CONTEMPORARY ADVANCES IN PHYSICS 177 



answers. Foucault and Fizeau reported that she had repHed: the 

 former. But they had not heard distinctly ; for her actual response was : 

 tieither. 



Michelson's experiment however came after the idea of group- 

 velocity as distinp;uished from wave-velocity had been invented and 

 established. The refractive index of carbon bisulphide varies with 

 wave-length. On determining the wave-speed or phase-speed v from 

 the refractive index (by the equation N = c/v) and then the derivative 

 dv/d\, it is found '" that in the region of the visible spectrum, the term 

 \{dvld\) amounts to about seven per cent of the term v, on the right- 

 hand side of equation (20) — that is, the group-speed should be some 

 seven per cent lower than the wave-speed in carbon bisulphide. In 

 air, however, group-speed and phase-speed are sensibly the same. 

 The ratio of the group-speeds in air and CSo falls close to Michelson's 

 value." 



Coming as it did, therefore, the Michelson experiment merely showed 

 that those who had subtilized the Huyghens' theory by introducing 

 sine-waves had incidentally invented something able to move with the 

 measured speed of a light-flash, though nothing of the sort had been 

 available in the original form. Had it come earlier — well, there is no 

 way of knowing what would have been inferred; but people might have 

 come to think that after all a wavefront-theory or a corpuscle-theory 

 of light may have some use and value, even though the speeds assigned 

 to the waves or the corpuscles do not agree with those actually meas- 

 ured. Such an attitude of mind would be rather advantageous, 

 today. As a corollary for the present I submit: in picturing a jet of 

 free negative electricity as a beam of waves or a stream of corpuscles, 

 we should not be too confident that either the speed of the waves or 

 the speed of the corpuscles is the speed with which a segment dissected 

 from the jet would move from place to place, until someone succeeds in 

 making actual measurement of this last. Fundamental theory has 

 something to say on this point, which we will presently consider. 



'" I take all the numerical values in this section from a re\'ie\v of Michelson's 

 work by J. Willard Gibbs {Am. Jour. Sci. 31, pp. 62-64; 1886) which so far as I know 

 is the latest critical discussion of the data. 



'1 The problem is more complex than I have intimated, not only because Michelson 

 observed light covering a very wide range of wave-lengths so that i' and dvjdX both 

 extend over wide ranges of values, but also because different parts of a wave-front 

 are reflected from different parts of the mirror at different moments, and therefore 

 from differentl y-incl ined parts. Quite a controversy went on during the eighteen- 

 eightiesin the pages of " Nature" as to what it was that Foucault had really measured. 

 Rayleigh at first {Nature 24, p. 382; 1882) thought it was g; then changed his mind, 

 (25, p. 52; 1882) and decided it was t'-/g; then was convinced by Schuster {3i, pp. 

 439-440; 1886) that it was really i'V2(t' - g). J. W. Gibbs then took a hand {Zi, 

 p. 582; 1886) and contended that after all it was really g. The contro\"ersy seems to 

 have rested there. It may be added that Michelson's data eliminate v'-jg, but do not 

 quite discriminate between g and Schuster's expression. 



