252 Mean Temperature of twenty-seven different places 



so that all the labour of doubling 6 and c and adding d may 

 be saved, without perceptibly affecting the daily averages. 



We may therefore regard the daily averages in ihe report 

 as equivalent to the mean of the three ordinates of the 

 daily curve at &^ a. m., 3'^ p. m., and an hour past sunset ; 

 so that if we suppose an hour after sunset to be a little colder 

 than the mean temperature of the day, the periods adopted 

 should give very nearly the mean daily temperature ; for as the 

 maximum temperature takes place about 3'^ p. m., and the mi- 

 nimum before Gi^ a. m., the mean of these two will be a little 

 higher than the mean temperature of the day, and this mean 

 being again combined with the observation after sunset, which 

 we suppose a little lower than the mean temperature, will give 

 a result not very far from the mean temperature required. 

 Had the third observation been at a fixed hour, we could 

 have calculated exactly by means of the Leith results the dif- 

 ference between the averages in the report and the mean tem- 

 perature of the twenty-four hours. 



The following table contains the mean monthly temperature 

 of the twenty-seven places above-mentioned, the annual mean 

 temperature, the annual range, and the highest and lowest 

 during the year. 



In the year 1826 the mean temperature of ten of the above 

 places was _ _ - _ 490. 4 



In 1828 the mean temperature of twenty-three places is 49 .99 



The mean temperature of a point in the State of New 

 York, corresponding to the mean position of all these 

 places, is, according to Dr Brewster's general Formu- 

 la, - - - - - 49 . 8 



So that the formula must err in defect, as the mean altitude of 



the different places must be considerable. 



