SURVEY OF THE BRTTISH ISLES FOR THE EPOCH JANUARY 1, 1886. 
279 
yet the disturbance diminishes from 23°‘8 E. a,t Station XXIII., near the top of the 
hill, to 1°’6 E. at Station IL, near its base. 
These conclusions are completely borne out by observations made on the Coven- 
tina,” in which we visited the island in 1884 and 1888. 
On leaving Ganna for Loch Boisdale in 1884 we sailed as close as possible to the 
northern face of the island, and took frequently the compass bearings of points on 
Skye. We were unable to detect the smallest deviation of the needle. 
In 1888 observations were made under still more favourable circumstances. We 
approached the island from the north, and, w^hen about three miles distant, the yacht 
was directed towards a mark on Bum, by which its coiu'se could be kept without 
reference to the compass. We were then sailing magnetic S. E., in the most favour¬ 
able direction for the effect of Compass Hill (if any) to be detected. We passed it 
within 200 yards of the shore, but observed no deviation of the compass; and we are 
quite certain that, if there was any, it was less than one-eighth of a point, i.e., less 
than 1°'5. 
The net result of our observations is that the basaltic cliffs of Ganna are powerfully 
magnetic, and may deviate the needle of a comjDass placed near them by about two 
points, i.e., about 23°, but that the effect diminishes very rapidly with the distance, and 
is inappreciable on a ship’s compass 200 yards from the base of the hill, to which 
tradition ascribes, and in which we have ourselves detected, the most powerful 
magnetic properties. 
We have adopted 2 2°'8 as the mean value of the Declination at the less disturbed 
stations in Ganna in August, 1888 ; this leads to 23° 13' for January 1, 1886, which is 
only 6' in excess of the calculated value. This is interesting, inasmuch as Plate IX. 
shows that neighbouring stations have Declination disturbances of opposite signs, 
and indicates that a line of no regional disturbance runs near to Ganna. 
Thus the four stations, Kyle Akin (No. 35), Soa (No. 46), Ganna (No. 14), and 
Loch Boisdale (No. 10)_lie very nearly in a straight line, and the disturbances of the 
Declinations vary continuously, being 28''0, 18''4, 6''0, and — 45'’9, which proves 
that Ganna lies near an attractive centre or ridge. This is in harmony with the fact 
that the disturbance of the Vertical Force is positive, and is enormously great, 
amounting to 0‘1839, or about 0'04 of the whole Vertical Force. 
A comparison of the results obtained at Malvern and Ganna points very clearly to 
the otherwise probable conclusion that far-reaching effects are to be expected quite as 
much from the great mass and uniform magnetization of rocks as from their being 
highly magnetized. The basalt of Ganna is far more susceptible and more magnetic 
than the Malvern syenite, but we doubt if a mile and a quarter from Gompass Hill it 
would produce an effect on the Declination needle at all equal to that which we liave 
shown is due at that distance to the Malverns. 
