g Gen. Roy’s Account of 
firft to be 482, and the laft 403 feet, above low water at the 
fea, the refraction on the morning of the 1 ft of September, 
i 7 8 7 , as obferved at Shooter’s Hill, was which is be- 
tween f d and |th part of the contained arc. 
If the circumftances had permitted the reirachon on the 
diftance between Dover and Calais to have been repeatedly 
tried, at the bottom of the cliff, at the top of the cliff, and 
again at the top of the caftle, we (hould probably have found 
the refraftion at the three ftations confiderably different, with 
tire fame length of arc, or one only varying infenfibly. 
But, in order to be enabled to make conclufive obfervations 
of this fort, the. operation fhould become a diftinfft one, or at 
moft only comprehend fuch others as aie conneaed with the 
modifications of the atmofphere. For purpofes of this kind 
very fine levels would be requifite ; and fome of the higheft 
mountains of Scotland, fituated near the fea, fuch as Ben 
Nevis and Cruachan Ben , where the relative heights of the 
ftations might be accurately afcertained by levelling, would 
feem to be eligible fituations. 
Tablb 
