34 
PUTNAM. 
hausted to about 2.4 inches (60 millimeters) with a portable 
air pump. The flash apparatus is for the purpose of observ¬ 
ing coincidences between a chronometer and a swinging 
pendulum. An electromagnet in circuit with a break circuit 
chronometer moves a shutter at the end of each second, thus 
throwing a flash of light through a narrow slit. The image 
of this slit is seen in an observing telescope as reflected from 
two mirrors, one moving with the pendulum and the other 
fixed. As the pendulum is slightly slower than the chro¬ 
nometer, the image reflected from its mirror will be seen in a 
somewhat retarded position at the end of each second, and 
once in five or six minutes the two images will be in the 
same horizontal line, which is the moment of coincidence. 
Between two such phenomena it is evident that the pendu¬ 
lum makes one less oscillation than twice the number of 
seconds beat by the chronometer, and hence its period is 
easily deduced. These pendulums are of the invariable 
type—the theory of their use depending on the length re¬ 
maining constant except for changes due to temperature. 
Their periods at different stations are then inversely pro¬ 
portional to the square root of gravity, after allowance has 
been made for all variations of conditions which influence 
the time of oscillation. To eliminate these disturbing effects 
the observed periods of the pendulums require five correc¬ 
tions, known as the arc, temperature, pressure, rate, and 
flexure corrections. By applying these the periods are re¬ 
duced to what they would have been under certain standard 
conditions, those adopted for this work being: arc, infinitely 
small; temperature, 15° C.; pressure, 60 millimeters at 0° C., 
true sidereal time and inflexible support. The arc correc¬ 
tion is based on the usual theoretical reduction and the 
assumption that the arc diminishes in geometrical ratio as 
the time increases in arithmetical ratio. The effects of 
changes of temperature, pressure, and flexibility of support 
were investigated experimentally and coefficients deduced. 
An elegant comparative method, first used by Airy,* was 
*Phil. Trans, of the Royal Society, 1856, p. 297. 
