T. C. Mendenhall—Acceleration of Gravity at Tokio, Japan. 131 
placed ten times to cover the whole length, it is plain that an 
great degree of accuracy must have been difficult to obtain, 
and this is especially true when the measurement of that por- 
tion of the wire which hangs over the wheel is considered. 
Their 26th experiment was made on the 25th of January, and 
the 58d on the 21st of February, from which we may infer that 
the entire time of suspension was at least two months. As 
only one measurement is spoken of, it is probable that it was 
measured at the conclusion of the series of experiments, an 
it seems hardly likely that its length would have remained con- 
stant during that length of time. In getting the time of vibra- 
tion the first method used was what might be termed the 
method of coincidences by electricity, and which, so far as I 
know, was first described by Professor Pickering, in his excel- 
lent “ Physical Manipulations.” This was afterward rejected, 
however, and the vibrations were counted by means of a 
Morse instrument. The authors speak of measuring the frac- 
tion of a vibration, but evidently this could not be done with 
accuracy by the use of such an arrangement, and there is also 
the objection that the pendulum was obliged to do the work of 
breaking the circuit at every vibration. Messrs. Ayrton and 
Perry give the time of vibration of their pendulum for only 
three experiments, and it is a little difficult to understand 
exactly how these were obtained. The time, taken from the 
chronométer, is given and also the number of vibrations. Any 
one who will take the trouble to divide one by the other, will 
obtain results differing very materially from those given in the 
paper. The only way out of this difficulty that occurs to me, 
is to consider the times given as the apparent times corrected 
for clock error, and this I shall do, alth 
corrected for buoyancy, but the same wide range would exist 
