J. G. QoodcMld—On Drift 507 



Glacier de Lechaud, and wliicli was afterwards recovered "with tlie 

 iron not rusted, nor the wood decayed," — all show that under favour- 

 able circumstances bodies may drift for years in the ice, and yet be 

 little the worse for their journey. Why, then, should we assume 

 that a few even fragile shells might not be preserved in the great ice- 

 sheet while they were drifted from the sea-bottoms to the tops of the 

 mountains ? 



When the ice began to melt away, the largest amount of liquefac- 

 tion would take place at the surface, but terrestrial heat must also 

 have had its share in melting the ice at the bottom. 



Wherever channels of any kind existed beneath the ice, — and we 

 have abundant proof that the rock surface itself was very irregular, 

 — the water resulting from the liquefaction of the ice would follow 

 the descent of the surface along these channels towards the lowest 

 ground it could find. By this means the stones and mud which 

 were being liberated as the bottom of the ice melted, would be 

 swept away and the channels kept open. On the banks the case 

 would be otherwise. Here, as the mud and stones melted out of the 

 ice, the accompanying water would drain into the lower channels, 

 while the clay and stones would be left behind, and it is easily 

 conceivable how sheets of water-sorted materials might occasionally 

 alternate with the unstratified clays in places where a shallow 

 channel had been temporarily formed between the continually rising 

 bottom of the ice and the surface of the sediment, if one may thus 

 use the word, which was accumulating between it and the under- 

 lying rock. The detritus which was being liberated from the surface 

 of the melting ice must also have occasionally found its way down 

 through crevasses or otherwise to be mingled with, or arranged 

 amongst, the more clayey deposits which were forming beneath. 



It is clear that close to what one may term the margin of the 

 physical basins of each district, the volume of water of liquefaction 

 flowing past any given spot would be less than would be found 

 nearer the centre and the lower end of each basin, because, in addition 

 to that derived from the melting of the ice in the immediate vicinity, 

 there would be also all that which was flowing downwards from the 

 higher parts of the basin. It would therefore follow, that the farther 

 from the main water-shedding lines, the larger would be the propor- 

 tion of water- worn and water-sorted material. 



High up in the valleys the deposition of the clay and stones, 

 between the continually thinning ice and the underlying rock, would 

 go on uniformly until all the ice had melted ; because there would 

 be little more water passing beneath than that resulting from the 

 melting of the ice immediately above. But where the volume of 

 seaward-passing water was very great, as it must have been in the 

 ' bottoms of the valleys, it might be expected that occasional varia- 

 tions in the velocity of the sub-glacial streams would cause them 

 to cut channels into the older banks of clay which were left when 

 the streams were at a lower ebb. The increased transporting power 

 of the streams would tend to fill these channels with water- worn 

 drift of different degrees of coarseness ; and when the streams again 



