SURVEY OF THE BRITISH ISLES FOR THE EPOCH JANUARY 1, 1891. 645 
200 1 
when V = 0’47, H = 0-15, sin 8 = 0‘4 about, x = \ mile, = -- mile, 
i / OU oo 
^3=5 + ^ miles. 
Hence the force becomes 2'8/c. 
To show how enormously these numbers may exceed the true forces, we have only 
to remember that if we had supposed the visible and submerged surfaces of the island 
to be planes at the same height, say 300 feet, above and below the compass respec¬ 
tively, they would have neutralized each other’s effects, and the Force would have 
been only O'IIk. 
This latter supposition is much more likely to be in accord with fact. The island 
of Canna is merely a hump on a lava stream. Sir Archibald Geikie informs us that 
the thickest visible mass of bedded lavas among the Western Isles occurs at Ben More 
in Mull, where these rocks somewhat exceed 3000 feet in total depth. The three 
soundings given in the Admiralty Chart nearest to the north, east, and south 
sides of Canna are 40, 75, and 40 fathoms respectively. The mean of these is nearly 
300 feet, so that the hypothesis that the island may be regarded as a block, the total 
height of which above the main lava sheet is 600 feet (300 below and 300 above 
water), is justified. 
It is true that on the western side, remote from that on which our observation was 
made, there is a patch of shoal water several miles in area, the sounding nearest to 
the island being only 14 fathoms. If we assume this shoal water to mark a submerged 
extension of the more elevated basalt, it would be represented by a sheet 300 feet 
thick and a few square miles in area laid on the top of the main mass. It will be 
shown hereafter that, even when the observations are made on land with the most 
delicate apparatus, the Disturbances due to such thin sheets become insensible at 
distances from their edges equal to a few multiples of their thickness. The points at 
which we began and ended our observations were about 8 and 5 miles from the shoal 
respectively, or about 140 and 88 times its assumed thickness. The difference 
between the effects produced could not have been detected by a magnetometer, and 
would be absolutely inappreciable on a compass. 
Hence the only consideration which we have omitted, which might have tended to 
increase the force, is that we have supposed the eastern face of the island to be 
vertical instead of sloping towards the sea. The Force dae to this face is, however, 
but a small fraction of the whole, and if we doubled or trebled it, no serious difference 
would be produced in our results. Considering, therefore, that all our assumptions 
have tended to make the Disturbance impossibly large, we may safely say that the 
Horizontal Force cannot be greater than 2•8 k, and may be 25 times less than this 
quantity. 
Tavo fragments of basalt from Canna gave k = 0'0015 and 0'0013 respectively. 
The mean of 13 specimens from Mull was O'OOIG, 
