SIX EQUIDISTANT SPECTRA AS A DIFFRACTION PROBLEM 73 



compute, which would give an exactly similar diffraction fan, but 

 abruptly broken off at the limit of the aperture. Theory shows that 

 a thread-shaped object which could yield such a particular diffraction 

 effect must (without considering other differences) be greater in 

 breadth than another one yielding the full continuous dissipation of 

 light ; in other words the actual thread, so inconceivably fine, 

 belonging to the Bacterium has produced a i diffraction effect ' 

 through the microscope, resulting in the appearance of a thread 

 which is the ' diffraction image.' But this latter is greater in width 

 than the actual thread or protoplasmic fibre would be could it be 

 seen directly without the aid of diffraction. 



(2) Whenever a portion of the total diffraction fan appertaining 

 to a given structure is lost, the image w T ill be more or less incomplete 

 and dissimilar from the object ; and in general the dissimilarity 

 will be the greater the smaller the fraction of light admitted. With 

 structures of every kind (regular and irregular) the image will lose 

 more and more the indications of the minuter details, as the peri- 

 pheral (more deflected) rays of the diffraction pencil are more and 

 more excluded. The image then becomes that of a different structure, 

 namely, of one the ivhole of whose diffracted beams would (if it 

 physically existed) be represented by the utilised diffracted beams of 

 the structure in question. 



At this place it is suitable to point out that Dr. Abbe em- 

 phasises to the present editor the importance of interpreting the 

 ' intercostal points ' shown by Mr. Stephenson in P. angulatum 

 (fig. 64) as not a revelation of real structure. ' The fact is that the 

 image, which is obtained by stopping off the direct beam, will be 

 more dissimilar from the real structure than the ordinary image. 

 It has already been shown that the directly transmitted ray is a 

 constituent and most essential part of the total diffraction pencil 

 appertaining to the structure ; it is the central maximum of this 

 pencil. If this be stopped off a greater part of the total diffraction 

 pencil is lost than otherwise, and the image, consequently, is a more in- 

 complete one, and therefore more dissimilar than the ordinary image. 



' The interest of the experiment in question is consequently confined 

 to two points, viz. 



i. ' It is an exemplification of the general proposition that the 

 same object affords different inages if different portions of the total 

 diffraction fan are admitted to the objective. 



ii. ' The image in question show^s to the observer what would be 

 the true aspect of that structure which will split up an incident beam 

 of light into six isolated maxima of second order of equal intensity, 

 suppressing totally the (central) maximum of the first order, as 

 fig. 65 ; a structure of such a particular and unusual diffraction effect 

 is theoretically possible, although it may be probably impossible to 

 realise it practically. Mr. Stephenson's experiment shows, in fact, 

 the true projection of the hypothetical structure. 



(3) ' As long as the elements of a structure are large multiples of 

 the wave-length of light, the breaking up of the rays by diffraction 

 is confined to smaller and smaller angles ; that is, all diffracted rays 

 of perceptible intensity will be comprised within a narrow cone 



