1863.] |g][ [Coppee. 



tion. This, in fact, is the great problem in all efforts to secure uni- 

 form motion of rotation. This problem was resolved theoretically in 

 many ways, several of which methods were executed mechanically 

 without success, as it was found that the machine stationed as a 

 sentinel to regulate the going of the disk was too weak, and was 

 itself carried off by its too powerful antagonist. The following- 

 method was, however, in the end, entirely successful. Upon the 

 axis of the winding key already mentioned, a toothed wheel was 

 attached, the gearing being so adjusted that one revolution of this 

 wheel should produce a whole number of revolutions of the disk. 

 The circumference of this wheel was cut into a certain number of 

 notches, so that as it revolved, one of these notches would reach the 

 highest point once in two seconds of time. By means of an electro- 

 magnet, a small cylinder or roller, at the extremity of a lever arm, 

 was made to fall into the highest notch of the toothed wheel at the 

 end of every two seconds. In case the disk was revolving exactly 

 once a minute, the roller, driven by the sidereal clock, by means of 

 an electro-magnet, fell to the bottom of the notch, and performed no 

 service whatever ; but in case the disk began to slacken its velocity, 

 then the roller fell on the retreating inclined face of the notch, and 

 thus urged forward by a minute amount the laggard disk, while on 

 the contrary, should the variation from a uniform velocity present 

 itself in an acceleration, then the roller struck on the advancing face 

 of the notch, and thus tended slowly to restore ths equilibrium. Let 

 it be remembered that this delicate regulator has but a minute 

 amount of service to perform. It is ever on guard, and detecting, as 

 it does instantly, any disposition to change, at once applies its re- 

 storing power, and thus preserves an exceedingly near approach to 

 exact uniformity of revolution. This regulator operates through all 

 the wheel-work, and thus accomplishes a restoration by minute incre- 

 ments or decrements spread over many minutes of time. 



'' With a uniformly revolving disk, stationary in position, we should 

 accomplish exactly and very perfectly, the record of one minute of 

 time, presenting on the recording surface thirty dots at equal angular 

 intervals on the circumference of a circle. To receive the time dots 

 of the next minute on a circle of larger diameter, required either that 

 the recording pen should change position, or that at the end of each 

 revolution, the disk itself should move away from the pen by a small 

 amount. We chose to remove the disk. To accomplish accurately 

 the change of position of the disk, at the end of each revolution, the 

 entire machine was mounted on wheels on a small railway track, and 



