508 MAGNETIC INTENSITY. 
The object of noting the time of every tenth vibration is to 
check errors in the counting, which will sometimes occur in the 
course of the 360, particularly with the very short and quick- 
moving needles of M. Hansteen’s very portable apparatus, and at 
stations of low dip, where the horizontal force is greatest, and the 
needle consequently moves most quickly. Several such mistakes 
evidently occurred. When the time of completing every supposed 
tenth vibration is observed with tolerable exactness, and the dura- 
tion of each pair of vibrations decidedly exceeds any irregularity of 
probable occurrence, apart from miscounting the number of vibra- 
tions, such mistakes can be discovered with ease, and rectified 
with certainty. This has been done in every case where no doubt 
could exist of a mistake of the kind having occurred ; such as 
when all the intervals are of nearly equal duration, with one or 
two exceptions, which differ as much as three or four seconds from 
the general body. There are two stations, however, Callao and 
Keeling Islands, where the rectification is not so clear, or the true 
result so obvious. At Callao there are three series of horizontal in- 
tensities, each of forty observed intervals, which should be of ten 
vibrations each. Several of these intervals are between 17,5 and 
18,5 seconds, and several others between 20,5 and 21,5 seconds. 
These can hardly represent an equal number of vibrations, because 
the difference between them is greater than can easily be supposed 
due to any uncertainty in seizing the particular beat of the chro- 
nometer at which the vibration was completed ; it is, moreover, 
about the time that would be occupied by two vibrations more or 
less. The question then arises, do the longer intervals represent 
12, and the shorter 10 vibrations, or do the longer represent 10, 
and the shorter 8? In the former supposition the intensity at 
Callao would be about 1.01(Paris=1.348): in the latter about 
0.75. The difference shews how great an error would be risked by 
either assumption. If we take a mean of all the intervals as they 
stand, the amount of error risked would be certainly lessened ; 
but we should assuredly not have the true time of three hundred 
vibrations, except on one supposition: namely, that the irregu- 
larities in question are not errors in estimating the number of 
vibrations, but that each interval really represented an equal 
number, and that some unusual and accidental cause occasioned 
the needle to differ so greatly in successive intervals. But this 
