184 PROCEEDINGS OF THE ACADEMY OF [March, 



will extend from close to the centre of objective to beyond its margin, 

 and correspond in all essential respects to those present in a fully 

 illuminated objective, as such widely diffracted beams must be derived 

 from dioptric rays which pass through that portion of the objective's 

 aperture diametrically opposite to them. It is true these beams 

 would be expanded if the full aperture was illuminated, but the expan- 

 sion would be principally outside the margin of objective and, as far 

 as present experiment is concerned, be immaterial. The following 

 results demonstrate that three such spectra as we are considering 

 will produce an image identical with that resulting from the whole six, 

 and the conditions may be summed up as follows : 



Through the right-hand tube of the binocular, the image will be 

 produced by a full dioptric beam, supplemented by diffracted beams 

 corresponding to those resulting from a small central illuminating cone, 

 while through the left-hand tube it is derived from diffracted beams 

 alone, corresponding to those present with a full cone of illumination. 

 On examination, it will be found that both images are fairly well 

 defined, but that the resolution of the fine structure is noticeably 

 sharper and more distinct in the diffraction image through the left 

 tube. 



It will also be noted that the diffraction image is blue in color, and 

 before going further it will be well to fully understand how the colors 

 of such images are to be accounted for. The diffracted beams seen 

 at back of objective correspond to any other diffraction spectra and 

 include light of such wave lengths as enter into their formation. The 

 images resulting from their recombination will accordingly have the 

 same color as the light supplied to the microscope, as modified by 

 absorption in passing through the object, provided complete spectra 

 are included within the aperture of the objective; but should part of 

 the spectra be cut off at its margin, the diffraction images will corre- 

 spond to color sensation produced on the retina by light included in 

 that portion of the spectrum admitted, and will accordingly depend 

 on two factors: the distance between the elements of the object and 

 the aperture of the objective. Under the conditions outlined in pres- 

 ent experiment it has already been noted that the diffraction image of 

 P. angulation is blue, which is due to the red ends of its spectra being- 

 cut off at margin of the objective's aperture. If Navicula Lewisiana 

 or a coarse N. rhomboides be substituted only the violet end of the spec- 

 tra will be admitted and the unresolved image be similarly colored. 

 On the other hand, PI. formosum, whose spectra will be completely 

 admitted, will appear just as white in the diffraction as in the directly 



