltered from the northerly impelling force that affected the upper part; and in consequence, the southeasterly-moving Eden-valley ice, meeting with less cross-resistance, flowed up the valley, striating the rocks and forcing the Eden-valley drifts up the ravine. This is still more evident with the Brockram drift in the same valley; the map shows that there is a deflection of the boundary-line of this drift two miles and a half out of the course it has taken to the south. It is as well to mention that this does not rest upon negative evidence, because the numerous driftsections to the east of the line (A) do not yield a single stone that may not have come from the Carboniferous district to the south.

We have therefore in these cases clear proof that while the higher strata of the ice sheet were moving steadily forward in a north-easterly direction, the lower layers, being sheltered from much of the northerly impelling force that urged on the upper strata, and being at the same time acted upon by a like amount of south-easterly impelling force as affected the ice to the north and south of the Bela, were forced in a direction nearly at right angles to the course taken by the upper part of the ice sheet over the same ground.

Only some such explanation as that given by Prof. Ramsay will account for the fact that while the ice near the low ground in the Howgill Fells and the adjoining parts of the Dale district was shed nearly along the line indicated by the dots on the map, some of the higher strata moved in directions nearly at right angles to the course taken by the underlying streams. It is impossible to give any satisfactory explanation of the stria on Swarth Fell and Baugh Fell if they were not caused by ice flowing from the high ground at the head of Mallerstang. It has been shown that in Mallerstang itself the drift moved towards the north-perhaps because the ice had no other course open to it; but at higher levels, where there were fewer obstructions, the ice would flow everywhere away from the highest ground. That the Baugh-Fell, and especially the Swarth-Fell stria were produced by ice coming from the head of Mallerstang seems to be almost proved by Prof. Hughes finding Carboniferous sandstone in the drift on the north-western side of the highest ground of the Howgill Fells. Most of the adjoining Carboniferous Fell-tops are higher than any part of the Howgill Fells; it is therefore quite likely that some of the higher strata of

the Dale-district ice, laden with Carboniferous fragments, flowed over the Howgill Fells towards the lower ground of the upper part of Lunedale.

Owing to this diversity of directions taken by the various strata of the ice over any one place, it is next to impossible to draw any line that shall indicate precisely where the line of shedding was: that represented upon the map (Pl. II.) must be understood to be an approximation to the shedding-line of only the lower part of the icesheet thereabouts.

The instances of cross-glaciation mentioned as occurring in the Eden-valley seem to show that the component forces that resulted in the easterly turning of the Eden-valley ice-stream must have been very nicely balanced; so that where a stronger current from the south set in, the Eden-valley ice was sent further to the north, and, on the other hand, where the local northward-flowing ice exerted less power, the Eden-valley stream advanced towards the line of iceshedding. In this way we can easily account for the fact that none of the Eden-valley ice, after crossing Stainmoor, found its way down Arkendale into Swaledale. There is nothing whatever in the form of any of the surrounding ground to prevent a marine current from passing that way; on the contrary, every thing seems favourable for a dispersal in that direction; but, as a matter of fact, there is not a single fragment known of any rock in the drifts about Arkendale Head that may not have been derived from the rocks near at hand. One can understand this easily enough when it is known that a thick sheet of ice from the Fells at the head of Swaledale flowed right across the head of Arkendale, and kept the Stainmoor drifts a long way to the north of the line they would have taken had no such ice-sheet been in existence; but how are we to explain these facts by any theory of marine action?

No reference has yet been made in this paper to the dispersal in a southward direction of Shap granite in the basins of the Mint and Lune on the south side of the Lake-district watershed. This was almost certainly owing to the lowness of the ground between the Shap-granite area and the Howgill Fells, in consequence of which the local ice had not sufficient bulk to keep the upper part of the Eden-valley ice-stream far enough to the north to prevent its overflowing into the area south of the watershed. The striated rocksurfaces indicated on the granitic area itself seem to bear out this conclusion, as the rounded sides of the rock face to the north, as if the ice came from that direction. There seems, however, to be some reason for thinking that this cannot always be relied upon as a test of the way the ice flowed, as in one instance, in Garsdale, where it seems tolerably certain which way the bulk of the ice went, the rounded surfaces face away from the source of the drift. This may be due to a local variation in the ice-current; but at any rate it will serve to caution us not to trust entirely to the appearance of the dressed surface as evidence of the direction of flow of the ice.

It seems that there are but few glaciated districts that afford an opportunity of making an approximation to the thickness of the ice

that covered them at the climax of the Glacial Period. But one may venture to assume that the very slightly glaciated rock-surfaces above the 2200 contour-line on the western side of the Dale district prove that the ice cannot have had a much greater thickness than the height of the highest ridge of land over which it can be proved the ice-sheet flowed. Apart from the evidence of the limited amount of glaciation on hard rock-surfaces, we seem to get other evidence in favour of this supposition in the fact that the soft shales and thin interbedded flagstones that occur on the highest ground of the Dale district nowhere exhibit any of that remarkable surface-crushing and contortion so often met with on the surface of beds of the same character at lower levels. After considering all the evidence, there does not seem much probability that the surface of the ice of the Dale district ever rose much above the 2300 or 2400 contour, if it ever was so high as that. If we assume that the ice reached an elevation of 2400 feet above the sea about the line of departure, the fact that it flowed away to the north would seem to prove that the Eden-valley stream must have had a lower surface. Had it been higher, some of the ice from the Dale-district Fells would have been ponded back, and must have flowed southwards from the Eden valley. It is clear that the lower strata of the ice did not do so; but there is only the negative evidence that no Eden-valley driftboulders have gone south of line B anywhere, and that the ground rises more than 1500 feet above the sea, to prove that the upper part of the ice did not flow southwards from the Eden valley.

From what has been stated about the causes that impelled the Eden-valley ice over Stainmoor, it will be seen that the thickness of the ice that came from the southern uplands of Scotland need not have been greater than the local ice of the Lake district. Mr. Ward finds no striæ above the 2500 contour-line. As this ice flowed steadily away northwards, it would seem to prove that in the north-western part of the Eden valley the Scotch ice cannot well have exceeded 2400 or 2500 feet in thickness-a conclusion which harmonizes well with that derived from the investigation of the Daledistrict glacial phenomena.

Glacial erosion.

Such a sheet of ice as that with which the area treated of in this paper was enwrapped must have exercised a very powerful denuding force upon the low-lying parts of the country; but there seems as yet no satisfactory means of determining what thickness of rock was removed from any given spot. One thing, however, is tolerably clear: although the great ice-sheet did undoubtedly deepen many valleys where these happened to lie in its course, there are other cases in which it can be shown that the ice has tended to make them relatively shallower by grinding down the intervening ridges, in districts where the lower parts of the ice were forced across the lines of drainage.

In one part of the Eden valley, near Crosby Garret, we have satisfactory proof that the deep hollow in which Scandal Beck now