346 DISTRIBUTION OF BOULDER STONES. 



must not confound rocky masses which pierce the soil in 

 the midst of the desert, and of which mention has often 

 been made by travellers, with mere scattered fragments. 

 These facts seem to prove, that the blocks of Scandina- 

 vian granite, which cover the sandy plains on the south- 

 ern side of the Baltic, in Westphalia, and in Holland, 

 are owing to a particular debacle which proceeded from 

 the north, to a purely local catastrophe. The old con- 

 glomerate (gres rouge), which covers a great part of the 

 Llanos of Venezuela and of the basin of the Amazon, 

 contains, without doubt, fragments of those same primi- 

 tive rocks of which the neighbouring mountains are com- 

 posed ; but the convulsions of which these mountains 

 present undoubted evidences, do not seem to have been 

 accompanied with circumstances favourable to the trans- 

 portation of great blocks. This geognostic phenomenon 

 is so much the more unexpected, that nowhere in the world 

 does there exist a plain so continuous, and which is pro- 

 longed with fewer interruptions to the abrupt declivity 

 of a purely granitic cordillera. Before my departure 

 from Europe, says Humboldt, I had already been struck 

 with the observation that there are no primitive blocks 

 in Lombardy, nor in the great plain of Bavaria, which 

 i$ the bottom of an ancient lake, having an elevation of 

 250 fathoms above the level of the ocean. This plain is 

 bounded on the north by the granites of the Upper Pa- 

 latinate, and on the south by the alpine limestones, tran- 

 sition clay-slates, and mica-slates of the Tyrol. 



Boulders, or loose blocks of alpine rocks, are found in 

 the lower part of the Alpine valleys, which terminate in the 

 great principal valley that stretches between the Alps and 



