24ft" QPrpf. E. Taylor Jones on a 



arrangement worked very satisfactorily, the image of the 

 scale in the telescope being quite steady. Balance was 

 effected by varying IS, and this could be adjusted to 2 or 

 3 ohms. The tuning-fork was driven by two storage-cells 

 and had a ^-microfarad condenser connected across its 

 mercury interrupter to diminish the sparking. During these 

 measurements, the fork was in circuit with the primary of an 

 induction-coil whose secondary was connected to a vibrograph 

 consisting of a rotating and travelling cylinder covered with 

 smoked paper and driven by a motor. One terminal of the 

 coil was connected to the axle, the other to an insulated needle- 

 point placed at a few millimetres from the cylinder. At each 

 break of the circuit a spark passes from the needle-point 

 leaving a small white mark on the paper. A second induction- 

 coil was connected to the laboratory clock through a drop 

 of mercury placed underneath the pendulum and at about 

 the centre of its swing. A point attached to the pen- 

 dulum made and broke the circuit once every second. A 

 suitable condenser was connected across this interrupter to 

 diminish the spark. The secondary of this induction-coil was 

 connected to a second insulated needle-point, so that another 

 series of marks is made on the cylinder at intervals of one 

 second. After each determination by Maxwell's method the 

 cylinder was sparked while turning at a fairly uniform speed 

 for 40 or 50 seconds. In subsequently determining the 

 frequency from the record the cylinder was turned slowly 

 back (the needle-points being still in position), the marks 

 counted, and a table drawn up giving the readings opposite 

 the first needle-point corresponding to all the clock marks. 

 The differences of the readings for the longer intervals (at 

 least 30 seconds) were used to give a mean value of the 

 frequency of the fork. Only even-numbered intervals were 

 used in case the drop of mercury was not exactly at the 

 centre of the swing of the clock pendulum. 



In one case, for example, taking all the even intervals 

 above 40 seconds, sixteen values of the frequency were 

 obtained, the least of which was 48*835 and the greatest 

 48*847, and nine lying between 48*840 and 48*842. On 

 another occasion the fork was placed on a different table 

 along with the storage-ceils used for driving it. Three papers 

 taken on this occasion with the exciting current, the ampli- 

 tude of vibration, and all other circumstances the same, 

 gave for the frequency of the fork the mean values 48*706, 

 48*707, and 48*707. The frequency depends to some extent 

 on the manner in which the fork is supported and on the 

 current driving it. 



