Experiments on Positive Rays. 219 



origin* If the fields are not coterminous the lines will be 

 bent, and the bending will be complementary to that observed 

 with the secondaries of the type previously considered. 

 Thus, suppose the magnetic field overlaps the electric on the 

 camera side of the observation-tube, then the rays which 

 experience magnetic and not electric deflexions will be those 

 which have previously travelled in their charged state all 

 the electric and most of the magnetic field before reaching 

 the critical region ; they will therefore have suffered a 

 considerable deflexion, and the effect of the overlapping 

 magnetic field will show itself in the most deflected portion 

 of the rays, and not, as in the previous case, on the least 

 deflected portion. Thus the cu^es for these rays will have 

 the shape shown in the lower curve in fig. 4 a instead of that 

 of the upper one in the same figure, If the velocities of the 

 two kinds of rays were the same, we should have for each 

 primary two secondaries of the kind shown in fig\ 4 a. This 

 disposition of secondaries is shown on many of the plates ; 

 an example of it is given in fig. 12 (PL IX.) . 



This type of secondary, since it is produced by the com- 

 bination of a positively charged particle with a negative 

 corpuscle, might be expected to occur more readily with 

 slowly moving particles than with fast ones : a particle 

 moving faster than a certain velocity would not combine 

 with a negative corpuscle, so that there would be a superior 

 limit to the velocity of this type of secondary. There does 

 not seem, however, any reason why there should be an 

 inferior limit to the velocity, if the slow particles have re- 

 tained their charge up to the time of entering the electric 

 and magnetic fields ; and, as a matter of fact, this type of 

 secondary often (as in fig. 5) shows itself as the edge of a 

 patch of fogging on the plate rather than as a sharply defined 

 line. There are however cases, notably with the mercury 

 lines, where this type of secondary is more sharply defined 

 than we should expect, and where there are, in the primary 

 curves, particles with a smaller velocity than can be detected 

 in the secondaries of this type. 



In addition to the secondaries already mentioned, there are 

 on many plates, for example that reproduced in fig. 14, 

 prolongations of the primary which do not reach right up to 

 the vertical axis but stop after going halfway. These pro- 

 longations of the primaries are parabolic and are continua- 

 tions of the primary parabolas : they are therefore due to 

 particles which in the deflecting fields have the same value 

 of e/m as the particles which produce the primary parabolas. 



