as applied lo the measurement of Heights. 343 



After watching the sympiesometer in its most fluctuating 

 state, it is pleasing to find, in a series like the one before us, 

 that in its remarkable steadiness, which is greater than I ever 

 observed, our supposition is completely backed by the unifor- 

 mity of temperature very singular for such a time of day. The 

 cistern was opened at the instant of the first observation at 

 the lower station, and after two more observations, it seems to 

 have acquired the temperature of the air, which continued for 

 45^ with an extreme variation of only 0°.8, which occurred but 

 in one instance. We may therefore easily account for the ex- 

 traordinary unity of the columns of the sympiesometer, though 

 it is not so easy to explain why the height is so much greater 

 than it ought to be ; but that is not to our present purpose. 

 The only variation which breaks the column of inches between 

 Obs. 72 and 82, is of .01 at Obs 77, and is immediately suc- 

 ceeded and accounted for by a sudden fall of 0°.4 of the ther- 

 mometer, which is actually double of any fall of the thermome- 

 ter within the same limits. This seems to prove to demon- 

 stration the accuracy of our explanation. 



Table VII.— -/wZ^ 30^^, 1825. 



95 — 3.35 74.5 29.63 196 26 



