AND THE OTHER PLANETS OF THE SOLAR SYSTEM. 655 



must be a reason for a more approximate equability of temperature in those regions than if the 

 transference of heat were entirely vertical. Thus in the equatorial region the heat must constantly 

 be carried by atmospheric currents and other causes both to the north and to the south, and the 

 mean temperature of that region must be materially diminished, while that of the regions imme- 

 diately contiguous must receive a corresponding increase. This is probably the principal cause of 

 the anomaly which has given rise to these remarks. Let us examine how far it will be removed 

 by an assumed amount of the change in the values of the mean temperature at the equator 

 and on the parallel of 30° N. For this purpose let us take the following values of u , instead 

 of those above given (Art. 21), as assumed approximate values of the mean annual tempera- 

 tures of different parallels, supposing that no horizontal transmission of heat should take place 

 by atmospheric or local causes, so that the corresponding values of s may give the real effect 

 of the Sun on the temperature of the air in different latitudes, in addition to the effect 

 produced upon it by the heating of the surface of the Earth, arising from the solar heat which 

 radiates directly upon it. 



m = 30° (C) at the Equator, 

 = l6°,50 in lat. 30, 



= 6°,35 48°. 50', 



= -7° at the Arctic Circle, 

 = - 15° at the Pole. 



I have here added 3°,5 to the observed mean temperature of the equator, and deducted 4°,5 

 from that of the parallel of 30° N., which is rather higher than the mean of the north and 

 south parallel of 30°. I have increased the temperature at the pole by 1°,5 : the other tem- 

 peratures are not altered. We should thus have a depression of temperature of 13°,5 (C) in 

 passing from the equator to the parallel of 30° ; while it would equal 20°,5 between that 

 parallel and that of 60°, and 11° between the latter parallel and the pole. 



23. We may also here make another correction of which we know the sign, but can only 

 assume the amount. Poisson has calculated the numerical value of h by observations made at 

 Paris ; but since it depends on the intensity of solar radiation after the heat has passed through 

 the atmosphere, he suggests, as before stated, that its value must be somewhat greater at the 

 equator and somewhat less at the pole than at Paris. Instead of using 34°,4 for the value of 

 hQ at the equator, I shall now use 36°; and instead of 14°,4 at the pole, I shall substitute 

 12°,4. This is equivalent to increasing the value of h about 1°,7 at the equator, and dimi- 

 nishing it about 5° at the pole. Changing the values of h, also, for different latitudes in a 

 proportionate degree, we shall have 



hQ= H = 36° (C) at the Equator, 



H x = 29° in lat. 30°, 



H 2 = 23°,76 in lat. 48°. 50', 



H a = 16°,5 at the Arctic Circle, 



H t = 12°,4 Pole. 



Vol. IX. Part IV. 84 



