ON IMPROVEMENTS IN ASTRONOMICAL CLOCKS. 57 



Now let us follow the action of this escapement. 



First suppose the battery to be attached when matters are in the 

 position shown in fig. 1. The effect will bs that the arm A will be di-awn 

 against p^- If ^^ now set the pendulum swinging to the right the 

 impulse arm w will follow the pendulum as far as the arm A will allow it 

 to do so, but on reaching this limit, the pendulum will leave the impulse 

 arm and continue to swing to the right alone. 



The instant, however, that the contact between w and P is thus broken, 

 M is no longer an electro-magnet, and the arm A is drawn by the spiral 

 spring to the position of fig. 2 ; the pendulum continues its swing to the 

 right, comes to rest, and returns. On its return it encounters the impulse- 

 piece w, not where it left it (viz., at its lowest limits, the arm A resting on 

 P2), but as in fig. 2, the arm A resting on p,. When P and \v encounter, 

 the immediate result is that, contact being formed, M becomes a magnet, 

 and the arm A is drawn against jjj. whilst the impulse-piece w continues 

 its motion towards the left, along with the pendulum, and returns again 

 to the right with the pendulum till it is stopped by encountering the arm 



A pressing against ^2- 



Simply stated, the impulse is this : — The pendulum in swinging 

 against the impulse-weight picks it up at _pi, and in swinging with the 

 impulse-weight it carries it on pastp, as far as p<i. The effective impulse 

 is, therefore, that of the fall of the resolved horizontal force of w in falling 

 from p, to j)2- 



This force is absolutely constant. 



There is no locking or unlocking, and no friction, and no element of 

 change except such as may be due to the electric contact between w and P. 

 Such contacts are liable to wear and to stick, and it was not until some 

 prospect offered of overcoming this fault that we ventured to request 

 a grant from the Association. The plan of escapement had already been 

 contrived and tried experimentally by Mr. Gill ; but it was in conse- 

 quence of an idea of Mr. Gimingham's that it first seemed possible to 

 overcome the outstanding difficulty and attain a nearer approach to 

 perfection. 



Mr. Gimingham's idea was to construct a relay which could be 

 worked by radiation. This relay he first contrived for the purpose of 

 registering the number of revolutions of a radiometer. 



The form which this relay has now assumed, after a variety of expe- 

 riments, is shown in fig. 4. 



Fig. 4. 



K is a very light arm of aluminium, mounted on needle-poinfcB. 



B is a fan of mica, coated on one side with lampblack. 



c is a carbon point attached to k. 



By means of an aluminium ring r, fitting spring-tight into a glass tube, 

 the supports of the needle-points of k are fixed in position — the supports 

 being attached to the ring. 



Another ring, t, carries a small carbon anvil, against which the carbon 

 point c can come in contact. 



