-1859] WRITINGS OF JOSEPH HENRY. 177 



If our instruments consist of a maximum and a minimum 

 self-registering thermometer, exposed to the air in the way 

 we have indicated, it will be sufficient in order to obtain the 

 average temperature of the day approximately, to note the 

 temperature of each but once in twenty-four hours. If we then 

 add together the maximum and minimum, and divide the 

 sum by two we shall have approximately the average tem- 

 perature; but this is not precisely the quantity required for 

 meteorological and agricultural purposes, or that which en- 

 ables us to judge of the heat of different days or different 

 periods, since the thermometer may at different times of the 

 day be suddenly elevated or depressed and not reach its 

 maximum and minimum gradually, as is usually the case. 



To determine these points with more precision, and the 

 average temperature of the air during the day, we must ob- 

 serve the thermometer at very short intervals; for example 

 every quarter of an hour. If we add these into one sum 

 and divide by ninety -six we shall have the mean or average 

 temperature of the day. Before division however caution 

 is to be observed in combining the observations taken in 

 winter, or when the temperature sinks below zero, to subtract 

 the sum of the observations with the minus signs from the 

 sum of those with plus signs. 



In running our eye down the column of a series of obser- 

 vations of this kind, we can mark not only the maximum 

 and minimum temperature for the day, but also the time 

 at which they occurred. If we continue these observations, 

 during the month of thirty days for example, we shall obtain 

 thirty maxima and as many minima, and an equal number 

 of mean temperatures. If we now add these thirty observa- 

 tions of the same kind together, and divide by the number 

 thirty, we shall obtain the maximum, the minimum, and the 

 mean of the month. Similar observations continued through- 

 out the year and thus combined will give us the mean of all 

 the maxima, of all the minima, as well as the general means 

 of all the three hundred and sixty-five or three hundred and 

 sixt3''-six days of which the year may be composed. 



There is still another way of combining these observa- 



12-2 



