256 On the Mean Temperature of the 



can otherwise explain this appearance, so much at variance witli 

 our ideas of the distribution of temperatures in the surface of the 

 earth. It is clear, that, if the coldest point of the polar sea is 

 just under the pole, the colder water must move below from 

 north to south, and the warmer on the surface from south to 

 north, the first of these currents would be changed by the rota- 

 tion of the earth into a south-west, the second into a north-east 

 one ; as it must be the surface-water which affects the floating 

 of the ice, it must take place towards the north-east, viz. in the 

 opposite direction, to what really happens. But, if the coldest 

 point is some distance south from the pole, then the surface-cur- 

 rent must take a southerly direction, or rather a south-west, on 

 account of the rotation of the earth. 



I believe, that, in future, we will find more connexion be- 

 tween thephenomenaof currents and the distribution of the earth's 

 temperature; but the latter may exert some influence on the 

 distribution of the intensity of the earth's magnetism. I have, 

 in a former treatise, endeavoured to establish the probability of 

 the magnetism of the earth being seated at its surface ; if this 

 is the case, then certainly the distribution of the earth's tempe- 

 rature must influence that of the magnetic intensity. But we 

 have here the choice of two hypotheses, either the earth is to be 

 regarded as a magnet itself, and then the intensity of its mag- 

 netic power decreases with the increase of temperature, or it 

 receives its power from without, and is, as it were, a mass of soft 

 iron, which is rendered magnetic by a foreign body, and then 

 its magnetic force increases with the heat. 



Although the first of these hypotheses has been that hitherto 

 universally received, yet the second gains some probability from 

 the newly discovered magnetic influence of the sun's rays, and 

 the dependence of the daily changes of the declination on the 

 course of the sun. We will immediately see, that the knowledge 

 we have obtained of the distribution of the earth's temperature 

 gives us a means of deciding the question with greater certainty. 



Let us suppose, first, the globe of the earth as a heated mass, 

 extremely capable of magnetism, and whose surface has almost 

 a uniform temperature, to be rendered magnetic by the power 

 of a distant heavenly body (the sun). It is clear, that the dis- 

 tribution of the magnet will only, in such a body, shew a great 

 regularity, and the lines of equal inclination Avill correspond with 



