58 



jVA TURE 



[November i6, 1893 



surface as source of light ; the pencils may be separated 

 as far as desired ; its range of difference of path between 

 the interfering pencils is unlimited ; and when properly 

 adjusted the position of the interference fringes is per- 

 fectly definite, so that there is no uncertainty on account 

 of parallax, and no difficulty in counting the number of 

 fringes passing a given point. Finally, it may be added, 

 that this is probably the only form of instrument which 

 permits the use of white light (and consequently of the 

 identification of the fringes) in the determination of the 

 position or inclination of a surface without risk of distur- 

 bance due to contact or close approximation. 



As shown in Fig. 8, the refractometer consists essentially 

 of a plane parallel plate of optical glass Gj and two plane 

 mirrors M^ Mj. The beam of light to be examined falls 

 on the plate Gj at an angle, usually 45°, part being re- 

 flected and part transmitted.^ The reflected portion is 

 returned by ihe mirror M.,, and passes back through the in- 

 clined plate. The transmitted portion is returned hy the 

 mirror .M[,and is reflected by the inclined plate, and from 



Fic. 8. 



this point it coincides with the other beam, so that the two 

 are in condition to produce interference fringes.- 



A little consideration will show that this arrangement 

 is in all respects equivalent to an air-film or plate between 

 two plane surfaces. If the virtual distance between these 

 surfaces is small, white light may be employed, and inter- 

 ference fringes may be observed similar in all respects 

 to those between two plates of glass pressed nearly into 

 contact.^ 



' The front surface of the plate Gj is lightly coated with silver. The 

 light which leaves the refractometer is a maximum where the thickness of 

 liie silver film is such that the intensities of the transmitted and reflected 

 portions are equal. The silvering has another impjrtant advantage in 

 diminishing the relative intensity of the light reflected from the other sur- 

 face ; and lorlhis reason the thickness ol the film may be advantageously 

 increased, which permits also a more uniform surface. The ultimate ratioof 

 intensities of the two pencils is not afTeeted, for what is lost by transmission 

 on entering the plate is made up by reflection on leaving it. 



■-' One of the beams has to pass twice through the thickness of the glass 

 plate Gi, and in order to equalise the two paths, a similar plate G; is intro- 

 duced in the path of the other beam. 



1 If the plate Gi be not silvered, the colours follow the same order as those 

 of Newton's rings, but if the silvering be sufficiently heavy, the colours are 

 complementary ; this, if the plates Gi and Go are exactly equal and parallel. 

 Otherwise, the excess of path in-glass of one of the pencils disturbs the order 

 of colours by the effect of achromatism due to the dispersion of the glass, 

 as was first pointed out by Cornu. 



If, however, the distance exceeds a few wave-lengths, 

 monochromatic light must be employed. In this case the 

 fringes are in general invisible, unless they be viewed 

 through a small aperture. If, however, the two surfaces 

 are very accurately parallel, the fringes are always dis- 

 tinct, and it follows from the symmetry of the conditions 

 that they are concentric rings. Their diameters increase 

 as the square root of the order of the ring. 



These rings are not formed at the surface of the mirrors" 

 (as is the case when the distance between them is small), 

 but are perfectly distinct when the eye or the observing 

 telescope is focussed for parallel rays. 



In the preceding comparison between the refractometer 

 and the telescope, microscope, or spectroscope, the "ac- 

 curacy" has been increased at the expense of '"defini- 

 tion.'' When, however, the object viewed is beyond the 

 "limit of resolution ' of the instrument, its form and 

 distribution of light can no longer be inferred from that 

 of the image. Thus, if the object be a disc, a triangle, 

 or a double star, the appearance in the telescope is the 

 same. Similarly in the spectroscope, a source of great 

 complexity cannot be distinguished from one which pro- 

 duces a single spectral line. So that for such objects, 

 even in the ordinary sense of the word "definition," the 

 more familiar optical instruments cannot claim any ad- 

 vantage over the refractometer ; but if by "definition" 

 is meant not the actual resemblance of the image to the 

 object, but the accuracy with which the form or the dis- 

 tribution of light in a minute source may be inferred, 

 then it can be shown that all the advantage rests with 

 the refractometer. 



As an illustration of such an application of inter- 

 ference methods, let us consider the celebrated experi- 

 ment of Fizeau, in which Newton's rings are observed 

 with a sodium flame as source. The light, consisting of 

 two separate systems of radiations differing by about 

 one-thousandth in wave-length, each system produces its 

 own series of interference fringes. When the surfaces 

 are nearly in contact, the difference of path is very nearly 

 the same for both systems, and the fringes coincide, and 

 the clearness is a maximum. When, however, 

 the difference of path reaches about 500 waves 

 for one of the systems, it is a half wave more 

 for the other; and the maxima of intensity of 

 the one coincide with the minima of the other; 

 hence at this point the fringes are faintest. But 

 ^^-^ when the difference of path of the first system 

 ^-Va is about 1000 waves, it is a whole wave more 

 V ^ for the second, and the fringes coinciding, there 

 is again a maximum of distinctness. M . Fizeau 

 has counted 52 such periods, corresponding 

 roughly to a difference of path of 50,000 wave=. 

 Suppose, now, that this double line were so close that 

 it could not be resolved by the spectroscope ; then from 

 the evidence furnished by the variations in distinctness 

 of the interference fringes as the difference of path in- 

 creases, the duplicity of the line could be readily detected. 

 But beside this, it can be shown that the relative inten- 

 sities of the components, their distance apart, and even 

 the distribution of intensities within the component lines 

 can be inferred. 



Thus it has been shown {Philosophical Magaziiie for 

 September, 1892) that among some twenty radiations 

 wnich were examined (though all give simple lines in the 

 spectrum) the great majority are shown to be highly 

 complex. Thus, the red hydrogen line is a double 

 whose components have the intensity ratio 7 : 10, and 

 whose distance is about a fiftieth of the interval between 

 the sodium lines. Each component of the yellow sodium 

 lines is itself a double whose components are in the ratio 

 7 : 10, and whose distance is about one-hundredth of that 

 between the principal components. Thallium gives a 

 double line whose components are in the ratio i : 2, at a 

 distance of about a fiftieth of that of the sodium lines, 



NO. 1255, VOL. 49] 



