72 
MR. A. W. RtiCKER AND DR. T. E. THORPE ON A MAGNETIC 
The following table gives the mean values of P obtained for each year during the 
survey. 
Magnetometer. 
60 
! 
61 i 
1884 . 
0-000817 
1885 . 
0-000866 
1886 . 
0-000828 
0-000753 
1887 . 
0-000809 
0-000692 
1888 . 
0-000800 
0-000706 
The ordinary method of determining the Inclination with two needles gives two 
independent values, the agreement of v'hich furnishes a test of the accuracy of the 
observations. It is, however, usual to give only one value of the Horizontal Force 
deduced from the means of the values of w/H and inH calculated from the deflection 
and vibration experiments. The constancy or regular change of the value of m affords 
a means of determining whether the observations are sufficiently good, but it seems to 
us that nothing is for this purpose so satisfactory as the agreement or disagreement 
between two independent experiments. The results of the ordinary deflection and 
vibration observations can readily be presented in such a form as to satisfy this 
condition with only an insignificant addition to the labour of reduction. 
The vibration experiment furnishes twelve independent determinations of the time 
of 100 vibrations, six of which are taken when the scale of the vibrating magnet is 
apparently moving to the right, and six when it appears to be moving to the left. If 
we take the mean of the six observations furnished by the first and last three of each 
of these groups, we have two independent determinations of the vibration period, 
based in each case on six observations, of which three were taken when the movement 
appeared to be to the right, and three when it appeared to the left. 
If we agree always to combine the first and second of these groups with the values 
of 77 i/H obtained when the distance between the magnets is 0'3 and 0'4 metre 
respectively, we obtain from each experiment two values of m and H. Except in so 
far as errors may arise in the determination of the temperature, or in the adjustment 
of the instrument (levelling, &c.), these are absolutely independent, and thus furnish 
a satisfactory check on the accuracy of the observations. It would, however, be hardly 
worth while for the sake of this advantage to go through the labour of repeating the 
reduction of the vibrations. As the two vibration periods are very nearly the same, this 
is unnecessary. Let T be the time deduced from all the observations, and T e/T the 
times given by the first and second groups of observations selected as above described. 
Then if D be the value of mj'H. given by the deflection experiment, 
