134 SUMMARY or CURRENT RESEARCHES RELATING TO 



existence of any fact is attested, viz. the medium of some one of tlie 

 senses. But may it not be true that we have not yet reached the 

 fulfilment of the conditions necessary to visibility? It certainly 

 cannot yet be safely asserted that it is impossible to see a material 

 particle which has, in one direction, a magnitude not exceeding 

 1/500,000 in. Photography offers the evidence, somewhat negative 

 in its character, that the limit of visibility is reached with lines having 

 a width of about 1/200,000 of an inch. Lines of this width are 

 the finest that have ever been photographed. But the most conclusive 

 (evidence against the certainty of being able to produce lines as fine 

 as 500,000 to the inch consists in the fact, repeatedly proven in my 

 own experience, that lines which appear to be excessively fine often 

 have a real width two or three times as great as they appear to have, 

 as has been proved conclusively by filling the lines with graphite, 

 which brings out the real limit. This phenomenon will come up 

 again in connection with the subject of resolution. 



I have already stated my belief that the limit of resolution has 

 been so nearly reached that, though it is quite possible under a com- 

 bination of favourable circumstances to obtain a resolution a little 

 beyond 113,000 to the inch, the uncertainty which must always attend 

 observations of this character is so great that the certainty of resolu- 

 tion cannot be safely asserted. In consideration of this uncertainty, 

 and of the fact that so little progress has been made in resolution 

 compared with the recent advance in the construction of objectives, I 

 beg to propose as a test the visibility of single-ruled lines in place of 

 the resolution of these lines in close combination. Instead of bands 

 of lines of the Nobert pattern, I propose a series of bands, each having 

 the same interlinear unit, but with the lines of each successive band 

 finer than those of the preceding band. The space between the lines 

 should not be so great as to interfere with their easy detection, nor 

 so small as to require any effort in resolution. One micron (/x) is a 

 convenient unit. A heavy line should precede the band, in order to 

 facilitate finding it. 



According to my own experience there are four facts which 

 must always throw grave doubt upon any reported case of difficult 

 resolution : — 



1. It is well known that by the manipulation of the light, every 

 other condition remaining the same, it is possible to vary the apparent 

 number of lines in a given band of coarse rulings. Can any one offer 

 a reason why there shotxld not be the same difference with bands of 

 fine lines closely ruled ? 



2. I have many times ruled bands of lines with the interlinear 

 spaces distinctly marked, but in which each line was in reality con- 

 siderably wider than the space between the lines, as I have proved 

 by extending single lines beyond the others and filling them with 

 graphite. The only explanation of this singular fact which I can 

 suggest is that the diamond may possibly cut square down at one 

 edge of the line and for the remainder of the line produce only an 

 abrasion of the surface of the glass, which is so slight as not to 

 interfere with throwing up a furrow upon the remaining portion. 



