a 



h 



c 



Theories of Microscopical Vision. By A. E. Conrad y. 623 



tig. 96, c ; that is to say, the bright line becomes much brighter, but 

 also narrower, exactly as it ought to do, and this effect will be 

 the more pronounced the narrower the actual slits are, because the 

 second spectrum will be relatively brighter accordingly. There 

 will, however, be a tendency to the formation of a feeble false line 

 midway between the true ones. Generally, these ghosts will be 

 invisible, but I believe they have 

 been "glimpsed." 



(b) When the slit really is equal 

 in width to the dark interval, the 

 second spectrum disappears, and 

 the result produced by the direct 

 light and the first spectrum is left 

 unchanged ; and so it ought to be, 

 for we saw that the image pro- 

 duced by these alone is just right 

 in this case. 



(c) When the slits are wider 

 than the dark interval, we saw that 

 the second spectrum is opposed in 

 phase to the direct light, hence we 

 now get an intensity-curve like 

 fig. 96, d, from the combination of 

 direct light and second spectrum, 

 and combining this again with the 

 effect produced by the direct light 

 and the first spectrum, we get a 

 curve like fig. 96, e — i.e. the lines 

 are diminished in brightness, but 

 broadened out ! 



I think it will be unneces- 

 sary for the present to carry this 

 analysis further to spectra of higher 

 order ; it must already be obvious 

 that this improvement in the veri- 

 similitude of the image wilb be 



maintained solely through those phase-relations — which I believe 

 I am the first to point out, at any rate in their bearing on the 

 formation of the microscopical image. 



It will also be unnecessary to deal at any length with the case 

 of direct light of less obliquity, or perchance proceeding along 

 the optical axis ; we shall then have some or all of the successive 

 spectra which can enter the objective present in pairs ; but it is 

 easy to see that, owing to the complete symmetry of the object 

 we are considering, the other spectrum of any one order can only 

 emphasise the effect produced by its fellow. The main facts will 

 remain as previously discussed for extremely oblique light. 



d 



Fig. 96. 



