164 AN EMPIRICAL STUDY OF GYRATING BODIES. 



happened to have it. But in fact the wheel, in pro- 

 portion to its mass, is more efficient. The results are the 

 same with either. 



A is a conical steel point, very fine and hard, resting 

 in a polished cup of hardened steel. Two stout wires 

 are firmly screwed into the brass semi-circle, and by the 

 two weights attached, lower the centre of gravity, and 

 give stability to the instrument when balanced on the 

 point A. Two small hooks, h h, are used when it is 

 desired to susjDend the machine from the ceiling. The 

 arrows denote the direction of rotation. 



Now set the wheel to revolving in the direction of the 

 arrows o o. It will run perhaps ten minutes, but no 

 other motion will be produced. Attach a weight to the 

 pin, B. The instrument will at once start off in a hori- 

 zontal movement in the direction indicated by the di- 

 rectional ellipse around the standard. As the instru- 

 ment is carefully balanced in every way, and moves with 

 very little friction, a fiftieth part of the same weight on 

 the pin before the wheel was set revolving would at once 

 have tilted it till the axis was vertical, but now, ap- 

 parently, it does not tilt at all. This, however, is an 

 illusion. It is slowly tilting so that the weight is de- 

 scending all the time — a fact that in a minute or two be- 

 comes very evident. Here, as elsewhere, we have the 

 law. No fall, no gyration. 



We have in the Bohnenbergher some of the problems 

 that met us in the gyroscope. The rationale is equally 

 simple. The w r eight on B causes the axle to tilt down a 

 little ; that makes the lower o move from you ; as it 

 rises in its rotation to the horizontal line, having lost no 

 energy, it pushes that end (the right hand of the figure) 

 from you. This causes the gyration ; and, as it rises to 

 the top, it pushes that side from you. And, pushing 

 the upper side from you, means pushing against the 

 weight that seeks to pull B down. The upper o goes 



148 



