6 
MR. H. M‘LEOD AND LIEUTENANT G. S. CLARKE ON THE 
To count the number of lines passing across the horizontal hair of the eyepiece 
during an experiment it is necessary to know the number of complete revolutions, 
and also the additional fraction of a revolution which the drum has performed during 
the observation. The number of the revolutions is obtained by an electric counter 
attached to the axis of the drum, a pin on which, shown at a on the diagram, 
Plate 3, fig. 9, completes the circuit by pressing two springs together. To measure 
the fraction of the revolution, an electro-magnet, b, is made to produce a mark on 
the drum both at the commencement and conclusion of the experiment. Above the 
chronograph drum is a saddle supporting the electro-magnet; to the armature of the 
latter is attached a finely pointed tube, t, containing a solution of magenta mixed 
with an equal volume of glycerin to prevent it drying and so stopping up the jet. 
An electric current from the clock is made to pass through the electro-magnet, both 
at the beginning and end of the experiment, thus making two dots on the white 
paper covering the drum, the distance between the dots being the fraction of the 
revolution required. To find the number of lines in this fraction it is only necessary 
to place a straight-edge along the drum and to note the numbers of the lines 
corresponding to the two marks. The saddle carrying the electro-magnet is moved 
forwards or backwards parallel to the axis of the drum by a screw, which can be 
turned by means of a cord and pulley near the operator’s hand, so that two con¬ 
secutive dots may not be made on the same line on the drum, and the paper round the 
drum thus serves as a permanent record of the observations made. The electric 
current which makes the marks also starts and stops the counter. The circuit of the 
counter is broken at two mercury cups, c, which are connected by a copper wire 
attached to a small magnet resting between the poles of an electro-magnet, cl. By 
means of the reversing switch, e, which is turned in one direction before the commence¬ 
ment of the experiment, the copper wire is dipped into the mercury cups, so that when 
the current is sent by the clock relay the counter is automatically started, and before 
the end of the experiment the reverser is turned over, so that the contact is broken 
and the counter stopped also automatically when the second mark is made on the 
drum. 
In each determination of a fork, five experiments are usually made, lasting one, two, 
three, four, and five minutes respectively. The object of this is to diminish any error 
which may be due to variation in the duration of movement of the electro-magnets, and 
also to reduce to a minimum the errors of observation. The forks are placed in the 
vice many hours before the determinations are made so that they may attain a uniform 
temperature throughout, and that this temperature, as shown by the thermometer, 
may be as correct as possible. The glass case surrounding the fork prevents air 
currents from affecting the thermometer, and a large glass plate placed vertical on the 
table diminishes the rise of temperature due to radiation from the observer’s body. 
When a determination is to be made the counter is set at 0, and the velocity of 
rotation of the drum regulated by means of driving weights and by the fan until the 
