529 



he of interest, as this method of ilkimination was so usual now in 

 examining, say, living bacteria with high powers. This method 

 has been styled ultra-microscopic, but this is an implied contra- 

 diction ; all that dark-ground illumination does is to render 

 visible by contrast in illumination that is, very translucent 

 objects are rendered visible, and sometimes strongly visible, by 

 dark-ground illumination. His remarks were based on the Abbe 

 theory as applied to gratings, as the only other theory Airy's 

 spurious-disc theory had never been carried beyond Airy's 

 determination of the form of a star-image. Mr. Conrady 

 demonstrated that the full resolving power of an objective was 

 only obtainable when the dark-ground illuminator had three 

 times the numerical aperture of the objective, that otherwise the 

 resolving power was equal to one-fourth that of the objective 

 plus one -fourth that of the condenser. No higher resolving 

 power can be obtained with dark-ground illumination than will 

 "be given with an objective having a numerical aperture of 0*47. 

 {In practice, 0*6 was found to be the limit.) Particular stress 

 was laid on the necessity for the dark-ground stop to be but little 

 in excess of the absolutely necessary size. 



Mr. W. B. Stokes (Hon. Sec.) said that Mr. Conrady had not 

 given them any practical results, an omission which he himself 

 was fortunately able to fill. It seemed to him that Mr. Conrady's 

 assumptions were not such as were generally accepted. Dark- 

 ground illumination was used to display objects which were 

 thought to refract light fairly strongly, and it was mainly by the 

 light refracted by the details of the object that the image was 

 formed. Silicious fibrils, as in diatoms, bacteria, scratches 

 {rulings) on glass, etc., deflected the course of the light which 

 pas>ed through them, and it would be difficult to predict the 

 position of spectra formed by this light. The object appears as 

 if self-luminous, even in the case of rulings on glass, when the 

 individual lines shine and the clear interspaces are dark. Mr. 

 Conrady's assumption was that the light was not refracted, and 

 his ideal object would probably be a grating with opaque bars 

 and clear interspaces, an object which, on dark ground, would 

 give a very disappointing image. It would not be surprising, 

 therefore, if Mr. Conrady's theoretical results did not agree in 

 practice. Mr. Stokes added that, finding his own- experiments 

 did not agree with the spectrum theory, he had asked Mr. Nelson 



