88 Mr. F. T. Trouton on the Influence of 



each oscillation, that is to say, with this edge at right angles 

 to the electric component. And, indeed, this is as we should 

 expect, since on comparison we see that in this case there is 

 less area which can act effectively as a source of reflexion 

 through the passage on it of current each oscillation. For 

 all the electricity of course does not go to the very edge, but 

 there is gradual accumulation of charge over the surface ap- 

 proaching the edge, increasing no doubt up to it. Conse- 

 quently the current diminishes in intensity towards the edges, 

 and thus it is only those parts of the reflector remote from 

 the charged edges which can be considered completely effec- 

 tive in producing the magnetic component of the reflected 

 beam. Each centimetre of charged edge, so to speak, reduces 

 the effective area of a reflector, so that with a rectangular 

 sheet strongest reflexion is obtained when the edge charged 

 is shortest. 



A very slight increase in the nodal distance is seen by the 

 table to occur when the reflector is somewhat less than half 

 a wave-length wide ; this may require further experiments of 

 a more sensitive character before being finally accepted. If 

 it be the case, it would seem as if there were really an accele- 

 ration of phase on reflexion from mirrors small in the mag- 

 netic direction also, but so slight as to be almost entirely 

 masked by the inward tendency due to divergency in the 

 reflected beam. If we look at it from a purely optical point 

 of view, Fresnel's theory indeed represents a linear reflector 

 as having an acceleration of phase of ^ it ; however, this is 

 omitting all considerations of direction in the vibration ; 

 but we have seen that the whole phenomenon requires a 

 treatment introducing a difference depending upon the mag- 

 netic and electric directions. An acceleration of J 7r might 

 suit very well in the magnetic case, but it is certainly insuf- 

 ficient when the reflector is narrow in the electric direction, 

 as is seen from Table I. With a quarter wave-length of 

 17 cm., the effect of a change in phase of ^ir on reflexion 

 would be to shift the node out to 21*25 cm., that is out 

 through -j^g the wave-length; but this is seen by the Table to 

 be much too little, especially so when we remember that the 

 observed minimum sparking-position must always be some- 

 what nearer the reflector on account of the divergency effect 

 than the true position of opposite phases. 



The result of diminishing the reflecting sheet simultaneously 

 in both directions was also examined, but though it was very 

 evident that the reduced dimension in each direction tended 

 towards possessing its own respective effect, no simple rela- 

 tion between them exhibited itself; thus with a reflector 



