{ 244 ) 



Thus we Jire led to considei' the elements of / as the qiuiiilities wliicli 

 determine the tensions in the medium. In the electrostatic and iji the 

 purely mag-netic field these expi-essions agree with those given by 

 Maxweij.. l^nt in the general case they difïei' from those values. In 

 a plane wave for instance the tension in the direction of propagation 

 becomes zero. At first sight this may seem strange. For M.wwfjj, 

 deduced the existence of the pressure of the radiation from his value 

 of the tensions, and it appears that in putting /,,. =z [x being the 

 direction of propagation) we deny the existence of that pressure. 

 Yet this is not the case. For often we deduce the existence of the 

 pressure of radiation from the momentum of the electromagnetic 

 field without making use of the tensions. Properly speaking these 

 two explanations of the pressure are contradictory; or at least one 

 of them is superfluous. If both the tension and the momentum existed 

 in the medium, the effect of these two causes ongiit to be added 

 and we siiould iind the double value for the i)ressure. 



This difficulty does not exist if we ascribe the above values to' 

 the tensions. According to them a force exerted on a body is to be 

 ascribed either to a tension or to the momentum of the medium. 

 And if they both exist, their elfect must be added. liet us consider a 

 ray of light refiected on a |)erfect mirror. In the ray we do not 

 assume any longitudinal tension, bnt at the surface of the iniiTor the 

 normal component of © is zero, and our expression for the tension 

 is by no means zero, but coincides with that of Maxwell. The effect 

 of a ray of light on a mirror is therefore cpiite analogous to the 

 effect of a jet of water on a surface by which it is thrown back. 

 In the jet there need not be any pressure, but on the surface where 

 the water is thrown back, a pressure does exist. 



The tensions t, which we introduced, are therefore quite analogous 

 to ejastic tensions in bodies; the tensions of Maxwell on the other 

 hand are analogous to the absolute' tensions, as Lauk, calls them, i. e. 

 of those quantities whose divergence is equal to the change in 

 momentum of a stationary element of volume. This change is occasioned 

 bv two causes: 1*^ the tensions t, 2"^^ the transport of momentum 

 through the surfaces of the volume element. 



The result of our general considerations is this, that we — it is 

 Xyi^q — deny the existence of l)odies with a constant mass, and that 

 our assumptions differ in this respect from those of classical mechanics. 

 But on the other hand the law of conservation of energy warrants 

 that the total amount of mass is constant, so that the only 

 difference is that we assume that the mass can be transferred with 

 the energv from one body to another. Moi'eoxer we have reassumed 



