360 SIR DAVID BREWSTER ON THE HOURLY 



The hours of mean temperature have a considerable range in the monthly 

 curves, varying in the morning from half-past 8 to half-past 10, and in the even- 

 ing from 7 o'clock to 9. 



IV. On the relation between the Mean Temperature of the Day, and that of any single 

 Hour, or pair of similar or homonymous Hours. 



It was long the practice of meteorologists to observe the thermometer at two 

 convenient hours, so that if the one gave a temperature greater than the mean, 

 the other might give a temperature as much less, and in this way several registers 

 were kept with considerable accuracy. The hours of 10 h a.m. and 10 h p.m., sug- 

 gested by the Rev. Dr Gordon, were frequently used, and gave a result nearer to 

 the mean of the maximum and minimum than any other pair of convenient hours. 



Upon computing the mean temperature of every pair of similar or homonymous 

 hours, I found, as shown in the following Table, that they differed very little 

 from the mean temperature of the 24 hours : — 







Diff. from Mean 



Temp, of Day in 



Hours of Observation. 



Thousandths of a Degree. 







Leith. 



Inverness. 



5 h a.m. and 



5 h P.M. 



-0-134 



-0-434 



6 6 





-0281 



-0-543 



7 7 





-0372 



-0-552 



8 8 





-0421 



-0-396 



9 9 





-0-285 



-0-113 



10 10 





-0086 



+ 0174 



11 11 





+ 0176 



+ 0-374 



12 12 





+ 0-374 



+ 0-555 



1 1 





+ 0-367 



+ 0-550 



2 2 





+ 0-366 



+ 0-389 



3 3 





+ 0-252 



+ 0173 



4 4 





+ 0059 



-0-175 



Hence it appears that the defect or excess of the mean temperature of any pair 

 of similar hours, when compared with that of the 24 hours, is always in the 

 Leith observations less than half a degree. It appears, also, that the mean of 4 h 

 and 4 h approaches nearest to the daily mean, and 10 h and 10 h next to it. 



I have added to the above Table the results of the Inverness hourly observa- 

 tions. The deviations are very slightly greater, but the law is the same ; and it 

 is interesting to observe the interchange of the signs at 10 h and 10 h , and 4 h and 4 h , 

 a proof of the singular equality between the mean temperature of the day, and 

 half the sum of the mean temperature of these hours. 



In speaking of this law, as given in the Report upon the Registers for 1824 

 and 1825, Humboldt says, — 



" We are surprised, at the first glance, by the generality of this law. The 

 homonymous hours are very inequally distant from the hour of the maximum of 

 the daily temperature It is a thing truly remarkable, that from the 



