168 



EIGHTH REPORT — 1838. 



Table LXVIII. 



Hence, instead of comparing the other results of C3lincier L (A) 

 with the mean of the initial and final observations in Dublin, 

 they are to be compared with the initial observations alone ; the 

 final observations not being comparative with the rest of the 

 series. The loss of force sustained by the cjdinder L {b) being 

 •006, the amount of the correction is 



5/, = --003 X A; 

 h denoting the horizontal intensitj'^, as originally deduced, and 

 Z h its correction. 



A correction of a similar kind (that is, depending on the rate 

 of vibration at the base station) seems to be required also in the 

 series of results obtained in the west and south of Ireland in 

 the summer of 1835. In reducing the observations of this series, 

 I had taken as the Dublin time, the mean of the initial and final 

 times, without regarding the number of separate observations ; 

 but, if we suppose the difference between these times to be owing 

 to errors of observation, or to ?i\\y fluctuating source, it is ma- 

 nifest that we should take, as the Dublin time, the mean of the 

 separate results themselves. This seems to be the proper course 

 in the present instance. The initial time is the result of a sin- 

 gle observation only, and that taken under the disadvantage of 

 an uiuisually high temperature ; so that the difference bet^^een 

 it and the final time (which difference is nearly the same for the 

 two cj'linders) is probably due either to the irregular fluctuations 

 of the horizontal intensity, or to error in the coefficient of the 

 temperature correction. 



It is easy to determine the amount of the required correction. 

 If T denote the time of vibration at any station, 'l^ tliat at the 

 base station, and h the ratio of the horizontal intensities, 



h-Tl 



