376 Dr BREWSTER on the Register of the Thermometer kept 



The determination of the exact times of mean temperature 

 throughout the year, furnishes us with the two best times of the 

 day for recording the indications of the thermometer. These 

 times at Leith are obviously 9 h 13' A. M. and 8 h 27' P.M. ; for, if 

 any of the observations is accidentally omitted at one of the hours, 

 the mean of the remainder will approach nearer to the mean 

 temperature of the year, than if any other two hours had been 

 taken, and similar omissions made. 



There is, however, another advantage of this determination, 

 namely, that the mean temperature of the year may be obtained 

 with great accuracy by a single observation made every day at 

 one of the times of mean temperature. Let us suppose that we 

 wished to determine the mean temperature of the year 1825 at 

 Leith, and that we had possessed no other observations than a 

 single daily one made at 9 h 13', then the mean of these 365 daily 

 ones would have been 48.944, the very same result that has been 

 obtained in 1825 from 24 observations every day. 



If we examine the annual curve, and also the monthly curve, 

 it will be seen, that the ascending or morning branch is more re- 

 gular in its progression than the descending or evening branch. 

 On this account, we would prefer a single observation every day, 

 made at the time of the morning mean, to a single observation 

 made every day at the time of the evening mean. 



It must be carefully observed, that the hours of mean tem- 

 perature which we have now been considering, are only mean 

 results for the whole year. If we wished to deduce the mean 

 monthly temperatures from an observation made once a-day, it 

 would not answer to take 9 h 13' A. M. and 8 h 27' P. M. ; for the 

 times of mean monthly temperature occur at different hours of 

 the day throughout the year, as will appear from the following 

 Table : 



