Section III., 1908. [3] Trains. R. S. C. 



I. — Note on an Improv&men4 in the Method of Determining of 

 Visibility Curves. 



" By C. S. Weight, B.A., 



1851 Exhibition Scholar, 1908, University of Toronto. 

 (Communicated by Prof. J. C. McLennan, and read May 26, 1908.) 



One of the chief objects of Michelson in designing the interfero- 

 meter which bears his name, was to analyse the source of light, and 

 determine, if possible, the distribution of intensity in that source. This 

 is arrived at by estimating the visibility of the fringes formed by the 

 interference of the two beams of light in the interferometer for various 

 differences in path of the two rays. 



The method of observation consists simply in moving back one 

 interferometer plate until the difference in path A between the two 

 interfering beams is such as to give a visibility of one-half the maxi- 

 mum — the latter being obtained for A = C. The half-width of the 

 spectral line used as the source of illumination is then given by 

 0.22 r 

 A 



While determining the visibility curves for the spectral lines of 

 some of the elements it occurred to the writer that if some standard 

 could be used giving fringes of a known visibility, and in the same 

 position as those due to the interferometer, the method of determin- 

 ing the half-width would reduce to a very simple photometric deter- 

 mination. 



As the standard a mica plate was found very convenient. This plate 

 was placed in the path of the light entering the interferometer and 

 inclined at such an angle as to give straight line fringes perpendicular 

 to those given by the interferometer and in approximately the same 

 focal plane. A simple adjustment of the angle of the mica plate was 

 all that was necessary to alter the width of the mica fringes or the angle 

 of intersection of the two sets so as to make them similar in every 

 respect. 



It was found in practice that when the fringes formed a system of 

 squares the points of equal intensity could be determined with consid- 

 erable exactitude. 



For absolute measurement of the half-width of the spectral line 

 this standard would require to give fringes of a visibility one-half 

 the maximum. But if comparative measurements only were required, 



