322 SMITHSONIAN MISC^IyLANEOUS COLLECTIONS " VOL. 52 



Note. — This hypothesis seems to be rendered doubtful by the discrep- 

 ancy between the maximum velocity of the rays and that of the sources of the 

 light showing the Doppler effect. If true, since the canal rays must probably 

 be rendered luminous by the collision which ionized them, and emit most light 

 while speeding up, no intensity minimum would be expected. 



Second hypothesis: Sources are gas molecules hit and ionised by the rays. 

 To see how this explains the curve of maximum velocities Fig. 61, consider 

 the case of H rays and H2 rays, having velocities u and 0.71 u respectively. 

 Assuming perfectly elastic collision, maximum possible velocity is, for the 

 collision of 



(i) H ray with H atom, i.oo u; (4) H ray with H molecule, 0.71 u; 



(2) H, " " H " o.94u;(5)H. " " H " 0.67 u ; 



(3) H " " H2 ray, 0.50 u; (6) H " forming " 0.71 u; 

 Assuming collision is not perfectly elastic, energy being lost in radiation and 

 ionization, and that collisions of types (i), (2), (3), are less important with 

 the higher cathode falls, the curve is accounted for. 



Now to see whether this hypothesis explains the intensity minimum. 

 Assuming; that gas molecules hit squarely enough to be ionized, alone emit 

 light, the canal rays being mostly neutralized by the collisions ; that ioniza- 

 tion occurs only when the energy imparted exceeds a certain minimum ; and 

 that the intensity of the light emitted is proportional to the momentum given 

 to the molecule as a result of the collision; the author has calculated by a 

 laborious statistical method (starting with 10,000 canal rays and computing 

 the directions and magnitudes of the velocities of the gas atoms hit in five 

 generations of collisions) the probable distribution of intensity in the resulting 

 Doppler effect. One set of curves is shown in Fig. 63. 



The intensity minimum is seen to be distinct and of fairly constant width, 

 in spite of the fact that the number of sources with small velocities is much 

 greater than the number of the swifter sources. The importance of more data 

 regarding the deflection streak and the Doppler effect so as to decide between 

 these two theories is obvious. 



Carriers oe Band Spectrum. (Stark's hypothesis.) 



Not the positive atoms while in motion since light shows no Doppler 



effect: S5(464, 894); SH(95). 

 Probably neutralized atoms formed by the collision of charged rays with 

 gas atoms, the former being stopped by the collision : 

 S4(6o5); S5(46i, 893); Si2(355); Si3(43, 425); 

 Si9(399) ; Ew3(3i4). 

 Why canal rays neutralized by electrons and retaining their velocity 

 do not emit the band spectrum is not explained. 



Doppler Eefect, Intensity Minimum. 

 Explanation. Either 



(1) Rays of slow velocity are relatively few: Ws(56i, 263) ; W6(588) ; 

 W8(663) ; S2(583) ; Si3(3i, 412) ; Sti(686) ; Tm3(s69). This assump- 

 tion fails to explain the constancy of width of the intensity minimum; or 



(2) Intensity of radiation is a function of the velocity: 810(253); 

 Si3(3i, 177, 180, 412, 435, 439) ; ICn3(36) ; Ps2(2S9). 



Velocity must exceed a certain minimum or no displaced line is ob- 

 tained: Si3(i8o, 439). 



