1859] WRITINGS OF JOSEPH HENRY. 59 



of a like action are witnessed on the higher peaks of the 

 Appalachian chain. Besides the causes of the general per- 

 turbations of the atmosphere, which we have thus given in 

 considerable detail, some authors have added magnetism and 

 electricity, and others have indeed attributed some of the 

 principal effects we have mentioned to these agencies ; but 

 the present state of science does not warrant us in consider- 

 ing these as true or sufficient causes, except in the case of 

 thunder storms, and perhaps tornadoes, in which the elec- 

 tricity evolved by the action of the storm itself may modify 

 some of the results. Electricity however probably plays a 

 subordinate part ; since it is itself a consequence, and not a 

 cause. 



Terrestrial magnetism has not been shown in any case to 

 affect meteorological phenomena ; it is a force which never 

 produces translation, but merely direction of the needle. 

 The air in its natural condition is not magnetic in the proper 

 sense of the term, any more than a piece of steel wire is so 

 before the power has been developed in it by a magnet. 



We are not allowed in strict scientific investigations, to 

 explain a phenomenon by referring it to any agent, unless 

 we show, in accordance with the laws of that agent, that it is 

 capable of producing the result; and consequently magnet- 

 ism is here not admissible. 



Currents of the Ocean. — We have seen the effect of the un- 

 equal heating of different parts of the earth by the sun in 

 giving rise to great gyrations of air ; and it must be evident 

 that there is a tendency to produce a similar result in the 

 aqueous envelope of the globe. Let us first suppose the ocean 

 to cover the whole earth to a uniform depth, and to be unin- 

 terrupted by continents. If the earth were at rest and the 

 heat of the surface at the equator could extend down suffi- 

 ciently into the depths of the water, the latter would be 

 expanded and would stand higher in the equatorial regions 

 than in those of the poles; a current therefore, as in the 

 case of the air, would be established toward the north and 

 south, from the equator, which would be cooled in its pas- 

 sage, would sink to the bottom, and return again to its 



