150 



NOTES. 



An Experimental Proof of Phase-Reversal in Diffraction-Spectra. 

 By A. E. Conkady, F.R.A.S., F.R.M.S. 



In the paper which I read at the November Meeting on micro- 

 scopical theories, I supplied the mathematical proof of an important 

 theorem applicable to gratings, and showed how this theorem 

 explained the formation of correct images of gratings by means of 

 the light diffracted by them. 



As that proof, being mathematical, is not likely to appeal to the 

 majority of practical microscopists, and as some of these may have 

 been misled by the attack which has been directed against the 

 validity of that proof, I have much pleasure in describing and 

 showing a convincing experimental proof of the theorem in question. 



In my paper I showed theoretically that the second spectrum 

 was the lowest which was subject to phase-reversal, and I pointed 

 out in detail how the reversal occurred as soon as the width of the 

 slit exceeded the width of the dark interval, and how this reversal 

 of phase accounted for the relative width of the slits being correctly 

 shown as soon as the second spectrum was admitted. I moreover 

 pointed out that, owing to the preponderance of the direct light 

 and of the first spectrum, the slits were always shown in their 

 correct position, i.e. coincident with the ideal geometrical image, 

 and that the admission of the second spectrum merely altered the 

 width of the lines in the image in the proper direction. It should 

 be borne in mind that all these deductions were purely theoretical, 

 for I had not had an opportunity — and, indeed, had not sought for 

 one — to try the experiment. I wanted to prophesy from theory 

 what should happen. In deference to the leading idea of my 

 treatment of the Abbe theory, viz. to limit myself to normal v:ork- 

 ing conditio7is, I did not discuss what would happen if the direct 

 light were cut off, and the image formed by the first and second 

 spectrums only, as would occur if dark-ground illumination were 

 adopted. Now it is easy to extend the deductions in this direction. 

 I showed in the original paper that, with relatively narrow slits, the 

 direct light and the light of the first and of the second spectrum 

 would all arrive at the centre lines of the geometrical images of 

 the slits in the same phase ; hence it follows that the exclusion of 

 the direct light would leave the two spectra still meeting in equal 

 phases, and producing a maximum of brightness along the same 



