November 12, 1908] 



NA TURE 



53 



dermic injections, which I find answer excellently for this 

 purpose. After getting through the needle, the positive 

 rays on their way down the tube pass betw'een two parallel 

 aluminium plates A A. These plates are vertical, so that 

 when they are maintained at different potentials the rays 

 arc subject to a horizontal electric force, which produces 

 a horizontal deflection of the patch of light on the screen. 

 The part of the tube containing the parallel aluminium 

 plates is narrowed as much as possible, and passes between 

 the poles P P of a powerful electromagnet of the Du Bois 

 type. The poles of this magnet are as close together as 

 the glass tube will permit, and are arranged so that the 

 lines of magnetic force are horizontal and at right angles 

 to the path of the rays. The magnetic force produces a 

 vertical deflection of the patch of phosphorescence on the 

 screen. To bend the positive rays it is necessary to use 

 strong magnetic fields, and if any of the lines of force 

 were to stray into the discharge-tube in front of the kathode 

 they would distort the discharge in that part of the tube. 

 This distortion might affect the position of the phosphor- 

 escent patch on the screen, so that unless we shield the 

 discharge-tube we cannot be sure that the displacement of 

 the phosphorescence is entirely due to the electric and 

 magnetic fields acting on the positive rays after they have 

 emerged from behind the kathode. 



To screen off the magnetic field the tube was placed in 

 a soft iron vessel W with a hole knocked in the bottom, 

 through which the part of the tube behind the kathode 

 was pushed. Behind the vessel a thick plate of soft iron 

 with a hole bored through it was placed, and behind this 

 again as many thin plates of soft iron, such as are used 

 for transformers, as there was room for, w'ere packed. 

 When this was done it was found that the magnet pro- 

 duced no perceptible effect on the discharge in front of the 

 kathode. 



The object of the experiments was to determine the value 

 of chn by observing the deflection produced by magnetic 

 and electric fields. When the rays were undefiected they 

 produced a bright spot on the screen ; when the rays 

 passed through electric and m.agnetic fields the spot was 

 not simply deflected to another place, but was dra.vn out 

 into bands or patches, sometimes covering a considerable 

 area. To determine the velocity of the rays, and the value 

 of c in. It was necessary to have a record of the shape of 

 these patches. This might have been done by substituting 

 a photographic plate for the willemite screen. This, how- 

 ever, was not the method adopted, as, in addition to other 

 inconveniences, it involves opening the tube pnd re-pump- 

 ing for each observation, a procedure which would have 

 involved a great expenditure of time. The method actuallv 

 adopted was as follows : — The tube was placed in a darl< 

 room from which all light was carefully excluded, the tube 

 itself being painted over, so that no light escaped from it. 

 In these circumstances the phosphorescence on the scr^-n 

 appeared bright and its boundaries well defined. The 

 observer traced in Indian ink on the outside of the thin 

 flat srreen the outline of the phosohorescence. When this 

 had been satisfactorily necomplished the discharge was 

 stopoed. the light admitted into the room, and the pattern 

 on the screen transferred to tracing-paper ; the deviations 

 were then measured on these tracmgs. 



Calculation of the Magnetic and Electric Deviation 

 of the Rays. 



If we assume the electric field to be uniform between 

 Ihe plates and zero outside them, then we can easily show 

 Ih.nt .-v, the horizont.il deflection of a ray the charge of 

 which is e, mass ni, and velocity i', is given by the 

 equation 



r^iX 



,/(/+2,/), 



where X is the force between the plates, 1 the length of 

 path of the rays between the plates, and d the distance 

 of the screen from the nearer end of the parallel plates. 



To find the deflection due to the magnetic field, we have, 

 if p is the radius of curvature of the path at a point where 

 tlie magnetic force is H, 



""'' = 11..., 

 P 



XO. 2037, VOL. 79] 



p f/lV 



If y is the vertical displacement of the particle, we have 



I cf-v ... 



- = -^ approximately, 



p (h- 



where : is measured along the path of the ray. Hence 



H: 



""^Lio io "J 



(I) 



In these strong fields there are considerable variations 

 of H along the path, so that to calculate the integrals we 

 should have to map out the value of H along the path of 

 the ray. This would be a very laborious process, and it 

 was rendered unnecessary by the following simple method, 

 which, while not involving anything like the labour of the 

 direct method, gives much more accurate results. The 

 method is shown in Fig. 3. The part of the tube through 

 which the rays pass was cut off, and a metal rod placed 

 so that its tip Z coincided with the aperture of the narrow 

 tube through which the positive rays had emerged. A 

 very fine wire soldered to the end of this tube passed over 

 a light pulley, and carried a weight at the free end. The 

 pulley was supported by a screw, by means of which it 

 could be raised or lowered ; a known current passed 

 through the wire, entering it at Z and leaving it through 

 the pulley. The pulley was first placed so that the path 

 of the stretched wire when undefiected by a magnetic field 

 coincided with the path of the undefiected rays. A vertical 

 scale, the edge of which was at the same distance from the 

 opening through which the rays emerge as the screen on 

 which the phosohorescence had been observed, was placed 



just behind the wire, and was read by a reading microscope 

 with a micrometer eye-piece. When the magnetic field 

 w-as put on, the wire was deflected ; and if T is the tension 

 of the wire, p the radius or curvature into which it is bent, 

 I the current through the wire, 



T = H/; 



P 

 or, if y, is the vertical displacement of the wire, 



€Zi=L.n 



dz- T 



Now if ^ =0 when c = o we have, \i }\ is the dispUcen'ent 

 dz 

 of the wire at the scale, 



y,=U'.rHdz 

 ' .' u .'0 

 Hence, tomparing (i) and (2) we have 



(2) 



(3) 



