172 MR. GEORGE W. WALKER ON THE INITIAL ACCELERATED 



With No. 1 we get 



= 171xl0 7 x^ 3 ' 13 



and 



nir 



= 171x10'. 



The difference between the values of -, ' r from the two investigations is not 



(m + m!) 



perhaps very serious, hut the relative magnitudes of m and m' in the two cases is 

 more important. In the fii'st set m appears to be about \m' , while in the second set 

 m is about equal to m'. Some latitude must, however, be allowed in the value of the 

 constant K, in the first set and the ensuing values of k. The function f(k) is 

 extremely sensitive to changes of k, when k is approaching unity, and thus relatively 

 large changes in the calculated ratio of 111 to in' will be produced. More accurate 

 experiments are necessary to decide this point. 



We may, however, claim that formula No. 1 provides a substantial explanation of 

 KAUFMANN'S experiments, and assigns to the real mass of the particle a value 

 comparable with the electric mass. 



If we wish to hold the view that the mass of an electron is wholly electric we must 

 conclude that the particles in KAUFMANN'S experiments are not electrons, but are 

 either charged particles with a real constitutional mass, or electrons which have 

 become attached to gross matter. 



The analysis, on the other hand, is distinctly against ABRAHAM'S formula and 

 No. 2. 



We have no right to conclude that the particles are conductors, as it is still 

 probable that the assumption of perfect insulation would explain the experiments 

 (see Section 9). We may only claim that the assumption of perfect conductivity 

 does not disagree with the facts. 



We may, however, fairly argue from the experiments that condition (2) cannot be 

 maintained along with the view that the particles are conductors, while condition (1) 

 with this hypothesis adequately explain the facts. 



In forming a judgment of the results of this application of theory to experiment it 

 may be well to recall the concluding paragraph of Section 4. 



Since this analysis was made an investigation by BuOHKRER (' Phys. Zeit,' 1908, 

 p. 755) has appeared. He gives the results of experiments agreeing well with 

 LORENTZ' formula m'/(l F) 1/2 for a " contracted electron," but not in agreement with 

 ABRAHAM'S formula. I may say that KAUFMANN'S experiments also agree excellently 

 with LORENTZ' formula, just as they do with THOMSON'S formula, when a proportion 

 of ordinary mass is admitted. The reason is that both formulae contain an infinity of 

 the form l/(l F) 1 ' 2 . KAUFMANN'S results thus seem to me not inferior in accuracy 

 to those of BUCHERER. 



