548 



Transactions of the Society. 



section of lines b and lines d will be brought out as bright dots ; 

 reference to fig. 125 shows that these again exactly correspond to the 

 true position of the actual perforations ; this new combination of 

 maxima, therefore, further improves the image. 



3. Finally, we may have four maxima such as direct light and 

 spectra, A lf Bj and D x admitted simultaneously — fig. 127e. 



We again obtain very bright dots at the points of intersection 

 of lines a, b and d in fig. 3 — i.e. in the position of the actual dots. 



This small increase of aperture is, therefore, sufficient to 

 emphasise the bright dots to such an extent as to render them 

 unmistakable. 



The qualitative method of discussing the results to which I am 



Fig. 128. 



at present limiting myself is not adapted to bring out all pecu- 

 liarities of the different partial images, such as the secondary 

 maxima (intercostals) resulting with combinations of maxima like 

 figs. 127c, 127d, and 127e, nor to show that these secondary effects 

 are different with the different combinations, and are more or less 

 neutralised when they are all superposed. These finer points must 

 be left for a rigorous quantitative treatment of a few judiciously 

 selected cases which I hope to bring forward on a future occasion. 

 What has been stated above will suffice for the present to show 

 that with a perforation pattern also there must be rapid improve- 

 ment of the image with increasing aperture. 



