Royal Astt'ononiical Society. 147 



Suppose the last, wheel of the moving- train to be a beveled wheel 

 whose plane is horizontal, carried by a vertical spindle ; and suppose 

 the spindle of the centrifugal balls to be above the moving-train 

 spindle, in the same vertical line, but separate, and carrying a be- 

 veled wheel similar to the other, but in the opposite position ; and 

 suppose the connexion between these two beveled wheels to be 

 made by an intermediate beveled wheel, whose plane is vertical, and 

 whose axis passes between the two ; the train-spindle and the ball- 

 spindle will evidently rotate in opposite directions. The axis of the 

 intermediate wheel must not be fixed, but must be capable of turning 

 in a horizontal plane, and it must be pulled by a weight (acting by 

 a line over a pulley, or a right-angled lever), in the direction of 

 turning the ball-spindle ; it must also be connected with the water- 

 valve in such a way that the same pull will tend to open the water- 

 valve. It is evident now that, while the axis of the intermediate 

 wheel is free to move, the force driving the beveled wheel of the 

 ball- spindle is rigorously uniform, and is equal to half the weight 

 which pulls the axis of the intermediate wheel. If the moving power 

 is suddenly increased, the immediate efi^ect of this increase is, not to 

 accelerate the ball-spindle, but to drive the intermediate wheel in 

 such a manner that its axis lifts the pulling- weight, and partially 

 closes the water-valve. An opposite motion takes place if the moving 

 power is suddenly diminished. In this manner the water-valve is 

 maintained in that state which supplies the force that exactly pro- 

 duces a determinate moving power upon the ball-spindle. 



In some localities it may be difficult to obtain the proper supply 

 of water, and in some there may be danger of obstruction by frost. 

 The Astronomer Royal had, therefore, endeavoured to efi^ect the 

 same object by the use of weights. By a remontoir- train the uni- 

 formity of rotation of the balls might undoubtedly be secured ; but 

 then the motion of the primary wheels of the train would be inter- 

 mittent. The following plan, however, appears to the Astronomer 

 Royal to be perfect. Let the arrangement of the three beveled 

 wheels and the pulling- weight be exactly the same as that above 

 described, but let the train spindle carry a broad flat disc, whose 

 plane is horizontal : upon this let a lever press by a definite projec- 

 tion from its lower side, and on the upper side of the lever let there 

 be a sliding weight (as a ball running in a groove) ; let the spindle 

 of the intermediate beveled wheel command the place of this ball 

 by a fork ])rojecting downwards, loosely including it ; and let the 

 tendency of the pulling weight be to draw this ball towards or over 

 the fulcrum of the lever, so as to diminish the pressure of its definite 

 projection upon the flat disc. Then, if the moving force is suddenly 

 increased, its immediate effect will be to drive the intermediate 

 wheel in such a direction as to push the running ball further along 

 the lever, and thus to increase the pressure of its projection upon 

 the flat revolving disc, and to produce a friction which will absorb 

 the excess of power. 



There are still two practical defects which it is desirable to re- 

 move. 



L2 



