SURVEY OF THE BRITISH ISLES FOR THE EPOCH JANUARY I, 1886. 267 
If we take the values of H and 6 at AVick and St. Leonards, and 'write for dd the 
circular measure of 1' multiplied by x, the unknown number of minutes by which the 
calculated Dip has been taken too large, we get at 
Wick. dV — 0-0048a:, 
St. Leonards. dY = 0'0036x. 
Hence in the north the error will be greater than in the south, that is, the excess 
of the calculated over the true undisturbed Vertical Force is greatest in the north. 
Consequently the difference betw’een the observed and calculated Vertical Force will 
be more frequently negative in the north, the excess of the error amounting to 
0'0012 metric unit for every minute by which the calculated Dip is too large. 
If, then, apart from local attractions the country, as a whole, attracts the needle so 
that the Dip is everywhere 10' greater than it would be if the British Isles were 
replaced by sea, the error in the calculated Vertical Forces deduced from calculated 
Dips, obtained on the assumption that no such defect exists, would be 0‘0120 greater 
in the north of Scotland than in the Channel, so that neo;ative disturbances would 
largely predominate in the north. 
As from the geological character of Scotland it is probable that the error in the 
calculated Dip would not be constant, but would be greater than in England, the 
validity of this explanation is even more probable than this calculation indicates. 
It is not, therefore, safe to draw conclusions based on the relation between the 
Vertical disturbing Forces at distant stations. 
In the calculation of the disturbing forces we are not aiming' at, and could not attain 
to, results which would give more than a general idea of their direction and magnitude. 
It may be possible to conclude with certainty that an attractive mass exists near 
certain stations, even if the directions of the forces are wrong by 10° or 15° and their 
magnitudes are inaccurate by 50 per cent. 
To show that a degree of accuracy of this kind is attained we collect here the results 
at stations where we have taken a full set of observations at two or more places, or on 
different occasions. There are of course localities where it would be hopeless to get a 
good result from a single observation. Such places are Ganna and Portree. At these 
we should never have observed except for the sake of investigating the disturbances 
which we knew were very great, and we therefore do not now take them into account. 
It is sufficient if the agreement bet’ween different observations is satisfactory at stations 
which we regarded at the time of observation as normal. 
At Stornoway we made four sets of measurements, the first in 1884 in the Castle 
Grounds, the other three in 1885 and 1888 on Ard Point, about a mile distant from 
the former station. The ground 'was not good, as when we oljserved simultaneously 
about 50 yards apart on Ard Point the declinations differed by 18'. 
At Loch Aylort we observed in 1884 and 1888 as nearly as possible on the same 
station. At Oban we observed on the mainland in 1888, and on the island of Kerrera 
2 M 2 
