THE RESOLVING POWER OF THE MICROSCOPE. 27 



Dr. Johnstone Stoney's experiment shows in the most conclusive 

 manner that, under the conditions in which the experiment is made, 

 — viz. with a narrow beam or cone of light — the one part of the 

 object markedly influences the image of another part. The next 

 question therefore is, can the effect be eliminated if we use a wide 

 cone and critical illumination 1 We have to test the point 

 experimentally to see whether under the altered conditions the 

 resolution of the two lines will outlast that of the twelve-line 

 band, or whether they will disappear together. I have not yet 

 been able to satisfy myself as to the result. Some preliminary 

 trials clearly showed the effect to be mucli less marked when 

 using a wide cone of light, but it was still to be perceived. 

 Evidence on this point from microscopists accustomed to work 

 with critical illumination would be valuable. 



In conclusion I would say that it is the usefulness of the results 

 to which a careful study of the newly demonstrated fact may 

 lead that must be my apology for having detained you so long 

 in discussing and explaining an intrinsically trivial matter ; it 

 will be borne in mind, however, that it is only by close attention 

 to little things like this that the complicated questions connected 

 with microscopic vision can be brought a step forward. 



I am sure it will interest you if I quote a few words received 

 from Dr. Johnstone Stoney this morning. I had sent him the 

 MS. of my paper to make sure that my statements regarding 

 himself were correct, and he says : — 



" I have suggested two corrections as to dates. It was in the 

 end of 1893 or in 1894 that I discovered the new resolution of 

 light into undulations of flat wavelets, and one of the first 

 results I got out by it was that a pair of lines, or a pair of 

 dots, should be resolved by an objective of less aperture than 

 that required for the resolution of a ruling of lines or a row of 

 dots equally spaced, and that, when so resolved, the spacing of 

 the pair of lines should appear to be greater than that of the 

 ruling. This I got out by the theorem in reversal represented 

 graphically by your Fig. 6. 



" So far as I know, until my application of this principle of 

 reversal, aided by the new method of resolving light, to such 



