396 Prof. J. H. Poynting on 



pressing out sideways — that is, in the direction of propagation. 

 They press against the source from which they issue, against 

 each other as they travel, and against any surface upon which 

 they fall. Or we may take Professor J. J. Thomson's point 

 of view *. " Let us suppose that the reflecting surface is 

 metallic ; then, when the light falls on the surface, the varia- 

 tion of the magnetic force induces currents in the metal, and 

 these currents produce opposite effects to the incident light, 

 so that the inductive force is screened off from the interior of 

 the metal plate : thus the currents in the plate, and therefore 

 the intensity of the light, rapidly diminish as we recede from the 

 surface of the plate. The currents in the plate are accompanied 

 by magnetic force at right angles to them ; the corresponding 

 mechanical force is at right angles both to the current and 

 the magnetic force, and therefore parallel to the direction of 

 propagation of the light." In fact, we have in the surface 

 of the reflector a thin current-sheet in a transverse magnetic 

 field, and the ordinary electrodynamic force on the conductor 

 accounts for the pressure. 



In sound-waves there is at a reflecting surface a node — a 

 point of no motion, but of varying pressure. If the variation 

 of pressure from the undisturbed value w r ere exactly propor- 

 tional to the displacement of a parallel layer near the surface, 

 and if the displacement were exactly harmonic, then the 

 average pressure would be equal to the normal undisturbed 

 value. But consider a layer of air quite close to the surface. 

 If it moves up a distance y towards the surface, the pressure 

 is increased. If it moves an equal distance y away from the 

 surface, the pressure is decreased, but by a slightly smaller 

 quantity. To illustrate this, take an extreme case, and for 

 simplicity suppose that Boyle's law holds. If the layer 

 advances half way towards the reflecting surface, the pressure 

 is doubled. If it moves an equal distance outwards from its 

 original position, the pressure falls, but only by one-third of 

 its original value ; and if we could suppose the layer to be 

 moving harmonically, it is obvious that the mean of the 

 increased and diminished pressures would be largely in excess 

 of the normal value. Though we are not entitled to assume 

 the existence of harmonic vibrations when we take into 

 account the second order of small quantities, yet this illustra- 

 tion gives the right idea. The excess of pressure in the 

 compression half is greater than its defect during the extension 

 halt, and the net result is an average excess of pressure — a 

 quantity itself of the second order — on the reflecting surface. 



* Maxwell's * Electricity and Magnetism/ 3rd edition, vol. ii.p. 441, 

 footnote. 



