the Filar Anemometer. 85 



interrupted at high speeds if both springs lie in a plane 

 through the axis and are of proper lengths and curvature. 



Fig-. 1. — Vertical Spring Microphonic Contact on Air-Wire. 



As / is insulated, the current of four Leclanche cells passes 

 through/, e, d, c, a, g, and the Bell telephone included in the 

 circuit. The sounds are very much increased (and unfortu- 

 nately the thunder due to breaks of contact also) when the 

 primary of a transformer is placed in the circuit with the 

 telephone in the secondary, in the now well-known way*. 

 A great advantage is gained by filing the flat surface of e 

 with a smooth file, probably owing to the carbon powder 

 which lodges in ihe pores. Polishing the surface on a stone 

 is harmful. High tones seem to require firmer contacts than 

 low tones, and this state of things is brought about by the 

 action of centrifugal force as stated. The position of the 

 spring cd on the wire aa seems to be immaterial. I made 

 tests throughout 30 cm. without marked results. Nor is 

 much advantage gained from the shape or material of the 

 sounding-board. Curiously enough, the latching of d and e 

 often favoured the intonation of low sounds. No rules can 

 be given for the bending of the springs, and the best contact 

 is usually the result of chance, and is obtained only after 

 many trials. 



It is with the above arrangement that I once noted a 

 peculiar phenomenon for which I have failed to find an 



* The reader is referred to either Preece's or Du Moncel's Treatise on 

 the Telephone. 



