30 THE PRESIDENT'S ADDRESS. 



and with white light (line E, the same as used in R. M. S. 

 tables), will be 



96,410 according to Abbe theory. 



88,450 „ Rayleigh experiment. 



79,044 „ Airy theory. 



Now the results from experiments both with full and 3/4 

 cones go largely to corroborate Mr. "Wright's conclusions. 

 It is common knowledge that when a full cone is em- 

 ployed the resolving power falls off, and it has been cus- 

 tomary to account for this falling off in the resolving power by 

 the outstanding spherical aberration in the objective. To test 

 the accuracy of this current notion a critical image was set up, 

 and matters arranged so that access could be obtained to the 

 back lens of the objective without disturbing any of the adjust- 

 ments. When a full cone of light was used the resolving 

 power fell off, and when a 3/4 cone was employed it was 

 as usual restored again; a stop was then placed at the back 

 lens, cutting off the peripheral unilluminated annulus. We 

 had, therefore, an objective of less aperture, but illuminated by 

 a full cone. Under these circumstances one would have 

 expected to see a critical image, but not so, and this is the 

 crucial point. In order to obtain the maximum resolving 

 power for that reduced aperture the illuminating cone had to 

 be reduced until only three-quarters of the back lens was illu- 

 minated. This is a most important fact, because it shows that 

 spherical aberration is not playing the role commonly assigned 

 to it, and the blotting out of structure has a deeper meaning. 

 Reading this in the light of the new theory we see that when a 

 full cone is used the image comes under the physical or Airy 

 limit, but the moment we use a 3/4 cone we have diffraction 

 spectra in the peripheral annulus. The picture therefore 

 obeys the Abbe limit with its greater resolving power. 



Probably spherical aberration is present as well, and pro- 

 duces a certain amount of indistinctness of image which helps 

 to obliterate the fine detail, but the above experiment proves 

 that spherical aberration does not account for the whole pheno- 

 menon as it was previously thought to do. With regard to the 

 3/4 cone illumination, it should be remembered that the areas 

 of circles are in the proportion of the squares of their diameters ; 

 therefore the area of the peripheral annulus where the spectra 



