( 393 ) 



curve goes through its extreme range in three months ; the principal minimum 

 occurring about the middle of July. The remaining fluctuations are compara- 

 tively insignificant, but indicate a slight maximum about the middle of October. 



III. The epochs of temperature are retarded with the depth below the sur- 

 face in the usual manner, and, at the same time, casual fluctuations disappear, 

 and the ranges diminish. At 12 French feet, the principal maximum occurs five 

 weeks later than in the open air, and the range is still at least a degree and a half. 



From these facts, it is easy to infer that the phenomena of the propagation 

 of heat into the ground near the equator resemble those of temperate latitudes, 

 though modified in extent and character. Mr Caldecott's experiments con- 

 clusively establish (as he himself has pointed out) the error of the doctrine of 

 BoussiNGAULT (at least for the eastern hemisphere), that the annual tempera- 

 ture near the equator remains unchanged at the depth of a foot below the surface 

 in the shade. This mistake it is the more important to correct, because M. Poisson 

 has attempted to confirm his mathematical theories of heat by applying them 

 to this alleged fact.* 



Mr Caldecott's experiments appear farther to prove a considerable excess 

 of the temperatm-e of the earth above that of the air at Trevandrum. This result 

 is in opposition to the opinion of Kupffeb, which supposes the earth temperature 

 to be less than that of the air between the tropics, and that of Boussingault, 

 which supposes them to be the same. 



The results of Mr Caldecott are confirmed in both particulars by Captain 

 Newbold of the Madras Army, in a paper lately published in the London Philo- 

 sophical Transactions, f 



Theorie de la Chaleur, p. 508. 



t For 1845, p. 125. 



