182 PROCEEDINGS OF THE ACADEMY OF [March, 



the image seen through the ocular, but to remove the latter after 

 focussing, and modify the illumination until two spectra at right angles, 

 as well as the dioptric beam, are visible in the back of the objective. 

 Then on inserting the ocular the resolution should be apparent at once, 

 or by making any necessary spherical corrections by means of adjust- 

 ment collar or tube length ; but as long as both spectra are not seen, it 

 is utterly useless to endeavor to effect resolution by altering the adjust- 

 ment. If after satisfactory resolution be secured one of the spectra 

 be obscured by an intruding point, only lines at right angles to the other 

 spectrum will remain visible. 



Another well-known experiment is available with any objective 

 capable of resolving Pleurosigma angulatum with a very small central 

 cone, in which case the six characteristic diffraction spectra, free from 

 refracted rays, may be seen at margin of objective. A cover glass, 

 on which have been marked with India ink six dots that will cover 

 the diffraction spectra, may now be placed behind the objective, and 

 no trace of resolution will be visible; but if the cover glass be rotated 

 30° so as to bring the dots between the spectra, the resolution will be 

 as good as before. We here have a case where radically different re- 

 sults are obtained without changing the character of the screen behind 

 the objective, and any argument that it acts as a diffraction grating 

 is invalid, because if this were the case it should so act and produce 

 similar results in one position as well as the other. It is now evident 

 that there is some connection between these diffracted beams and image 

 formation in the microscope, and the question becomes whether the 

 formation of the image is dependent on their presence, or whether they 

 are merely accompaimng phenomena which happen to appear con- 

 comitantly with conditions which would permit of similar image for- 

 mation in their absence. 



That they are actually image forming rays can be readily proved in 

 several ways, the easiest being to throw the objective out of adjustment 

 sufficiently that the rays from its outer zone, which will be the diffracted 

 rays, are brought to a different focus from those of the dioptric beam 

 in the centre, and the two images can thus be examined successively 

 by a slight change in focus. A more convincing method is that of 

 Rheinberg, in which the rays from outer zone are refracted out of the 

 optic axis and the two images can be examined or photographed 

 simultaneously. Still a third method will shortly be referred to in 

 connection with the case of an objective illuminated by a solid cone of 

 light of its own aperture, which must now be considered. 



This is a condition we rarely meet in practical microscopy except 



