on Magnetic Attraction * 295 
viations, when its face was north and south, taken for different 
heights of the ball. A horizontal compass was likewise placed 
at the distance of 18 inches from the centre of the table, and ha- 
ving its centre at the same height as the centre of the dipping 
needle. The following are the results obtained : 
Height of centre 
of ball above 
centres of needles. 
Mean deviation 
of 
dipping needle. 
Horizontal Needle. 
Mean observed 
deviation. 
Deviation computed 
from deviation of 
dipping needle. 
10 inches. 
2° 05 
6° 10' 
6° 13' 
5 inches. 
1 36 
5 00 
4 47 
0 
0 05 1 
0 10 
0 17 
This agreement in the observed and computed horizontal de- 
viations was the more striking, from the smallness of the arcs v <r. 
The several deviations of the dipping needle, and of the horizon- 
tal needle, in all positions, being then the necessary consequences 
of such an hypothesis, I concluded, that when a mass of iron is 
removed beyond a few inches from the ends of a magnetic needle, 
so that they are beyond the influence of any accidental magne- 
tism in the iron, the deviation of the needle arises from the ac- 
tion of the iron on magnetic particles in a line, passing through 
the centre of the needle in the direction of the dip. 
In the conclusion of the paper, I pointed out the application 
of this theory to the determination of the deviations of compasses 
on board of ships. I also hinted, that it might be applied to the 
changes which have taken place in the variation and dip of the 
needle, mentioning that the computations founded on it, which 
I had made, agreed to within less than half a degree with those 
observed in London, during a period of more than 200 years. 
I have since made similar computations for the changes in the 
variation at Paris, which give results equally near to the obser- 
vations during the same period. 
Having stated the hypothesis which I adopted, and the gene- 
ral results by which it is supported, I propose to apply the same 
principles to phenomena, which have been observed in different 
latitudes ; and, for this purpose, I have selected those published 
by Mr Lecount, made with great care and perseverance at St 
