50  Bappickrr—On the Influence of Magnetism on the Rate of a Chronometer. 
easily to the steel parts of the chronometer, still some caution ought to be used. 
Here lies a great advantage of the “ azimuth” and ‘ soft iron” experiments over 
the magnetic ones, since the former cannot possibly do any hurt to the chrono- 
meter, and may, therefore, be carried on safely with the best and most sensitive 
instruments. It is, however, not probable that a distance of the magnet cor- 
responding to a compass deviation of 45° or even 90° will have any other than a 
transient effect upon the chronometer. It would be particularly interesting to in- 
vestigate how far and in what proportion variations of the magnet’s distance are 
accompanied by variations of rate. In order to ascertain this, the distance should 
be varied so as to alter the deviation of the needle successively by 5 degrees. 
This should at first be done with the magnet in position a; but it would obvi- 
ously be also of great interest in the following arrangements (é to d@). 
J. Delamarche and Ploix’s experiments with different magnets acting simul- 
taneously (cf. pp. 42, 43) should be repeated. It should especially be ascertained 
whether the position of magnets, compensating each other’s effect upon a compass, 
differs much from that of the same magnets when mutually neutralizing their 
influence upon a chronometer. This experiment would probably lead to a know- 
ledge of the point of equilibrium, mentioned before (p. 48), round which, strictly 
speaking, the chronometer ought to be turned. 
2. All the above experiments under 1 are, if possible, to be made with both poles 
of the magnet alternately, since we have seen that a difference between their 
actions frequently takes place, and is always to be presumed when the balance or 
the spiral have acquired fixed magnetism. 
3. A question of special importance is, whether the effects of a magnet 
upon a chronometer, when applied in exactly the same way at diferent tempera- 
tures, would differ sensibly from each other. Generally speaking, some difference 
seems to be probable, since the actions of temperature and magnet upon balance 
and spring may lie either in the same or opposite directions. 
4. Finally, we may repeat our remark, that all the above experiments would gain 
considerably in importance if they could be made upon chronometers, the steel parts 
(especially balance and spiral) of which have been separately tested as to fixed 
magnetic qualities, and are known to be either unmagnetic or to have permanent 
magnetism in a special distribution, which has been previously ascertained. In 
such a case the experiments should be arranged with reference to the magnetic 
axis of the balance instead of to its line of symmetry. 
I do not dwell upon the variation that could be brought into these experiments 
by replacing the steel balance and spring by others of different metals, since this 
could not be done without considerable expense, nor without the assistance of an 
experienced chronometer-maker. 
After the preceding suggestions, I should propose the following arrangement 
as an instance of a series of experiments upon a chronometer, which may form the 
