J. G. GoodcJiild — Motion of Land-ice. 21 



should do Tnuch to convince even those who are not field-geologists 

 that the lower parts of masses of laud-ice really do move forward, 

 and are therefore quite competent to score the rock-surface over 

 which they pass, and to score it to almost any extent. 



Let us consider what happened with any of the larger affluents 

 of the Ice-sheet in, say, North West Yorkshire, or in Cumberland 

 and Westmorland. During the Glacial period every one of the 

 valleys in those parts was filled to the brim with masses of ice, 

 which, in most cases, can be shown to have moved outwards from 

 the mountain centres towards the lowlands. The inland and uphill 

 movements, which affected certain parts, were exceptional, and took 

 place only at or near the climax, just before the Ice-sheet began to 

 melt away, and long after most of the glacial erosion had been 

 accomplished. There is reason for believing that for a long period 

 the thickness of the ice in places was not less than two thousand 

 feet. All the rock previously carried seaward by subaerial denuda- 

 tion from those valleys was thus replaced, as it were, by a solid 

 mass of other material, — ice. Doubtless, in those days, as is the case 

 with the smaller glaciers of to-day, the surface layers were slowly 

 impelled downhill and seawards by the alternate expansion and 

 contraction of the ice, aided by the expansive action of water 

 freezing in crevasses, as well, of course, by the pressure of the 

 higher masses of ice behind. All these forces, acting separately or 

 in combination, tend to make the line of swiftest flow coincide with 

 the centre line of the upper layers of each stream. 



The tendency to flow most rapidly nearest the surface is, however, 

 counteracted by other causes. Every one seems to have overlooked 

 the fact that when a deep valley is filled by ice, that ice tends to 

 act in much the same way in conducting terrestrial heat outwards 

 towards the surface as if there were rock there in the valley instead 

 of ice. The isogeotherms in such a case, instead of following nearly 

 the contour of the valley as they would do if the ice were absent, 

 would be carried to a higher level, whose exact position would vary 

 with the thickness of the ice and with the surface-temperature for 

 the time being. In other words, the lines of equal subterranean 

 temperature of the rocks forming the sides of the valley would 

 enter the adjoining ice itself, and would connect the isogeotherm of 

 one side of the valley with its corresponding line on the other by 

 a downward curve, which would run through the lower part of the 

 ice, instead of through the rock below upon which it lay. 



While the surface-temperature and the terrestrial-temperature 

 were alike in degree, it seems possible that there might be no 

 movement of the ice at all, except that due to the downhill impulse 

 derived from heavier masses at a higher level. Where the surface 

 conditions were such as those described near the commencement of 

 this paper, then the locus of swiftest flow tended to the surface. 

 But when, as must generally have happened during the Glacial 

 period, the surface of the ice was kept in a prolonged state of 

 contraction through extreme cold, the higher temperatures prevailing 

 in the parts of the ice nearer the bottoms of the valleys caused the 



