Terrestrial Magnetic Intensity. 369 



ancient volcanic eruption*, to abound in hot ferruginous 

 springs-)*, to be so intersected by trachytes:}:, lavas J, and 

 diorites||, that there is distinct evidence of this tract being 

 nothing else than a crater of elevation, raised by the uphea- 

 ving force of the trachyte of Mont Elbrouz itself f, which he 

 states to be undistinguishable from the rock of Pichincha, the 

 great South American volcano**. Any one who has the 

 slightest acquaintance with the connection between magnetism 

 and volcanic rocks will be at no loss to explain anomalies even 

 greater than those which M. Kupffer has observed. 



42. It was from a persuasion of the entire inconclusiveness 

 of M. Kupffer's results, as well as of all preceding ones, that 

 I undertook the experiments already detailed, in the hope of 

 compensating for the imperfections of the apparatus by the 

 number and extent of the experiments. I own that until I came 

 to calculate the results by the method of least squares, I had 

 little confidence in having obtained any positive result. A 

 careful examination of the station marked on the map, will 

 show that they were almost invariably chosen so that an ele- 

 vated station lay between two others at a lower level, by which 

 the effect of change in latitude and longitude might be elimi- 

 nated. When we criticize these groups of three series, we 

 find for the most part an agreement greater than I had my- 

 self anticipated that the instrument could insure ; yet the com- 

 bination of all with two independent needles, and likewise in 

 two series in different countries and difFerent years, unite in 

 giving a negative coefficient to the height, which I believe to 

 be true and not accidental, though it could not safely be in- 

 ferred from one or two insulated observations. I should be 

 disposed to deduce its probable value thus, taking the circum- 

 stances of the observations into consideration : 



Coefficient of Varied Intensity 

 Weight. for 100 feet of Elevation. 

 Alps, Needle No. I. ... 4 -000033 



Alps, Flat Needle 2 -000027 



Pyrenees, both 1 -000053 



Probable mean -000034 

 Hence to produce a variation of -001, an elevation of 3000 

 feet is necessary. At the height to which Gay-Lussac ascended, 

 the change of intensity would be nearly -008 of the whole ; but 



* Voyage, p. 39. t P- 39, p. 44, p. 55. % P. 44, p. 61, p. 65. 



§ P. CO, p. 66. || P. 63. 1 P. 65. ♦* P. 35. 



Third Series. Vol. 11. No. 68. Oct. 1837. 3 B 



