Fizeaus Experiment and the uffither. 215 



and not by v(l — 1//U- 2 ), the numerical value of the new term 

 being much smaller than the numerical value of either 

 expression for the drift. The experiment was repeated with 

 extreme care and accuracy by Zeeman in 1915, and not only 

 was the existence of the new term verified, but also its 

 variation with the wave-length. 



It has hitherto been assumed, that the Lorentz expression 

 for the aether-drift can be derived only on the basis of the 

 theory of electrons, or from the equations of the theory of 

 Helativity. I desire here to call attention to the fact, that 

 it follows much more simply on the basis of the older 

 elastic theories of light. 



§ 2. Let us suppose that 



. »g= B 8 w 



represents the propagation of a light-wave in the aether, 

 y denoting the displacement of the aether, p its density and 

 E its elasticity; p is constant throughout all space. Let us 

 suppose that inside matter there are particles attached to 

 points in the aether, the mass of the distribution per unit volume 

 being a. These particles may be regarded as constituting a 

 kind of " virtual " aether. They execute vibrations about 

 the points to which they are attached. If rj denotes their 

 average displacement, their motion is given by 



dS 



?+%-y) = 0. (2) 



The force attaching them to the aether reacts on the latter, 

 so that instead of (1) we obtain 



pgMg^-^) (3) 



Add (2) and (3), and we obtain 



PW+ a dt 2 ~ E W (4) 



Assume now that y and rj vary as cos nt. Then (2) gives 

 and (4) becomes 



Q2 



