C. B. WARRING. L55 



the same time revolving around a vertical axis in a 

 direction dependent upon the way the wheel revolves 

 and the direction in which it leans. If either of these 

 be changed the instrument goes the other way. If both 

 are changed it is unaffected. 



The rationale of these movements is no more difficult 

 than of those already considered. In fact it is the same 

 thing slightly modified. Let fig. 33 represent our in- 

 strument with wheel in rapid rotation. Suppose it falls 

 a little towards the observer. Any molecule on the 

 upper side, e. g., m., will, by that fall, acquire a motion 

 in the same direction. SujDpose it to be carried in an 

 instant to the position p ; it will tend to push that side 

 of the instrument towards the observer, and since the 

 friction at the point where the curved edge rests on the 

 glass j)late is almost zero, q, will, at the same instant, 

 recede from him. In another exceedingly small fraction 

 of a second, q will be at the top while still receding, 

 and consequently it will tend to push the top back to a 

 vertical. 



If m had tilted from the observer at the start, the ro- 

 tation of B A, which in the first case was with the hands 

 of a watch, would have been in the opposite direction. 



It will be seen that horizontal rotation of B A is neces- 

 sary in order that the instrument should stay up. The 

 rotation, or angular motion may be exceedingly small, 

 but there must be freedom for it to exist, or the gyrostat 

 will fall. 



Experimental proof : Set the knife edge B A in a 

 groove so that it can move neither to the right nor left, 

 and it will fall as quickly when the wheel is rotating as 

 when it is not. 



Or, use the piece with pins, fig. 32, H A. Set the 

 wheel in motion, and place the instrument on wood. It 

 will fall at once ; place it on glass, it will stay up. In 

 the first case, the points entering the wood prevent hori- 



139 



