1 54 Mr J. D, FORBES on the Horary Oscillations 



presents to us, is the analytical method of discussing observations, 

 extracting, as it were, from the data which nature presents to us, 

 laws which shall best represent them, instead of framing clumsy 

 hypotheses, to which it became a secondary object, to apply the 

 facts. No better example of the method can be given, than 

 HUMBOLDT'S masterly essay on Isothermal Lines ; and the same 

 distinguished traveller was the first who threw into anything like 

 lucid order, the class of phenomena which it is the object of this 

 paper to discuss, the Horary Oscillations of the Barometer. 



3. These atmospheric tides were first discovered in a decisive 

 manner at Surinam, in 1722, by a Dutch philosopher, whose 

 name has not descended to us ; and the history of their subse- 

 quent development is too well known to require even an abstract 

 in this place. It has been reserved for the nineteenth century 

 to ascertain almost every thing connected with them besides their 

 bare existence, and even that was not established in the tem- 

 perate zone previous to the observations of RAMOND, little more 

 than twenty years since. But it was reserved for HUMBOLDT, 

 after much practical labour in the first years of the century, to 

 draw up the admirable analysis of the subject which appeared in 

 the third volume of the " Relation Historique," published in 

 1825. From that time a new impulse was given to the labours of 

 observers ; and during the last five years many important registers 

 have thrown light upon the modification which the barometric 

 oscillation undergoes in different latitudes, and at various heights 

 above the sea. The subject had always appeared to me one of 

 singular interest, and accidental circumstances first induced me 

 to prosecute it experimentally, during some months' stay at Rome 

 in the spring of 1827- The period was very short, but as I fre- 

 quently made from twelve to fifteen observations in a day, I was 

 enabled, very satisfactorily, to trace the variation from hour to 

 hour, and to ascertain the periods of maximum and minimum, 

 though too brief to determine the amount accurately, which came 



