MANNER OF TAKING THE MEAN. 135 



equal intervals during the day. From this, too, 

 it is evident that this process can only give correct 

 results, when the observations have been made in 

 number sufficient to give the true form of the 

 curve. 



Since it is very laborious to continue so great 

 a number of observations through a long period, 

 methods have been contrived which render it 

 sufficient to note the state of the thermometer 

 but a few times in the day. Experience has led 

 us to several aids of this kind, of which, indeed, 

 none serves with certainty for the determination 

 of the mean temperature of a single day ; but 

 several of them, when the object is to find the 

 mean of longer periods of time, are extremely 

 useful, and among these the following are the 

 best. 



Humboldt first remarked that the sum of the 

 highest and lowest temperatures in the day, 

 divided by two, gives almost exactly the true 

 diurnal mean. Thus, by the use of a good re- 

 gistering thermometer, the mean temperature of 

 each day may be readily determined. 



A second method consists in reading off the 

 height of the thermometer at any three or four 

 points of time, chosen at equal intervals : for 

 instance, at eight o' clock in the morning, at four 

 in the afternoon, and at midnight, or at four and 

 ten in the morning, and at four and ten in the 



