396 COSMOS. 



i. e. assume a direction from east to west, as bismutli and 

 phosphorus, but that this property is most feebly exhibited 

 in oxygen ; it has been shown by his latest researches, "which 

 were begun in 1847, that oxygen alone, of all gases, like 

 iron, assumes a position from north to south ; and that oxygen 

 gns loses a portion of its paramagnetic force by expansion, 

 and by elevation of the temperature. Since the diamagnetic 

 activity of the other constituents of the atmosphere, such as 

 the nitrogen and carbonic acid, are not modified by expan- 

 sion or by an elevation of temperature, it only remains for 

 us to consider the oxygen " which surrounds the whole Earth 

 as it were, like a large sphere of sheet-tin, and receives mag- 

 netism from it." The half of this sphere which is turned 

 towards the Sun, is less paramagnetic than the opposite half; 

 and as the boundaries of these halves are constantly altered 

 by their rotation and revolution round the Sun, Faraday is 

 inclined to refer a portion of the variations of terrestrial mag- 

 netism on the Earth's surface to these thermic relations. The 

 assimilation thus shown by experiment to exist between a 

 single gas (oxygen) and iron, is an important discovery of our 

 own age, 30 which derives additional value from the fact that 

 oxygen probably constitutes the half of all the ponderable 

 matters that occur in accessible portions of our Earth. 

 Without assuming magnetic poles in the Sun's body, or any 

 special magnetic forces in the solar rays, the central body 

 may, as a powerful source of heat, excite magnetic activity on 

 our planet. 



The attempts that have been made to prove, by means of 

 meteorological observations prosecuted for many years at 



30 Faraday upon atmospheric magnetism, in the Exper. 

 Researches on Electricity, series xxv. and xxvi. (Philos. 

 Transact, for 1851, part i.) 2774, 2780, 2881, 2892, 

 2968, and for the history of the investigation, 2847. 



