Effect of Electrical Con vection . 191 



condensing-plates, the distance between which was just great 

 enough so that the coil hung perfectly free. With regard 

 to this coil, two things were found absolutely necessary : 

 first, to have it completely shielded from all direct electro- 

 static action, and secondly, to have it so suspended as to be 

 perfectly free from mechanical jarring. The following 

 arrangement, finally adopted after considerable experimenting, 

 has proved entirety satisfactory. 



Two brass hoops, 1*5 cm. thick, were placed around the 

 coil, one hoop outside, the other inside. The coil had been 

 previously wrapped with insulating-tape. The space between 

 the two hoops was then carefully covered with tinfoil. In 

 this way was formed a closed metal case only slightly wider 

 than the thickness of the coil. The terminals of the coil 

 were brought out from this case and were carefully wrapped 

 in tinfoil. The metal case and foil were connected to earth. 

 To the outside hoop w r ere soldered two heavy brass strips, U, 

 fixed to the wooden block W provided with levelling-screws. 

 This frame containing the coil was stiffened by a brass rod 

 connecting the strips just above the coil. The block W rested 

 on a heavy marble slab suspended from the ceiling by four spiral 

 springs. To prevent the air-currents generated by the rapid 

 motion of the disk from setting this system in vibration, 

 cardboard strips were screwed down over the openings 

 between the condensing-plates, so that each disk was com- 

 pletely boxed in. 



The Galvanometer. — As has already been stated, the gal- 

 vanometer employed was an extremely sensitive one. It is 

 an astatic instrument of the ordinary four-coil form, with a 

 low resistance., 58*4 ohms. The needle, which together with 

 its mirror weighs only 25 mg., is suspended by an extremely 

 fine quartz fibre, invisible almost in ordinary light. The 

 period of the needle when vibrating in the earth's field was 

 from 3 to 5 seconds, depending on how perfectly the needle 

 was rendered astatic. The period was usually increased to 

 about 25 seconds by cutting down the earth's field by means 

 of control magnets. With this period a current of 1 x 10~ lu 

 amperes gives a deflexion of 1 mm. on a scale l'l metres 

 distant. The deflexion of the needle was read by the deflexion 

 of the image of an incandescent light filament reflected from 

 the needle mirror, on a ground-glass scale 1*1 metres distant. 

 The needle was absolutely dead-beat. 



A heavy marble block resting on a bed of excelsior ou a 

 marble slab bracketed to the wall formed a support for the 

 galvanometer which rendered it entirely free from mechanical 

 jarring. The galvanometer was also shielded magnetically by 



