SURVEY OF THE BRITISH ISLES FOR THE EPOCH JANUARY 1, 1891. 545 
We think then that the error in determining the datum-plane from wliich Vertical 
ITorce Disturbances are measured, is sufficiently great to make it impossible to draw 
conclusions from the results of the two smweys, without reducing them to the same 
set of terrestrial lines. Otherwise there would be a risk that the stations of the one 
survey would appear, relatively to those of the other, as places of high Vertical Force, 
or vice versd. 
If the Vertical Force Disturbances at two stations were found in the earlier survey 
to be too and 0 respectively, and if in the second survey the corresponding quantities 
at two closely neighbouring stations were 0 and — 100, the two results would be in 
accord as to the relative change of Vertical Force Disturbance on passing from the 
one neighbourhood to the other, but, inasmuch as the absolute values are different, 
they would appear discordant if combined without further correction. 
That agreement between the relative magnetic levels of neighbouring districts has 
been obtained is conclusively proved by the identity of the ridge lines and peaks 
shown in Map 4 with those depicted in our “ 1890 Memoir” ; it only remains, therefore, 
to decide how a common datum-plane shall be fixed. It might seem at first sight 
that greater weight ought to be given to the 1891 terrestrial lines since they have 
been found from a much larger number of stations, but the difficulty of determining the 
true lines is not so much due to errors of observation or to local disturbance at 
particular stations, as to the fact that the Disturbing Forces over large districts affect 
the elements in the same way. In other words, the divergence between the district 
and terrestrial lines is caused not by errors but by regional disturbances. 
This is proved by a comparison of the district and terrestrial lines of the two 
surveys, which display the same peculiarities in spite of the fact that lines corre¬ 
sponding to given values of the elements have shifted from one position to another 
between the two epochs. 
To prove this we may draw particular attention to the Florizontal Force isomag- 
iietics corresponding to D70 and l‘G5 metric unit, to the isoclinals corresponding 
to Dips of 69°, 70°, and 71°, and to the isogonals in the north of Scotland and in the 
south-east of England. 
It follows from this that, if only the number of stations is sufficiently large to make 
the disturbance at any one station unimportant, no great advantage is obtained by 
multiplying the number of stations from which the district curves are deduced. 
We, therefore, determined to give equal weight to the terrestrial lines furnished by 
each of the Surveys, and to take the mean of the numbers in Tables XVII.-XIX. as 
giving the mean terrestrial isomagnetics for January 1, 1891. The results are given 
in Tables XX., XXI., and XXII. 
It will be observed that the elements do not vary quite regularly from point to 
point. This is due to the facts that in some cases different formulse express the 
equations to the isomagnetic curves for the same element in different parts of the 
MDCCCXCVI.—A. 4 a 
