310 CELESTIAL DYNAMICS. 



carried off by all the thermal springs in the world is very 

 small in comparison with the quantities which we have to 

 consider here. 



Much more important is the effect produced by active vol- 

 canos. As the heat which accompanies the molten matter to 

 the surface is derived from the store in the interior of the 

 earth, their action must influence considerably the diminution 

 of the earth s heat. And we have not only to consider here 

 actual eruptions which take place in succession or simulta 

 neously at different parts of the earth s surface, but also vol- 

 canos in a quiescent state, which continually radiate large 

 quantities of heat abstracted from the interior of the globe. 

 If we compare the earth to an animal body, we may regard 

 each volcano as a place where the epidermis has been torn 

 off, leaving the interior exposed, and thus opening a door for 

 the escape of heat. 



Of the whole of the heat which passes away through 

 these numerous outlets, too low an estimate must not be 

 made. To have some basis for the estimation of this loss, 

 we have to recollect that in 1783 Skaptar-Jokul, a volcano in 

 Iceland, emitted sufficient lava in the space of six weeks to 

 cover 60 square miles of country to an average depth of 200 

 metres, or, in other words, about 1 cubic miles of lava. 

 The amount of heat lost by this one eruption of one volcano 

 must, when the high temperature of the lava is considered, 

 be estimated to be more than 1000 cubic miles of heat ; and 

 the whole loss resulting from the action of all the volcanos 

 amounts, therefore, in all probability, to thousands of cubic 

 miles of heat per annum. This latter number, when added 

 to Fourier s result, produces a sum which evidently does not 

 agree with the assumption that the volume of our earth has 

 remained unchanged. 



In the investigation of the cooling of our globe, the influ 

 ence of the water of the ocean has to be taken into account. 

 Fourier s calculations are based on the observations of the in- 



