48  Bappicker—On the Influence of Magnetism on the Rate of a Chronometer. 
The following remarks concern more the special arrangement of the experi- 
ments :— 
The chronometers should always, whether under the influence of the earth, 
soft iron, or a magnet, be dbserved in at least four different azimuths; if 
possible, their rates should be determined in more positions (especially as to 
the magnet), but then it must be adhered to as a general rule, that the 
detached rate is always to be derived in the same position of the chrono- 
meter relative to the magnetic pole of the earth, as was done with reference to 
the artificial magnet. Thus only would strictly inter-comparable results be arrived 
at. The turning of the chronometer is not to take place round the centre of 
the dial or box, but round the centre (or rather vertical axis) of the balance and 
spring. It is not probable that this will have any sensible effect upon the results 
of the observations with varied azimuths; but the values obtained by experiments 
with magnets will undoubtedly become more regular and systematic, and probably 
admit of a mathematical treatment similar to that proposed by Airy (p. 38). 
Obviously, the proper centre round which the chronometer should be turned, so as 
to produce equal and opposite effects upon the balance and spring in opposite 
directions, need not necessarily coincide with the centre of the balance (on account 
of the other steel parts of the instrument), but it will not be very far from it. A 
determination of this proper centre would, however, be very interesting: some 
method by which perhaps it may be done will be pointed out afterwards. The 
primitive or starting direction, from which the azimuths are to be counted, should 
not be the XII-VI line of the dial, but the direction which the cross-bar of the 
balance (being the line of symmetry of the balance) assumes when the balance is 
at rest. This line will probably coincide with the XII—VI line in most instru- 
ments (cf. p. 32), but not necessarily, and should always be ascertained previously 
to the experiments. The turning should be done in the succession N., E., 8., W. 
(cf. Airy, p. 388). The time during which the chronometer remains in each position 
should not be less than one day, nor exceed two (cf. p. 36): the most advisable 
time of exposure is probably one day, and the more so, as in magnetic experi- 
ments too long an exposure to the magnet in one direction should be avoided. 
The above remarks contain everything necessary for the observation of rates in 
different magnetic azimuths; it is only to be added that, in case there should be a 
sensible difference between these rates, Airy’s method of compensating the terres- 
trial influence — strictly arranged as described by him — should be applied 
(cf. p. 46). 
Experiments with soft iron, which should succeed the ‘‘ azimuth ” observations, 
are highly desirable on account of their immediate practical applicability. It would 
further be important to decide whether an eventual effect of terrestrial magnetism 
is altered sensibly by the addition of soft iron. Their arrangement is, perhaps, 
best made after Barlow’s plan (p. 17, seg., p.21): a ball (or bar or plate) of soft 
