528 APPENDIX. [Errors in variation, 
needle, and the same with the upper parts of magnetic lands in the 
southern hemisphere. But it is an universal law in magnetics, that powers 
of the same quality repulse, and dissimilar powers attract each other ; 
therefore the upper parts, both of the iron in the ship and of the land, 
should, like the north end of the dipping needle, repulse the north and 
attract the south end of the compass needle. Now the compass in, or 
upon the binnacle of a ship is raised above the greater part of the iron, 
and therefore more in a situation to be attracted by the upper, than the 
lower parts of the different pieces. The same will generally be the case 
with respect to the land; its southern polarity must often be lower than 
the depths of the sea, whilst the upper part, which attracts the south end 
of the compass needle, will be nearly on a level with, sometimes a few 
degrees above the ship. 
This reasoning from abstract principles is consistent with my obser- 
vations « Z near the coast* of Terra Austral,* ; aud if it he just, the 
contrary effects must take place in the northern hemisphere, at least in 
high latitudes : the upper parts, both of the iron in a ship and of land ^pos- 
sessing magnetism, will attract the north end of the compass needle. That 
it is the north point of a compass on the binnacle which is attracted by the 
iron of a ship in the northern hemisphere, has already been shown ; but 
whether the land do generally attract the same point, I have no knowledge 
from experience : answers to the following queries would probably be 
useful in the determination. 
Is the west variation on the coast of Holland and Germany consider- 
ably less than on the east coast of England and Scotland , in the sanu. 
latitude? . 
Is it sensibly less at Holy Head than at Dublin ; at 1 ort 1 atrick 
than at Caricfergus ? 
Is the variation as much, or greater on the Yorkshire, than on the 
Lancashire coast ? 
And generally in the northern hemisphere, is the west variation 
greater, or east variation less on the east sides of islands and projecting 
points than on the west sides ? . 
Observations made on ship-board for determining t iis or any o er 
general question of magnetism, will require, when the head is not at North 
