158 THE TOP. - 
opposite to the motion of the hands of a watch. Call 
the right hand arrow A, the upper one B, and the left 
hand C. Raising A makes the top travel from the ob- 
server and parallel to the side opposite A, and so on. 
By using glass plates of various forms, a variety of 
figures can be produced of great beauty. 
I have said that the radius of the curve on the up-hill 
sideis shorter than that on the down-hill side, or in 
other words, the radius is shorter on the side towards 
the end of the path. That sucha difference will 
account for the traveling of the instrument and for the 
space between the curves on the side from which it is 
going being greater than on the other side, can be 
proved experimentally by a process the result of which 
is shown in figure 6, described by a compass-pen. I 
merely shortened the radius for each of the curves to- 
wards A (commencing at the largest end, B). By 
changing the shortening operation from the front to one 
side, I caused the spirals to change their direction. 
The less the inclination of the plate the less the dis- 
tance between the spirals. They may thus be made to 
over-lap and almost coincide. 
_ This is very prettily shown inplate IL. In number 2 
(numbering the three by size) the plate was not touched 
during the tracing, consequently the distance from curve 
to curve is exceedingly uniform inits progress from first 
to last. But in number 3 the rate changes abruptly. For 
42 
