PROFESSOR FORBES ON THE TEMPERATURE OF THE EARTH. 197 



a view to its occasional verification ; the freezing point being accurately ascer- 

 tained. Nine years after (Januaiy 1846), this instrument being re-examined, 

 shewed no appreciable change in the position of the zero point. It could not 

 have amounted to A of a degree. A common thermometer would have altered 

 appreciably under similar circumstances. Is not this owing to the strength of the 

 glass bulb ? 



The permanent influence of the pressure of sand in the holes was suggested 

 to me as a possible source of a change of figure of the instruments ; but the con- 

 clusive experiments lately made in America and France on the pressure of sand, 

 convinced me that this could have no appreciable influence. 



The observations were made weekly ; generally on Mondays. As it was com- 

 monly supposed that the diurnal variations of the temperature disappear at the 

 depth of three feet, I did not take particular precautions to have aU the obser- 

 vations made at the same hour. I find, however, that, at the Observatory, the 

 readings were taken immediately after noon, at the Experimental Garden, about 

 2 o'clock, at Craigleith Quarry, regularly between 11 and 12 o'clock. The later 

 hour of the second series may account for some irregularities observable. The 

 obsei-vations, at all the stations, were made regularly from February 1837 down 

 to May 1 842, about which time, the thermometers at Craigleith were maliciously 

 destroyed ; but these five years of complete observations have yielded all the results 

 which were looked for in commencing them. The boxes covering the thermome- 

 ters, in the Experimental Garden, were blown over in the winter 1844-5, crushing 

 the thermometers. Those at the Observatory stUl remain in good order, and are 

 regularly observed. The numbers, from the original registers, are exactly given 

 in the table at end of this paper, together with the corrections applied to them, in 

 the manner to be described in the next section. It is not to be supposed that the 

 registers are without some errors, at least, in the case of the two less experienced 

 observers (at Experimental Garden and Craigleith) ; but they are only oversights 

 of the eye or hand, which can affect none of the conclusions, as the admirable coin- 

 cidence of the three independent series, in Plate VII., sufiiciently proves. 



The following table contains the data necessary to be known, respecting the 

 scales and dimensions of the thermometers, for correcting the temperature of the 

 stems and exposed columns in the manner which we shall next proceed to inves- 

 tigate. 



VOL. XVI. PART II 3d 



