Theories of Microscopical Vision. By A. E. Conrady. 547 



unsatisfactory, inasmuch as it discloses only a part of what is 

 capable of being shown by the object-glass. And this leads us to 

 the discovery of an important advantage to be derived from the use 

 of an extended source of light — or, in other words, of a large cone 

 of illumination. For in that manner we can obtain the effects 

 shown in fig. 127a and fig. 127b simultaneously and superposed, 

 leading to the formation of an image showing bright lines corre- 

 sponding to a and b of fig. 125, and with the points of intersection, 

 which will be noted to correspond to the actual dots, specially 

 bright as the light of both systems of lines is there added together. 

 The simple expedient of using a wide cone of illumination, being 

 equivalent to oblique light in all directions, has, therefore, at once 

 produced a tolerably good indication of the actual nature of the 

 object. We can derive yet another lesson from this observation. 

 -On inspecting figs. 127a and 127b it will be seen that direct light 

 in the central part of the aperture is useless for the purpose of 

 showing any structure, because no corresponding diffraction- 

 spectrum can enter through the available aperture of the object- 

 glass. Such light can, therefore, only form a general bright 

 Illumination of the field ; cutting it off by a central stop, and thus 

 producing annular illumination, must improve the clearness of 

 the image, and this would appear to be a perfectly legitimate 

 means of attaining the utmost distinctness in the image of struc- 

 tures close to the limit of resolution of an object-glass. 



We proceed to study the effect of an increase of aperture. 



No new spectra can enter unless the aperture is at least equal 

 to the diagonal of the squares into which the spectra of fig. 125a 

 arrange themselves. When that aperture is slightly exceeded, we 

 have the possibility of three distinct combinations of maxima 

 which can enter the increased aperture, viz. : — 



1. In accordance with fig. 127c we can have a beam of direct 

 light, and the two diffracted beams A x and B x derived from it, We 

 have thus three separate beams capable of interfering with each 

 other. The direct light and Ai alone would meet in equal phase 

 and produce bright lines corresponding to a in fig. 125 ; the direct 

 light and B x would similarly produce lines like b in fig. 125. When 

 all three are admitted at once, then they will all meet in the same 

 phase and produce a very pronounced maximum of brightness at 

 the points of intersection of lines a and lines b in fig. 125 ; in other 

 words, these three maxima lead to the formation of the correct dot 

 pattern. When added to the crossed-line effect— with enhanced 

 points of intersection— resulting from the combinations illustrated 

 in figs. 127a and 127b, they will further accentuate the dots, and 

 thus improve the verisimilitude of the image. 



2. We may have groups like that in fig. 127d— i.e. the direct 

 light, the spectrum, B x (or Ax), and one of the remoter spectra, D x 

 (or Cj). By similar reasoning we find that the points of inter- 



