160 AN EMPIRICAL STUDY OF GYRATING BODIES. 



ment is thrown back to the vertical, and by the peculiar 

 and indescribable twist or rolling motion, which is pro- 

 duced by rotation about the axis d a (the twist must be 

 seen to be understood), the instrument tilts the other 

 way, and the whole process is gone through again, 

 causing quite a number of vibrations, each of greater 

 amplitude than its predecessors, till at last the instru- 

 ment falls. 



Experiment : Replace the wire p d (fig. 33|), by a 

 wheel, the same as is attached to the other end of the 

 frame at o. The instrument is then mounted upon two 

 wheels (see fig. 34^, plate II), whose axes are approxi- 

 mately in the plane of the large wheel ; (there is no need 

 of exactness here). Set the wheel in motion, and place 

 the instrument on a smooth, hard surface ; (glass is best). 

 It will stand up with great firmness, and execute a re- 

 markable series of pirouettes. 



Rationale : There is nothing new here, and no further 

 explanation is needed. 



If we call the point at which there is freedom of motion 

 a joint, there is only one joint in a gyroscope, and one 

 in a top ; there is only one in the common gyrostat. 

 About these joints there is also horizontal rotation. 

 With the gyrostrat on wheels there are two joints ; but 

 at least one must have liberty to revolve around the 

 other. They both may move ; but if both are unable to 

 move, the instrument falls. With the gyroscope on 

 stilts, there must be three joints ; and the two points of 

 support must not change their position. If neither leg 

 has a free socket at top the instrument will fall, or if 

 both legs have it ; or, if with one free socket at top, 

 either or both bottom joints are free to change their 

 positions, the same result follows. 



GYROSTATIC TRAPESE AND TIGHT ROPE PERFORMANCES. 



By means of the loops n and m, fig. 32, attach the 



