the lower side of A, and the
pressure acting on the under side of B is sufficient to cause it to
lift the valve A, and to admit air from the atmosphere, both to the
brake-cylinder and the train-pipe, through the clappet-valve D, which
also rises because of the difference of pressure on its two sides. In
a graduated application, neither D nor A rises from its seat, but air
from the train-pipe finds access to the brake-cylinder by passing
around the peg C, which is so proportioned as to allow the necessary
amount of air to enter the brake-cylinder, and so obtain simultaneous
action of the brake throughout the train. When the handle E is turned
so as to prevent the clappet D from rising, the rapid action is cut
out and the brake acts as an ordinary vacuum automatic brake. A
modification of the device for obtaining accelerated action, described
above in connexion with the Westinghouse brake, is also applicable.
Accelerating chambers, again containing air at atmospheric pressure,
are provided on each vehicle and are connected with the train-pipe by
valves which open as the vacuum in the latter begins to decrease with
the operation of the driver's valve. The air thus admitted into the
train-pipe effects a still further local reduction of the vacuum,
which is sufficient to actuate the accelerating valve of each next
succeeding vehicle and is thus rapidly propagated throughout the
train.

Brake trials.

Famous tests of railway brakes were those made by Sir Douglas Galton
and Mr George Westinghouse on the London, Brighton and South Coast
railway, in England, in 1878, and by a committee of the Master Car
Builders' Association, near Burlington, Iowa, in 1886 and 1887. The
object of the former series (for accounts of which see _Proc. Inst.
Mech. Eng._, 1878, 1879) was to determine the co-efficient of friction
between the brake-shoe and the wheel, and between the wheel and rail
at different velocities when the wheels were revolving and when
skidded, i.e. stopped in their rotation and caused to slide. These
experiments were the first of their kind ever undertaken, and for many
years their results furnished most of the trustworthy data obtainable
on the friction of motion. It was found that the co-efficient of
friction between cast-iron shoes and steel-tired wheels increased as
the speed of the train decreased, varying from 0.111 at 55 m. an hour
to 0.33 w