chronometers by the proximity of masses of iron. 365 
my experiments ; indeed we are led strongly to suspect, that 
the remarkable change in the rates of the nine chronometers 
of the Dorothea and Trent, reported by Mr. Fisher, must 
have been produced by some extraordinary cause, not com- 
monly operating on ship board. 
I have already observed, that, according to the idea I enter- 
tain of the action of iron on the balance of a chronometer, it 
is actually necessary to conceive, that part of the machine, 
or at least its spring, to have acquired a certain polar or 
directive quality, whereby, independent of any other power, 
the balance would have a tendency to assume a certain direc- 
tion, when brought within the sphere of action of a given 
mass of iron ; and the amount of that tendency might, I 
conceived, be estimated, by counting the number of vibrations 
which a small magnetized needle would make in a given 
time, in any assigned situation, near the iron, and comparing 
the result with the number it would make under like circum- 
stances, and in the same time, when wholly removed from 
any attracting mass. 
In order to illustrate this view of the subject a little more 
particularly, let A B C D (fig. 1 . PI. XXV.) represent the ba- 
lance of a chronometer, s, s' its spring, and let D be that part 
of the rim which is attracted by the centre 0, of an iron ball 
or shell. If now we conceive the spring to be detached from 
the fixed part of the machine, it will be free, with the balance 
itself, to take any position. The point D will therefore be 
attracted towards 0 ; and if it be displaced from this position, 
it will have a tendency to oscillate on each side of the point D ; 
and the number of vibrations which it would make in a given 
