526 
APPENDIX. 
[ Errors in variation. 
Besides the errors which the attraction of the iron produced in the 
compasses at the binnacle of the Investigator, differences are frequently 
mentioned in the course of this voyage as having been found in the 
magnetic needle on shore, and on board the ship in the vicinity of land. 
That there are few masses of stone totally devoid of iron, and that all 
iron which has long remained in the same position will acquire magnetism, 
or a power of attracting one end of the magnetic needle towards one part 
of it, and the opposite end towards another, is, I believe, generally ad- 
mitted. The kinds of stone which I have observed to exert the greatest 
influence on the needle, are iron ore, porphyry, granite, and basaltes ; and 
the least, are sand or free stone, and calcareous rock, and the argillaceous 
earths very little. 
The iron in the ship attracted the south end of the needle in the 
southern hemisphere, and in the same part of the world it was the same 
end of the needle upon which the land had an attraction. The following 
are some instances : 
In King George’s Sound, the west variation was 6° greater on the 
western head of Michaelmas Island, than it was on the east side of a flat 
rock in the sound. The stone here is granite. 
On approaching the granite islands of the Archipelago of the 
Recherche, from the west , the corrected variation on board the ship was 
increased from 5° 25' to 22' west, contrary to the regular order ; but 
when Termination Island bore nearly West, and the principal cluster 
N. N.W., the corrected variation was no more than 0° 51' ; and after clear- 
ing the Archipelago some distance, it again increased to 4~° west.* 
* M. Beautemps-Beaupre (in Vol. I. p. 605, of the work before cited) gives the follow- 
ing instance of attraction in the stone of this Archipelago. The compass was placed 
upon one of the capes of the main land, to set the bearing of a point. “ When the bear- 
“ ing had been taken, the compass was removed six feet from its place, beyond a large 
“ stone ; where the vane being by chance directed to the same point, a difference of four 
“ degrees was found in the bearing, although the object were so far distant that the 
fC change of place should scarcely have produced a difference of one minute. Fully per- 
“ suaded that we had made an error, either in reading off’ the bearing or in writing it 
“ down, the first observation was verified ; but we had the same result within a few 
“ minutes as had been marked on the paper, and it was certain that the stone near which 
“ the observation had been made, had solely caused this great error.” 
j 
