('. B. WARRING. L79 



wards the ecliptic till the maximum is reached, and then 

 this force begins to die away, the pole once more recedes, 

 and so on forever. 



But apparently correct reasoning, especially when 

 dealing with such illusive movements as those of rotating 

 bodies, sometimes fails, either from error in our logic, 

 or from the unobserved omission of some important fact 

 from our calculation. 



I therefore asked myself, Will this bear the test of 

 experiment % 



After thinking over several plans for solving the 

 problem, I adopted the following : 



I attached to my large balanced gyroscope, at c, fig. 2, 

 and approximately in line with the axis, a small tin pail 

 (one-half pint or thereabouts), having an opening in the 

 bottom. I then adjusted the movable weight till the 

 axis of the wheel was horizontal. 1 The opening was 

 closed with an outside valve, so suxiported that setting- 

 lire to a thread would cause it to fall. I then put in the 

 pail two or three pounds of fine shot, and set the wheel 

 in motion. The instrument began immediately to move 

 laterally — quite rapidly, and also to tilt downward. 

 AVhen this had gone on a few moments, I opened the 

 valve, while the instrument was in motion, by bringing 

 a blaze against the thread. The shot at once began to 

 pour out ; the lateral motion continued in the same 

 direction, but the loaded end, which had been going- 

 down, began to rise. The precessional motion grew 

 slower and slower, as the loaded end went up, until, at 

 the moment the £>ail became empty, both precession and 

 tilting ceased. 



At another time, I closed the valve when perhaps one- 

 half of the shot had escaped ; the gyroscope at once in- 



1 To succeed best with this, the bottom of the pail should be conical, and the opening 

 three-eighths of an inch, or so, in diameter. The point of support should be only a very 

 little above the centre of gravity of the whole system, so that the instrument will oscillate 

 with very little force. 



163 



