696 [June 18, 



same hour during the 6 months by the number of observations, or (2) 

 by taking the arithmetical means of the monthly means. Which- 

 ever method be adopted, the results for some of the hours will, from 

 paucity of observations, be liable to much uncertainty. As a check 

 upon results obtained by these two methods, and to remove errors 

 arising from the greater average strength of electricity in some 

 months than in others, I have (3) divided the sum of observations 

 at each hour for each month by the mean of observations " at all 

 hours" for that month, and, after adding the corresponding sums 

 for the 6 months, have divided by the number of observations. This 

 method of reduction I conceive to be the fairest of the three, as it 

 amounts to multiplying each observation by a factor inversely pro- 

 portional to the mean electrical potential for the month in which it 

 is taken. It gives the mean potential at each hour, supposing the 

 general mean derived from observations at all hours to be unity. 

 The means obtained in this way are headed " Reduced Means " in 

 the annexed Table of Diurnal Range, those obtained by the other 

 two methods being given in the two preceding columns. 



All three methods agree in furnishing the following results : 



1. That between 7 and 8 A.M. the strength of electricity is below 

 the mean. 



2. That between 8 and 9 A.M. its strength is above the mean, and 

 takes a very decided maximum. 



3. There is apparently a minimum between 10 and 11 A.M. ; but 

 observations are few. 



4. That from 1 to 7 P.M. the strength is above the mean, with 

 the apparent exception of the hour 4-5. 



5. That from 7 P.M. to midnight the strength is below the mean. 

 The mean here referred to is the mean of all observations, and is 



probably not the true mean value for the 24 hours. 



[Subjoined is a selection from the Tables which accompany the 

 Paper; the complete series is preserved in the Archives of the 

 Royal Society.] 



