214 THE ATLANTIC. [chap. iv. 



to make the whole soil-cap, heavy blocks included, creep down 

 even the least slope. I will only mention one or two of these. 

 There is constant expansion and contraction of the spongy veg- 

 etable mass going on, as it is saturated with water or compar- 

 atively dry ; and while with the expansion the blocks slip 

 infinitesimally down, the subsequent contraction can not pull 

 them up against their weight. The rain-water trickling down 

 the slope is removing every movable particle from before 

 them ; the vegetable matter on which they are immediately 

 resting is undergoing a perpetual process of interstitial decay 

 and removal. In this way the blocks are gradually borne 

 down the slope in the soil-cap, and piled in the valley below. 

 The only other question is, how the soil is afterward removed 

 and the blocks left bare. This, I have no doubt, is effected by 

 the stream in the valley altering its course from time to time, 

 and washing away the soil from beneath. 



This is a process which in some of the great stone rivers in 

 the Falkland Islands must have taken an enormous time. I 

 fear that the extreme glacialists will see in it a danger to this 

 universal application of their beloved theory to all cases of 

 scratchino; and OToovinoj. I have known too much of the ac- 

 tion of ice to have the slightest doubt of its power ; but I say 

 that ice had no hand Avhatever in the production of these grand 

 moraines in the Falkland Islands. 



In the West Highlands of Scotland, and in many other parts 

 of the world, I have often noticed that w^hen a hill of such a 

 rock as clay-slate comes down with a gentle slope, the outcrop 

 of the vertical or highly inclined slates covered with a thick 

 layer of vegetable soil or drift containing imbedded blocks and 

 bowlders derived from higher levels, the slates are frequently 

 first slightly bent downward, then abruptly curved and broken, 

 and frequently the lines of the fragments of the fractured beds 

 of slate can be traced for a yard or two in the soil-cap gradual- 

 ly becoming parallel with its surface, and passing down in the 



