DISCHARGE OF ELECTKRITY. 253 



that the iron retaiiis its maiinetic jn-operties under these very rapidly 

 alternating- forees. The result explains the etieet of the iron in stop- 

 ping the (liseharge. As I am not aware that any magnetie properties 

 of iron under such rai)idly alternating forces have been observed, I was 

 anxious to make quite sure that the <lifference between iron and brass 

 was not due solely to the differeiu'es between their specific resistances. 

 The first experiment I tried with this object was to cover the iron rod 

 with thin sheet platinum, such as is used for Grove cells. As the re- 

 sistance of phxtinum is not very different from that of iron, if the effect 

 depended merely upon the resistance, slipping a thin tube of platinum 

 over the iron ought to make very little diii'erence. I found however 

 that when tlie ])latinum was placed over the iron, all the])eculiar effects 

 produced by the latter were absent, thus showing that the effect is not 

 due to the resistance of the iron. It then occurred to me that 1 might 

 test the same thing in another way by magnetizing the iron to satura- 

 tion, for in this state // is nearly unity; thus if the result depended 

 mainly on the magneti(i properties of the iron it ought to diminish 

 when the latter is strongly magnetized. I accordingly tried an experi- 

 ment in which the iron in the coil B was placed between the poles of a 

 l)Owerful electro-magnet. When the magnet was "off" the iron almost 

 stopped the discliarge in A; when it was "on" the discharge became 

 brighter, not indeed so bright as if the iron were away altogether, but 

 still unmistakably l)righter than when it Avas unmagnetized. This ex- 

 periment, I tliink, i)roves that iron retains its magnetic projjerties 

 when exposed to these rapidly alternating forces. 



Another result worthy of remark is that though a brass rod or tube 

 inserted in B does not stop the discharge in A, yet if a piece of glass 

 tubing of tlie same dimensions is coated with Dutch metal, or if it has 

 a thin film of silver deposited upon it, it will stop the discharge very 

 decidedly. We are thus led to the somewhat unexpected result that a 

 thin layer of metal when ex[)osed to these very rapid electrical vibra- 

 tions may absorb more heat than a thick one. I find, on calculating 

 the heating effect in slabs of various th.icknesscs, that there is a thick- 

 ness for which the heat absorbed is a maximum. - - 



The slight iii(;rease in tlie brigiitiiess of the discharge in A when a, 

 brass rod is placed in 1> is due, I tiiiuJi, to the dimiiuilioii in the self- 

 induction in the i)rimary <;ircuit i>rodiu-ed by this rod whose condnc- 

 tivity is so good that it absorbs practically no heat. 



We will now return tothe<;aseof bad conductors, where nn is small; 

 here the absor])tion of eneigy is proi)orti(»nal to the conductivity, and 

 we might use this method to compare the conductivity of electrolytes 

 for very rapidly alternating currents. I tried a few experiments of 

 this kind and found, as 1 did in the experiments described in the Fro- 

 ceedhifjs of the Roi/al iSoeiety, XLV, p. 200, that the ratio of the conduc- 

 tivities of two electrolytes was the same for ra]»idly alternating as for 

 steady currents. I was anxious, however, to see whether these rai)idly 



