[ % ] 



fpeak with marhematical accuracy it is liable to the fame ob- 

 jeftion as the laft, viz. that a fcale cannot be applied to it ; for 

 countries, however, of moderate extent in latitude, it is the beft 

 that can be devifed. 



In the circular projed:ion the parallels of lat. are circles, de- 

 fcribed from the pole as the common centre, with radii equal to 

 the complements of their refpedtive latitudes : the meridians pre- 

 ferve their due diftances on the parallels, becoming gradvially 

 curvilineal, and cutting the parallels more and more obllqiiely, 

 as they recede from the central meridian ; but as this alone can 

 be a right line, and cut the parallels perpendicularly, it is ob- 

 vious, that this projedlion muft produce diftortlon in the map, 

 and confequently an error in its area, which, if it is greatly ex- 

 tended in longitude, will at length become enormous : for coun- 

 tries, however, of moderate dimenfions, it is one of the leaft 

 objedlionable projedions, and a fcale may be adapted to it with 

 fufficient accuracy, though not with geometrical preciiion. 



There is another method of projedion, ufed by the San- 

 fons. in which the parallels are right lines, and the meridians, 

 which obferve their due dlftances on the parallels, are parabolic 

 curves. This is admxffible only in delineating countries of in- 

 confiderable extent in longitude, and Mr. I '/inville has preferred 

 It for his map of South America: In all other cafes it produces 

 much greater diftortlon than any of the above methods. 



