256 
THE TEACHING OF MECHANICS BY EXPEEIMENT. 
substantially the experiment by which Galileo satisfied himself as to 
the character of uniform acceleration, namely, by letting a ball roll 
down an inclined groove. 
Fig, 3, Fall rolling on inclined railway , 
Here the groove is formed by planing out a portion of a stout steel 
tube. A ball such as is now readily obtained for the purpose of ball 
bearings will do well. Placing it at the top against a trigger we are 
able to release it and at the same time to make an electrical signal 
upon any form of chronograph, showing the instant of its release, 
and when it reaches the end of its course it hits another trigger which 
makes tke second electrical signal. The chronograph then shows 
electrically the interval of time that the ball has taken to roll down 
the plane. W e may vary the distance that it rolls down the plane by 
setting the stop piece at various places along the groove. We may 
also vary the inclination of the path itself by raising or lowering a 
bracket on which one end of the grooved tube rests. A waterclock 
makes a convenient chronograph, or the record may be made by a 
Morse telegraph instrument reeling off paper at a nearly uniform 
rate. The length of the line printed in ink upon the paper will show 
the duration of time that the ball has taken to roll down, for the cur¬ 
rent is made when the ball is released and broken when it reaches the 
stop. 
What the student does is first t© investigate experimentally the 
relation between the length of the path and the time taken, and he 
finds that one is proportional to the square root of the other. He 
takes the time and the square root of the distance as his co-ordinates 
for plotting on squared paper, and he finds they give a straight line 
which however does not pass exactly through the origiu but cuts the 
axis of time a short distance away from the origin. That gives him 
the clock error due to time-lag in the mechanism of the chronograph : 
in the particular instrument that I am using at Cambridge the clock 
error is about J^th of a second, and it is determined with very con¬ 
siderable accuracy by seeing how near the straight line connecting 
these quantities passes to the origin. 
Another experiment in elementary dynamics, intended to illustrate 
simple harmonic motion, is arranged by having a weight hung from a 
