31 



tures at Edinburgh. These curves are now submitted to the So- 

 ciety, and the results may be thus briefly stated. 



I. The annual curve of temperature at Trevandrum is a very ir- 

 regular one, but has only one great well-marked inflexion, giving a 

 maximum temperature in the beginning of April, corresponding ex- 

 actly to the period of commencement of the rainy season. The tem- 

 perature of the air goes through its whole range in the course of the 

 following three months ; but there is no decided minimum in the an- 

 nual curve. 



IT. Tlie mean annual temperature of the air appears to be under 

 80°, whilst that of the earth reaches 85° — a remarkable difference. 



III. The accidental, as well as the main annual fluctuations of 

 temperature of the atmosphere, are faithfully reproduced in the curve 

 of temperature at a depth of three feet. They are diminished in in- 

 tensity, and have a slight retardation in point of time, not exceeding 

 ten or twelve days, at the depth of three feet. 



IV. The range at three feet is nearly 2° greater than at the 

 surface. This is a very remarkable circumstance ; and arises ap- 

 parently from the atmospheric maximum being prematurely checked 

 by the approach of the rainy season. 



v. At six and twelve feet, the range is rapidly diminished, and the 

 casual fluctuations almost entirely disappear, leaving maximum in 

 spring well marked. Notwithstanding the deficiency of some of the 

 maximum temperatures, it may be inferred, that the range at six feet 

 is 6i" ; at twelve feet, 2^° ; the range at three feet being 8°. 



VI. The retardation of epoch is also plainly indicated, although, 

 from the deficiencies just mentioned, it cannot be so accurately de- 

 termined. According to the best guess I can form, the three feet 

 maximum occurs about the 8th April ; the six feet maximum about 

 the 20th April; and that at twelve feet about the 11th May; 

 giving a pretty regular rate of progress of the heat downwards, of 

 one foot in three and a-half or four days. 



From these facts, it is easy to deduce, generally, that the pheno- 

 mena of the propagation of that into the ground near the equator, 

 resemble perfectly those in our own latitudes, though modified in 

 extent. Even at a depth of 12 French feet, the annual variation 

 has not nearly vanished, nor would so, even approximately at more 

 than double the depth. Mr Caldecott's experiments conclusively 

 establish the error of the doctrine of Boussingault (at least in the 



