238 Sir J. J. Thom: 



on on 



of the ions under an electric field is not larger than it is. 

 The smallness of the velocity can also be explained as due to 

 the charge on the ion, so that we could not deduce the 

 existence o£ these aggregates from considerations of the 

 velocity of the ion. It may be pointed out that the radiation 

 due to these aggregates is in every case faint compared with 

 that due to the simple atom, so that at the low pressures at 

 which the experiments on positive rays are made the number 

 of these complex aggregates is very small compared with that 

 of the simple atom. 



In marsh-gas there is a secondary radiation with an 

 electrical atomic weight of 36 ; an aggregate of three atoms 

 of carbon with one charge between them would explain this. 

 In other hydrocarbons, as well as in CO and ON, the 

 strono-est secondary radiation is one for which the measure- 

 ments of the magnetic deflexion give values ranging from 12 

 to 78. This is the strong secondary ^hown in fig. 18, the 

 gas in the tube being 00. An aggregate of six carbon 

 atoms, or a molecule of benzene, 6 H 6 , with one charge, 

 would give electrical atomic weights agreeing within the 

 errors of experiment with this. 



I have tabulated these results for the secondary radiation 

 of the gases I have so far examined in the following Table 

 (p. 239). The first column contains the electrical atomic weight 

 of the particles in the radiation, the second the sign of the 

 charge on these particles (thus + means that they are always 

 positively charged, + and — that they occur with negative 

 as well as positive charges), the third column contains the 

 name of the gas giving the radiation, the fourth the probable 

 composition of the particle, and the fifth whether it is a true 

 secondary , i. e. due to dissociation after the rays have passed 

 through the cathode or to recombination. 



Pr diary Rays. 



These are the rays which produce on the photographic 

 plate detached parabolic curves not passing through the 

 origin. The primary rays found in the gases investigated 

 up to the present are as follows : — 



Hydrogen. 



There are two kinds of primary rays having respectively 

 the electric atomic weights 1 and 2 corresponding to the 

 atom and molecule of hydrogen. The relative brightness of 

 the curves due to the atom and molecule varies greatly with 

 the circumstances of the discharge, in some cases that due 



