| Diurnal Variation of Temperature. 103 
the omission of the latest terms these errors will be equated out 
to any desired extent in the same way that this end could be 
attained by graphical methods, but with far more accuracy. 
And a general expression, which sufficiently represents the 
known mean temperature at intervals of an hour throughout 
the day, cannot fail to give that which corresponds to any inter- 
mediate moment with all the exactness needful in the present 
State of science. But since each variable term of the equation 
as above written contains two unknown quantities, its deter- 
mination demands, in the absence of other data, a knowledge of 
the temperature corresponding to two different hours at proper 
intervals, 
