478 Mr. Drach on the Horary Deviations of the Barometer 



The discrepancies in the other pairs are even removable 

 by referring to plate v. in the volume, where it will be no- 

 ticed that, first, the observed deviation (denoted by a *) must 

 be reduced to that of the curve by adding to 



2 a.m. about —15 3 p.m. about — 5 



9 +10 4 + 3 



5 -10 



and that, secondly, the p.m. upper curve appears also to be 

 somewhat too flattened at its summit in order to pass exactly 

 through the observed 10 p.m., which is remediable by apply- 

 ing to 



10 p.m. about 4-10, and to 11 p.m. about +7. 

 With these corrections, the sums of the homonymous hours 

 become for 



Hence may a rule be deduced of great value for observa- 

 tories and single travellers : — 



The mean height of the harometerfor the day of observation 

 is exactly equal to the average of four observations made at in- 

 tervals of six hours, whatever be the time of the first observatiim. 



This rule is far more accurate than the average of the 

 maximum and minimum, or of those at 9 hours and 3 hours, 

 4 hours and 10 hours, &c. ; as probably the time of mean 

 pressure is not the same for all points of the globe. The 

 same rule is applicable to thermometrical observations. In- 

 deed the height may be generally represented by the conver- 

 ging series h = 2o°°?«j^', and these may to some extent be 

 reduced to functions of sin it, cos it, leading to the rule in 

 question, as will now be shown. 



The Plymouth observations may be represented by the 

 formula h = H + A s\n t + a cos t + B sin 2t + b cos 2 1 

 + C sin 3 ^ + c cos 3 ^ + E sin 4 ^ + ^ cos 4 1, where // = the 

 height for the time t, reckoned as above ; H, A, a, &c. are 

 constants. 



Denote the height for t + — = ^ + 6 hours by h' 



t + TT = t + 12 ... h" 



t + ~ = t+ 18 ... h'" 



