PHYSICS. 361 



The weigbt of the cubic inch of aether would be only 2,467,500,000,^,000,000,000 

 ounce. (Phil. Mag., September, 1880, V, xxii, 255.) 



J. W. Gibbs has given a resume of the results obtained by Kewcouib 

 (see Nature, May, June, 1886, xxxiv, 29, 170) and by Michelsou in their 

 experiments on the velocity of light. As the final result of his Wash- 

 ington experiments Newcomb gives 299,800 i 30 kilometers ])er second 

 as the velocity of light i?i vacuo. Michelson's Cleveland experiments 

 give 299,853 ±60; a result substantially identical. Combining the first 

 of these values with Nyren's value of the aberration constant (20 492"), 

 the value 149.60 is obtained as the sun's distance in millions of kilome- 

 ters. Although both the above experimentors paid especial attention to 

 the question whether there was any difference between the velocity of 

 r. d and of blue light, not the least indication of any difference was ob- 

 served. A difference of one-thousandth in these velocities would have 

 given a -well-marked color to the retnrn image of the slit iu Newcomb's 

 experiments. But no such effect could be detected. Michelsou covered 

 one-half the slit with a red glass; the two halves of the retnrn image 

 were exactly in line. Gibbs also gives the values which have been most 

 recently obtained of the ratio between the electro- magnetic and electro- 

 static units, which, according to Maxwell's electro- magnetic theory of 

 ligbt, represents the velocity of this agent. These values, as corrected 

 for the true value of the ohm, are as follows: Ayrton and Perry, 1878, 

 290.1; Hockiu, 1879, 296.9; Shida, 1880, 295.6; Exner, 1882, 291.7 (?); 

 J. J. Thomson, 1883, 296.3; Klemencic, 1884, 301.88 (?). They should 

 be compared with the velocity of light in air, expressed iu millions of 

 meters per second— 299.778 according to Newcomb. Setting aside Ex- 

 ner's and Klemencic's values, the other four are closely accordant, their 

 mean being nearly identical with that of J. J. Thomson, which appears 

 to be by far the most worthy of confidence, and differing by only one 

 per cent, from the velocity of light. Michelson's experiments on the ve- 

 locity of light in carbon disulphide afford, the reviewer thinks, an inter- 

 esting illustration of the difference between the velocity of waves and 

 the velocity of groups of waves — a subject to which liayleigh has called 

 attention in an appendix to the second volume of bis "Theory of Sound." 

 The quotient of the velocity in vacuo divided by tbe wave velocity in 

 carbon disulphide, calculated from Verdet, is for the line D 1.624 and 

 for E 1.637 ; while tbe group velocity, when used as a divisor, gives 

 1.722 and 1.767, respectively. Michelson's exi)erimental result was 

 1.76 ± .02, agreeing well with the latter. (Am. J. Sci., January, 1886, 

 in, XXXI, 62.) Subsequently Shuster has argued iu support of Ray- 

 leigb's conclusion that tbe velocity measured by tbe revolving- mirror 

 method of Foucault is really neither the wave velocity V nor the group 



velocity U, but is jj- I^i^t he shows that if only one revolving mirror 



be used, the experiinent can not be performed in the way mentioned. 

 Hence the Foucault method really measures neither V nor U, nor yet 



