18,S2.] ^41 [Eddy. 



domain of tliermodjaiamics, let me disclaim such an interpretation en- 

 tirely, and say that my only wish is to add, if possible, to the exactness and 

 completeness of those theories, which are among the most important of 

 modern physics. 



Cincinnati, April 22d, 1882. 



[Note. — Professor Willard Gibbs has suggested to me that we are not 

 at liberty to assume that reflections or radiations taking i^lace at moving 

 surfaces, follow the same laws as from surfaces at rest ; and that a perfect 

 reflector moving in a medium through which luminous waves are being 

 propagated, may suffer a resistance which would require the expenditure 

 of as much energy as could be obtained by the proposed process. 

 Admitting for the moment the justness of these observations respecting 

 reflections and radiations from moving surfaces, I shall hope to show in 

 the first place that the syren may be so adjusted that no such resistance 

 need be encountered, and in the second place that it is possible so to 

 modify the syren that no reflections or radiations need take place from 

 moving surfaces. 



In the discussion of the first point, let us consider the case of a ray 

 fixUing perpendicularly upon a perfect reflector. The only numerical 

 magnitudes susceptible of variation in this radiation are its wave length 

 and amplitude, the velocity being assumed constant and dependent upon 

 the elasticity of the medium. When the reflector moves in its own plane 

 at right angles to the ray, it cannot, apparently, be seriously urged that 

 the reflected ray will have either its wave length or its amplitude changed 

 by the reflection. For, so far as can be seen, the wave length would suf- 

 fer a change and be shortened only by giving the reflector a motion 

 forwards the approaching ray, thus crowding the waves together. Neither 

 would the amplitude be changed, for to do this would require the movin"- 

 plane to impart tangential impulses to the ether such as can be com- 

 pounded with the transverse motions already existing. If "such be the 

 tangential action of the moving plane on the ether, we should be led to 

 the apparently inadmissable result, that since a moving plane may impart 

 tangential impulses to the lumniferous ether, a disk rotating with suf- 

 ficient velocity in vacuo would become self-luminous. It would seem 

 but reasonable in our present imperfect knowledge of the subject to con- 

 clude that the only resistance which a perfect reflector experiences, while 

 moving against a ray, is normal to its surflice, and to be represented by a 

 normal pressure. Even if this view be not regarded as entirely correct, 

 it may nevertheless be confidently affirmed that the tangential must be 

 small compared with the normal resistance, just as the fractional resist- 

 ance of a gas is small compared with that arising from direct pressure 

 upon a body moving through it. Hence, it is seen, that in spite of friction 

 it is possible to make a ray turn a mill whose vanes are perfect reflectors 

 in the same manner as the Avind turns a windmill ; and the energy ex- 

 pended will in that case be withdrawn from the ray itself. 



