448 



SCIENCE 



[N. S. Vol. XXV. Mo. 638 



firm of clock manufacturers is using invar 

 or nickel steel for its pendulum rods. 



Attempts have also been made to com- 

 pensate a pendulum to avoid changes in 

 the rate due to changes of the barometer. 

 At Greenwich this is accomplished by at- 

 taching a magnet to the end of the pendu- 

 lum and causing a second magnet sup- 

 ported below the pendulum to approach 

 or recede from the pendulum with changes 

 of the barometer. This second magnet in- 

 creases the acceleration due to gravity to a 

 greater or less degree as it approaches or 

 recedes from the pendulum. 



The most satisfactory way, however, of 

 freeing a clock rate from variations due to 

 changes of temperature and pressure is to 

 enclose the clock in an air-tight glass case 

 kept in a room of constant temperature. 

 A number of observatories now keep their 

 standard clocks under such conditions. 



Anchor or Recoil Es- 

 capement. 



Dead-beat Escapement. 



Another improvement in the manufac- 

 ture of clocks was accomplished by Graham 

 in the introduction of the dead-beat escape- 

 ment. 



In Fig. 5 are shown both the anchor or 

 recoil escapement and the dead-beat escape- 

 ment. In the first, shown on the left, the 

 pendulum moving to the left has just es- 

 caped a tooth at the left-hand pallet and 

 allowed a tooth to fall on the right-hand 

 one. The pendulum, however, still con- 

 tinues its swing to the left, and in conse- 

 quence the pallet pushes the wheel back, 

 thus causing the recoil that gives the name 



to the escapement. It is only after the 

 pendulum comes to rest and begins its ex- 

 cursion the other way that it gets any as- 

 sistance from the wheel, and the difference 

 between the forward motion of the wheel 

 and its recoil forms the impulse. In the 

 right-hand figure, the pendulum moving to 

 the right has just escaped a tooth from the 

 right-hand pallet while another has fallen 

 upon the left-hand one. As the pendulum 

 continues its motion towards the right, the 

 left hand-pallet slides over the point of the 

 tooth, but there is no recoil, as the 'dead' 

 face, as it is called, is the arc of a circle 

 whose center is the point about which the 

 anchor turns. As the pendulum returns 

 towards the left, the tooth traverses the 

 'dead' face in the opposite direction, and 

 immediately upon leaving this face it 

 passes to the 'impulse' face, and while pass- 

 ing along this face gives the impulse to the 

 pendulum. 



The great advantage of the dead-beat 

 escapement over the anchor or recoil type 

 is that, although a slight increase of force 

 on the escapement wheel increases the arc 

 of the pendulum, it does not sensibly in- 

 crease the time, while the time does sensibly 

 increase with the recoil escapement. 



At about this time also, Mudge intro- 

 duced the gravity escapement. With all 

 the previous escapements the impulse was 

 given to the pendulum by the driving 

 weight acting through the train so that any 

 irregularities in the train would cause a 

 variation in the impulse. With the gravity 

 escapement a weight is raised by the train 

 and the falling of this weight gives the 

 impulse to the pendulum. We thus have a 

 uniform impulse at each oscillation due to 

 the falling of a weight through a fixed dis- 

 tance. Simple and elegant as is this the- 

 ory, the application of it gave a great deal 

 of trouble and all gravity escapements were 

 regarded with suspicion, as having a tend- 

 ency to trip, until Mr. Denison designed 



