22 ON THE APPLICATION OF INTERFERENCE METHODS 



"resoh'iug power" of the method as compared with tliat of the grating. In 

 order that the comparison be quite fan-, however, it woiild be necessary to 

 take for a comparison spectrimi that of the substances here used, and imder 

 the same conditions. With the best iustnimeutal apphances now in use, it 

 is ihfficult to "resolve" lines as close together as the components of either of 

 the yellow sodium hues. It is evident, however, that by Light-wave Analysis 

 — if I may ventm-e so to call the foregoing method — a tenth of this distance 

 is obviously within the limit; indeed, if the width of the lines themselves be 

 less than their distance apai-t, there can be no limit. 



SUPPLEMENT. 



I, It has already been pointed out that in many cases it is difficult or 

 impossible to decide between two or more distributions of lines which give 

 veiy nearly the same visibihty curve; and when there are many lines in the 

 soiu-ce, the combinations of intensities and an-angoment of these, from which 

 a tyi)e may be selected, are enormously great. Indeed, even when the number 

 of hues is greater than thi-ee, excepting perhaps the cases where the hues 

 may be in pairs (as in the case of yellow sodium light), the residting visi- 

 bihty cun^e becomes so complex that it is very difficult to analyze. Doubt- 

 less ill many cases where the components are not too close, the grating 

 will give the infonnation necessary for the investigator to select the proper 

 combination. 



It may readily be shown that the fonnula 



'^ ~ pi ■> 



for the visihilifii ruivr duf to a distribution of light, // = 'f (r), is identical witli 

 that of the iuhiisifij curve at the fociis of a telescope provided with aper- 

 tures which produce this distribution in the light passing through. Accord- 

 ingly, if a telescope be prox-ided with a])ertures adjustable in width or length 

 and distance apart, the diffraction image of a distant illuminated slit will 

 give at once a representation of the whole ^^sibility cui'\'e; and by adjust- 

 ment of intensities and distances, any particular ^'isibihty curve may be more 



