NATU RAL HI STORY of NORWAY. 49 



redtion, then winds off to the right with afcents and defcents, 

 and in fome places growing narrower, and in others widening to 

 double its former breadth, which, according to my admeafure- 

 ment, was about four or five ells, and the height about three; 

 thus two perfons could go abreaft, except that they were now and 

 then obliged to ftoop, and even creep, and then they felt a damp 

 vapour like that of a burial-vault. This prevented my penetrating 

 fo far as I had intended. Another thing remarkable, was the ter- 

 rible roaring of the waters under us, the courfe of which was what 

 moft excited my wonder, as over it lies a pavement of fmooth 

 ftone, inclining a little like a vault on each fide, but flat in the 

 middle, and not above three fingers thick, with fome fmall cre- 

 vices, through which the water may be feen. If it be afked how 

 far this covered -wayreaches ? I make no queftion but its length 

 is equal to the courfe of the ftream, and that it has been pro- 

 duced by the falling of the water, which in length of time, has 

 perforated thefe rocks agreeably to the ancient maxim, 



Gutta cavat lapidem, non vi, fed fsepe cadendo. 



And this is more particularly confirmed by the many projections 

 which have been levelled, or undulated ilgures, which, as I have 

 before obferved, are to be feen on the roof, and along the fides. 

 If it be afked again, where is the fpring of this ftream ? the pea- 

 fants hereabouts fay, that on the uppermoft ridges of the moun- 

 tain,which is at leaft a hundred fathom high, almoft perpendicular 

 above the cavern, there is a ftanding-water of about a quarter of a 

 league in circumference, and unqueftionably formed and fupplied 

 by the frequent accefllon of the rain, and the melted fnow from the 

 other parts of the mountain. It is no difficult matter to judo-e 

 how the uppermoft dry vault comes to be of fuch a heio-ht over 

 the channel of the river, by which it is caufed; for the cavity in 

 its beginning could not have been fo high, but by length of time, 

 the ftream, of which the upper vault was then the bed, penetrated 

 to its prefent depth, and perforating the mountain, the particles 

 which it detached, as fend and gravel, fettled on the ground, 

 forming as it were a fmall and level pavement, which k now a 

 cover to that ftream, of which it had been the bed. I am the 

 more confirmed in thefe thoughts, by a fecond view I took of 



this 



