SURVEY OF THR BRITISH ISLES FOR THE EPOCH JANUARY 1, 1891. 511 
35''2 and 32''5 for each of them respectively. They thus increase as we pass north¬ 
wards until Scotland is reached, when a decrease follows. That the value 32''5 is not 
incorrect is proved by the fact that it is nea-rly the mean of the two numbers obtained 
(viz., 30'‘7 and 33'‘9) when the Scotch stations were divided into two” groups [cf. 
p. 504). On the other hand no such alternate increase and decrease was shown 
in the 1886 survey, and it is improbable that it really exists. On the whole, 
therefore, having brought the district and calculated curves into close agreement else¬ 
where, we thought it more probable that the large disturbances known to exist in 
Scotland affected the district lines, than that the formula suddenly ceased to be valid. 
We have two d posteriori proofs of the correctness of this judgment. 
(1) When reduced to the same epoch the calculated isogonals of the two surveys 
are in very good accord in Scotland. 
(2) The Horizontal Disturbing Forces, the calculation of which depends upon the 
Declination, point in Scotland, as elsewhere, to the regions of greate.st Vertical Force 
Disturbance the determination of which is independent of the Declination. 
It is impossible that the latter condition should be fulfilled if the calculated 
Declinations were affected with errors amounting to half a degree. 
Horizontal Force. 
The observations on Horizontal Force w^ere treated in exactly the same way as the 
Declinations. 
Tlie following table gives the positions of each of the central stations in the nine 
districts, together with the values of H at each, deduced from the weighted mean 
of all the stations in the district, and also the rates of change per degree of latitude 
and longitude in terms of metric units. 
Table IV. 
District. 
Central station. 
H'o. 
_ fZH' 
* dl ■ 
1 
1 
rfH' 1 
^ d\ ' ! 
Lat. N. 
Long. W. 
I. 
5°6 .38-2 
O i 
4 21-5 
1-6011 
- 0-03817 
- 0-00677 1 
II. 
.55 23-9 
3 23-7 
1-6519 
- 0-03290 
- 0-00587 
III. 
53 24-2 
2 5-9 
1-7313 
- 0-04042 
- 0-00593 - 
IV. 
51 47-4 
0 46-0 
1-8070 
- 0-04348 
- 0-00635 
V. 
54 2-7 
7 37-9 
1-6761 
- 0-04029 
- 0-00802 : 
VI. 
53 30-5 
5 35-2 
1-7083 
- 0-03801 
- 0-00535 
VII. 
51 5-3 
3 8-0 
1-8206 
- 0-04266 
- 0-00757 
VIII. 
52 51-5 
8 11-6 
1-7197 
- 0-03947 
- 0-00578 
IX. 
51 41-9 
4 32-6 
1-7863 
- 0-04137 
- 0-00628 
From these data the district lines were deduced by calculating the value of H at 
specified points in each district and taking the mean at points where the districts 
overlapped. 
Details are given in the following Table :— 
