November 12, 1908] 



NA TURE 



55 



the constancy of the velocity of the rays, indicated by the 

 straight edges of the deflected band, is a strong argument 

 against this explanation, and that the existence of the 

 negative rays is conclusive against it. These negatively 

 electrified rays, which form the faintly luminous portion of 

 the phosphorescence indicated in Fig. 6, are not kathode 

 rays. The magnitude of their deflection shows that the 

 ratio of e/m for these rays, instead of being as great as 

 i/Xio", the value for kathode rays, is less than lo'. 

 The particles forming these rays are thus comparable in 

 size with those which form the positive rays. The exist- 

 ence of these negatively electrified rays suggests at once 

 an explanation, which 1 think is the true one, of the con- 

 tinuous band into which the spot of phosphorescence is 

 drawn out by the electric and magnetic fields. The values 

 of elm which are determined by this method are really 

 the mean values of e, m, while the particle is in the electric 

 and magnetic fields. If the particles are for a part of 

 their course through these fields without charge, they will 

 not during this part of their course be deflected, and in 

 consequence the deflections observed on the screen, and 

 consequently the values of e/m, will be smaller than if the 

 particle had retained its charge during the whole of its 

 career. Thus, suppose that some of the particles con- 

 stituting the positive rays, after starting with a positive 

 charge, get this charge neutralised by attracting to them 

 a negatively electrified corpuscle, the mass of the cor- 

 puscle is so small in comparison with that of the particle 

 constituting the positive ray that the addition of the particle 

 will not appreciably diminish the velocity of the positive 

 particle. Some of these neutralised particles may get 

 positively ionised again by collision, while others may get 

 a negative charge by the adhesion to them of another 

 corpuscle, and this process might be repeated during the 

 course of the particle. Thus there would be among the 

 rays some which were for part of their course unelectrified, 

 at other parts positively electrified, and at other parts 

 negatively electrified. Thus the mean value of e/m might 

 have all values ranging from o, its initial value, to —a', 

 where o' might be only a little less than a. This is just 

 what we observe, and when we remember that the gas 

 through which the rays are passing is ionised, and con- 

 tains a large number of corpuscles, it is, I think, what 

 we should expect. 



At very low pressures, when there are very few ions 

 in the gas, this continuous band stretching from the origin 

 is replaced by discontinuous patches. 



Positive Rays in Hydrogen. 



In hydrogen, when the pressure is not too low, the 

 ■brightness of the phosphorescent patch is greater than in 

 air at the same pressure ; the shape of the deflected phos- 

 phorescence is markedly different from that in air. In 

 air, the deflected phosphorescence is usually a straight 

 band, whereas in hydrogen the boundary of the most 

 ■deflected side is distinctly curved and is concave to the 

 undfflected position. The appearance of the deflected 

 phosphorescence is indicated in Fig. 7. 



The result indicated in Fig. .S. which was also obtained 



-with hydrogen, shows that we have here a mixture of 

 two bands, as indicated in Fig. 4, the two bands being 

 produced by carriers having difl'erent maximum values of 

 ■c'm. The greatest value of e/iii obtained with hvdrogen 

 was the same as in air, 1.2x10', the velocitv was 

 j-Sxio' cm. per sec. The presence of the second band 

 indicates that mixed with these n^e have another set of 



NO. 2037, VOL. 79] 



carriers, for which the maximum value e/m is half that 

 in the other band, i.e. 5x10". The curvature of the 

 boundary generally observed is due to the admixture of 

 these two rays. 



Positive Rays in Helium. 



In helium the phosphorescence is bright, and the de- 

 flected patch has in general the curved outline observed in 

 hydrogen. I was fortunate enough, however, to find a 

 stage in which the deflected patch was split up into two 

 distinct bands, as shown in Fig. 9. The maximum value 

 of elm in the band a was 1-2 xio', the same as in air 

 and hydrogen, and the velocity was i-8xio', while the 

 maximum value of e/m in band b was almost exactly one 

 quarter of that in a (i.e. 2-9x10'). As the atomic weight 

 of helium is four times that of hydrogen, this result 

 indicates that the carriers which produce the band 

 b are atoms of helium. This result is interest- 

 ing, because it is the only case (apart from 

 hydrogen) in which I have found values of elm 

 corresponding to the atomic weight of the gas ; 

 and even in the case of helium, when the pressure 

 in the discharge-tube is very low and the electric 

 field very intense, the characteristic rays with 

 <;/m = 2-9Xio' sometimes disappear, and, as in 

 all the gases I have tried, we get two sets of 

 rays, for one set of which e/m = 10'' and for the "^' '' 



other 5 X 10'. 



Although the helium had been carefully purified from 

 hydrogen, the band a (for which e/m = 10') was generally 

 the brighter of the two. The case of helium is an interest- 

 ing one ; for the class of positive rays, known as the a 

 rays, which are given off by radio-active substances, would 

 a priori seem to consist most probably of helium, since 

 helium is one of the products of disintegration of these 

 substances. The value of e!m for these substances is 

 5X10', where we have seen that in helium it is possible 

 to obtain rays for which e/m = 2-9Xio'. It is true that, 

 at very low pressures and with strong electric fields, we 

 get rays for which e/m = 5Xio'; but this is not a 

 peculiarity of helium ; all the gases which I have tried 

 show exactly the same effect. 



Argon. 



When the discharge passed through argon, the effects 

 observed were very similar to those occurring in air. The 

 sides were perhaps a little more curved, and there was a 

 tendency for bright spots to develop. The measurements 

 of the electric and magnetic deflection of these spots gave 

 c'/m = io', the value obtained for other cases. There was 

 no appreciable increase of luminosity in the positions corre- 

 sponding to e'm = 10^/40, as there would have been if an 

 appreciable number of the carriers had been argon atoms. 



Positive Rays in Gases at very loxv Pressures. 



As the pressure of the gas in the discharge-tube is 

 gradually reduced, the appearance of the deflected phos- 

 phorescence changes ; instead of forming a continuous 

 band, the phosphorescence breaks up into two isolated 

 patches ; that part of the phosphorescence in which the 

 deflection was very small disappears, as also does the phos- 

 phorescence produced by the negatively electrified portion 

 of the rays. 



In the earlier experiments considerable difficulty was 

 experienced in working at these very low pressures ; for 

 when the pressure was reduced sufficiently to get the effects 

 just described, the discharge passed through the tube with 

 such difliculty that in a very few seconds after this stage 

 was reached sparks passed from the inside to the outside 

 of the tube, perforating the glass and destroying the 

 vacuum. In spite of all precautions, such as earthing the 

 kathode and all conductors in its neighbourhood, perfora- 

 tion took place too quickly to permit measurements of the 

 deflection of the phosphorescence. 



This difliculty was overcome by taking advantage of the 

 fact that, when the kathode is made of a very electro- 

 positive metal, the discharge passes with much greater ease 

 than when the kathode is made of aluminium or platinum. 



