SURVEY OF THE BRITISH ISLES FOR THE EPOCH JANUARY I, 1886. 161 
whereas in the above table they are reduced bj the mean coefficient of secular change 
at Kew (B'O) during the period of the survey, together with a mean correction for 
monthly variation. 
In the next place in the reduction on p. 58, no distinction was made between the 
diurnal variation and disturbance, the total divergence from the mean being read 
off directly. The corrections applied in the above table were obtained by reading 
off the disturbances only and afterwards adding the value of the diurnal variation. 
The differences are a little larger than we should have expected, but this m 
probably due to the fact that, as the corrections were wanted quickly, the constants 
of the curves were only determined provisionally. 
That the accuracy attained was practically sufficient is proved by a comparison of 
the numbers in the above table with those on p. 58. 
The mean difference between our results and those given by curves treated as 
accurately as possible is dh0''8], while the mean difference between the numbers 
given above and their mean is ff; 0'‘83. 
The Kew value for January, 1886, is 18° 16''3, and ours, from the above table 
when reduced to the Kew instrument by adding, the mean difference, is 18° 14''4 + 2'‘5 
= 18° 16'-9. 
If from each of our uncorrected observations given on p. 58, we subtract the 
difference between the corresponding Kew value and 18° 16''3, we reduce to 
January, 1886, by one operation, in which no distinction is made between disturbance 
and the diinnal, monthly, and secular variations. The largest of the numbers so 
obtained is 18° 16'T, and the smallest 18° ll'ffi, while the largest and smallest in 
the above table are 18° 16'‘6 and 18° l?/'5. Thus, not only are the mean results 
obtained by the tAvo methods in close accord, but the range of Amriation of the insults 
is in each case practically identical; in other words, they agree to within the limits 
of the error of experiment. 
Horizontal Force. 
- 
Date. 
Observer. 
Instrument. 
H. 
Ho- 1 
July 17, 1884 . 
R. 
60 
1-8075 
1-8100 
April 2, 1886 . 
T. 
61 
1-8082 
1-8078 
,, 2, „ . 
R. 
60 
1-8098 
1-8094 
)> 22, „ . 
R. 
60 
1-8101 
1-8096 
September 30, 1887 . 
R. 
60 
1-8112 
1-8082 
October 11, 1887 .... 
T. 
61 
1-8114 
1-8084 
„ 12, „ .... 
R. 
60 
1-8115 
1-8085 
„ 13, „ .... 
T. 
61 
1-8114 
] -8084 
» 18, „ .... 
T. 
61 
1-8123 
1-8092 
M 18, ,, .... 
T. 
61 
1-8125 
1-8094 
,, 19, ,, .... 
R. 
60 
1-8125 
1-8094 
Mean . 
1-8089 
MDCCCXC.-A. Y 
