BappickKER—On the Influence of Magnetism on the Rate of a Chronometer. 21 
The above experiments are, certainly, a great deal more scientific and 
systematic than those by Fisher; in some respects, however, they are decidedly 
defective, and lose much of their value as a base of practical conclusions. The 
greatest difficulty in all chronometrical experiments is well known to be that of 
separating the great number of disturbing influences according to their sources. 
This has not been sufficiently attended to. Thus the temperature is read once a 
day only, and no statement is given as to the maximum and minimum values, so that 
the mean temperature remains unknown. It is, moreover, not mentioned how far the 
compensation of the chronometers for changes of temperature was perfect; conse- 
quently the mean detached rates, on which the whole result depends, are rather 
uncertain and doubtful. Moreover, although it is very interesting and important 
to observe the chronometers in different azimuths, as Barlow did, still his obser- 
vations are of little value, since he did not vary the positions systematically, and 
especially since he did not observe the detached rates in different azimuths also. 
Scoresby (No. 8) thought apparently the same, when he says that in all the 
magnetical experiments made by others (which can only be those of Fisher and 
Barlow) some circumstances were neglected, ‘particularly the position of the 
watch or chronometer when its rate was determined.” And finally, almost the 
greatest objection to Barlow’s experiments is their want of uniformity: no two 
chronometers have been treated in the same way, so that the results obtained with 
the different time-pieces are only imperfectly, if at all, comparable with each other. 
These objections affect Barlow’s general deductions, in so far as they take a 
good deal of their evidence away. The fact, however, that iron affects a 
chronometer remains, and Barlow’s experiments show how desirable and important 
further experiments with soft iron would be. These would have to be made on a 
larger scalé, and with the most careful attention to every possible source of 
disturbance which might tend to obscure the result. Some suggestions as to the 
arrangement of such experiments will be made later on. 
VI. 
[W. Scoressy. | 
W. Scoresby’s (No. 8) general ideas about the effect of the ship’s iron on a 
chronometer are not unlike those of Barlow and Lecount, but he thinks that 
the éerrestrial magnetism has by far the greatest influence upon a chronometer’s 
rate, or, as he expresses it later on (with reference to Fisher’s observations in 
Spitzbergen), ‘‘the force of terrestrial magnetism acting upon a balance that is 
magnetic is fully sufficient to account for every change of rate observed.” For 
in the same proportion as the magnetism of the earth (or the directive force of the 
compass-needle) exceeds the magnetism of the ship (or the deviating force), the 
