MICROSCOPIC ILLUMINATION. 77 



fore even if the specimen is a periodic one (such as a ruled 

 grating or diatom) no visible interference fringes should be 

 formed in the upper focal plane of the objective. But it is an 

 essential feature of the Abbe theory to suppose that the whole 

 of the specimen is lit by coherent light (that is light between the 

 different parts of which a definite phase relationship exists) and 

 that interference occurs because of this fact. The ideals of 

 critical illumination, therefore, appear to have no justification 

 on the basis of the Abbe theory. 



With regard to the second question. Its ideals are unrealised, 

 for whereas with perfect critical illumination no interference 

 fringes (diffraction spectra) should be formed in the upper focal 

 plane of the objective, yet formed they undoubtedly are, as the 

 classic experiments of Abbe and Johnstone Stoney, and the con- 

 firmatory experiments of many other observers, suffice to show. 

 The following may be quoted : Abbe's experiments convinced 

 him that spectra produced by interference do exist in the upper 

 focal plane of the objective. Johnstone Stoney (3), in the case of 

 a spectroscope, under the most ideal conditions possible for the 

 realisation of critical illumination, proved that interference took 

 place to a degree that would have been impossible if critical 

 illumination had achieved its purpose. Its failure was due, it 

 appeared, to diffraction, since the optical system employed was 

 such that the aberrations usually found in microscopic apparatus 

 were almost completely eliminated. In reference to question 3 

 it will be observed that the ideals of critical illumination are 

 unrealisable, because even if the aberrations of the condenser 

 lens system are completely eliminated, diffraction (being due to 

 the nature of light itself) still remains, and is alone sufficient 

 (because of the representation of every point as a system of 

 rings) to destroy the realisation of point-f or- point illumination. 



In cases where the edge of a paraffin flame is used as a source 

 of light a further reason exists, as Conrady (6) has shown, why 

 critical illumination cannot achieve its ideals, namely, because 

 the depth of the illuminant prevents the simultaneous focusing 

 of all but one section of the flame. 



We must conclude therefore that critical illumination has no 

 theoretical advantage, and the question therefore arises as to 

 the cause of its continued employment. Two reasons may be 

 advanced ; firstly, it is (when properly applied) one method by 



