Atmosphere and the Earth. 253 



the whole year, break forth from under a covering of snow, how 

 can the small quantity of water, which, in summer, by the melt- 

 ing of the snow and fall of rain, sinks into the earth, elevate the 

 temperature of the subterranean water so many degrees for the 

 whole year ? 



In some places, such as marshes, the mixing of atmospheric 

 water with that of springs is evident ; also loose sands, such as 

 those of the Egyptian deserts, heated by the sun''s rays, elevate 

 the temperature of springs * ; but such observations have been 

 excluded in the above calculations. 



However well the formula a — 6 sin ^ ^ — ^ expresses the ob- 

 servations, yet it must not be forgotten that it is only an approxi- 

 mative formula, and that it may give false results for points far 

 removed from the places of observation. To such places belongs 

 the pole, for which all the four equations must give the same 

 value, but which is not the case. It may be admitted, that, in 

 the vicinity of the pole, the terrestrial temperature is at its 

 minimum, which this formula cannot indicate, as, at Z = it has 

 its greatest value, at Z = 90° its smallest. As the isogeother- 

 mal hne of 32° under the first meridian, approaches very near 

 the pole, if we can trust the formula even reaches it, so that 

 the space, which is included by the isogeothermal line of 32°, has 

 a considerable indentation, and seems almost to form two por- 

 tions, the middle points of which are to be viewed as two dis- 

 tinct poles of cold. One of these points probably lies in North 

 America, the other in the north of Siberia. Unfortunately we 

 still want observations on these places. The temperature under 

 these poles of cold cannot be much under 32°. 



With respect to the temperature at the equator, we see that 

 these points which are on coasts washed by the sea, or on islands, 

 have a lower temperature than those which are in the middle of 

 a great continent. The warmest point of the equator is in the 

 interior of Africa ; to the north of this point, at least in lati- 

 tudes which do not exceed 50°, the isogeothermal lines have a 

 considerable curve to the north. That point, which, in the 

 ocean included between great continents, falls in 60° east longi- 

 tude, has even a temperature of 3'37 lower. Those points, 

 finally, which lie next to the observations made on the west 

 coast of Africa (TenerifFe) and east coast of America (Cuma- 

 • Well at the great Pyramid fi8-2,'). 



