CLASSICAL THEORY OF LIGHT 731 



also that the instruments efficient in this field are the most delicate 

 and accurate which have ever been made for any purpose; that they 

 may be used to measure ordinary lengths and other physical quantities 

 with an almost unbelievable precision; that the theory of relativity 

 sprang from an experiment performed with one, and the only known 

 way of measuring the diameter of a star involves the use of another. 

 Surely, if this topic is not interesting, nothing in physics is interesting. 

 Methods of measuring wave-length are sometimes divided into those 

 which operate by interference and those which utilize diffraction. 

 Though to a thorough insight the distinction is only trivial, at the out- 

 set it is convenient. In an extreme example of what is specially called 

 "diffraction," a single train of plane parallel waves is sifted through a 

 sieve in the form of a grid or a sequence of slits; and each element of 

 wave-front which passes through a slit evolves and spreads thence- 

 forward according to the law of wave-propagation. Eventually — as 

 a rule, in the focal plane of the lens beyond the grating — a region is 

 reached where the light from the several slits intermingles; and here 

 occur the variations in amplitude which disclose the waves and the 

 wave-lengths. For, as I have said earlier, the eye perceives only the 

 amplitude of the light-waves, and not their phase; therefore, in a 

 plane-parallel beam where the phase is perpetually changing but the 

 amplitude is everywhere the same, the eye receives a uniform impres- 

 sion, with nothing wavelike in it; and to make such a beam reveal 

 that it is undulatory, we must cause the amplitude to vary from point 

 to point. This is what we accomplish by breaking the beam into 

 fragments, or lacerating it with obstacles, preferably with an obstacle 

 having a periodic structure of its own, which is a grating. But it may 

 also be accomplished by causing two plane-parallel beams to intersect 

 one another under proper conditions. The region where they overlap is 

 then a region of varying amplitude — indeed, the variations are as 

 great as one can imagine; for if the beams are equally intense, 

 there is a succession of parallel planes of no vibration and darkness, 

 which separate spaces where there is vibration and light. The widths 

 of these spaces, the fringes, may be computed from the wave-length, 

 and reversely the wave-lengths from the widths, very simply and 

 without any knowledge of the law of wave-propagation beyond the 

 familiar expression for plane waves. Therefore this method of meas- 

 uring wave-lengths by causing interference of two parallel beams is 

 much the easiest to grasp; but it does not differ in principle from the 

 method involving a grating, for that acts by interference between the 

 beams from the various slits. 



