DETERMINATION OF THE RATE OF VIBRATION OF TUNING-FORKS. 
9 
When this determination was made the rate of the clock was taken as *7 88 second 
86399‘212 
per day, losing; correcting for this rate, we have 256‘294x—gffdofi— = 256’292 vibra¬ 
tions per second at 15°*58 C. 
The maximum difference between the above numbers is '006 or ‘00234 per cent. 
An error of reading of one line, or mistaking one wave for the next, will produce, 
in a five-minute observation, a difference of ‘0033 in the determination of a 256 fork, 
while the same error in an observation lasting one minute will produce a difference of 
‘0166 ; the value of an experiment lasting several minutes is therefore apparent. The 
method is of such a nature that no bias in the mind of the operator can have the 
slightest effect on the result, as the final numbers show no indication of their direction 
until the positions of the dots are read off on the drum ; and, to prove that special 
training is not required for the observations, it may be mentioned that a sharp boy of 
16 succeeded in keeping the figure sufficiently steady after two trials with the 
machine. 
Three elements of possible error in the method above described suggested themselves 
—1st, the effect of the frequent bowing ; 2nd, the variation of amplitude ; and 3rd, 
the holding of the fork in a vice instead of fixing it, as usually employed, on a sounding 
box. These matters will now be dealt with seriatim. 
1 . Effect of bowing. 
In order to dispense with the use of the bow, and to keep the amplitude constant, 
we had contemplated the employment of Helmholtz’s method of maintaining the 
forks in vibration by means of an electro-magnet excited by an intermittent current 
from a contact maker worked by a large tuning-fork, the vibration number of which 
would bear a simple relation to the vibration number of the fork to be measured. The 
experiment, however, was not tried, as we feared that the current might to some extent 
control the fork, which would, of course, be inadmissible ; the bow was therefore 
retained. 
If the bowing altered the phase of the vibration, which seemed at first not impro¬ 
bable, there would be a sudden start of the wave figure, in one direction or the other, 
at each stroke of the bow ; this has never been observed. Again, if the bowing 
altered the phase, even to an extent inappreciable by the motion of the wave, its 
frequent occurrence at each application of the bow would necessarily alter the number 
obtained in the determination. In order to put this to the test six experiments were 
made, in three of which the fork was bowed every 20 seconds, two strokes of the bow 
being generally made on the prong, and in the three others the fork was bowed every 
five seconds. The determinations were made alternately, and the first four lasted 
five minutes, and the others three minutes. The following numbers were obtained :— 
MDCCCLXXX. 
c 
