Notes. 389 



arrange this so that the contact edges of the two halves of the com- 

 pensator just come between the zero maximum (or dioptric beam) 

 and the maximum of the first order (or first spectrum) on the one 

 side. We next cover up all the maxima except these two. Then, 

 looking through the Microscope, the image of the grating is seen, 

 and as we turn the micrometer-head of the compensator, thereby 

 altering the phase of the one maximum, the lines in the image 

 travel across the field. 



It is rather important that, under the conditions of this experi- 

 ment, with the admission of two contiguous spectra only, the actual 

 practical extent of the movement of any individual line is confined 

 to about twice its own width. This can be shown by studying the 

 effect of an isolated bright line, and it is desirable to bring the fact 

 out, because apparently any particular line of the grating moves 

 right across the field. 



But in this effect we have nothing more than a rather remark- 

 able optical illusion, by which the analogy between the shifting of 

 the lines when the grating is moved bodily, and the shifting when 

 the phase-difference is introduced by a compensator, becomes greater 

 than is actually the case. 



We can best understand this by regarding the image of any line 

 as the bright visible portion of a diffraction pattern. The pattern 

 may be represented by an intensity curve, and this curve would 

 show t hat besides the chief bright line there is a narrow bright line 

 on either side — almost unobservable under ordinary circumstances. 



Now when the grating is shifted on the stage, this intensity 

 curve representing the image is shifted "about bodily in the view 

 plane, i.e. the region in the view plane where the maximum occurs 

 may move from one portion of the field to another, whilst the con- 

 figuration of the diffraction pattern remains unaltered. When, 

 however, a phase difference is introduced by the compensator, there 

 is no change of region in the view plane where the visible portion 

 of the diffraction pattern is produced, but the configuration of the 

 intensity curve itself undergoes alteration. The narrow faint line 

 shifts along, becoming broader and brighter ; the central bright line 

 shifts along, becoming narrower and less bright, and a certain position 

 is arrived at where both lines are equally bright and equally 

 broad. This is the state of things when half a phase difference is 



wedges. Now think of these two halves being laid alongside of one another again, 

 with their edges in contact, so that the square slab is re-formed. It is evident that 

 the glass is of equal thickness in the two halves in any diameter drawn at riglit 

 angles to their contact edge. But if the one glass be slid against the other, then in 

 all such diameters the glass on one side will be slightly thicker or thinner than the 

 glass on the other. 



In the compensator the one glass wedge is therefore made immovable, whilst to 

 the other a movement is imparted by a micrometer-screw sufficiently gradually that 

 we can readily alter the relative thickness of the two halves by fractions of a wave- 

 length of light, thereby obtaining our difference of phases. 



