4 Transactions of the Society. 



conventional limit of resolution, the intensity curves are as shown 

 in fig. 1, the dotted curve being the resultant intensity of the 

 overlapping images, each of which has an intensity given by 

 the continuous lines. It will be seen that the intensity in the 

 middle is less than the adjacent maxima by about one-third. 

 Such a proximity of images will give a well-marked band of 

 separation of the images, yet in this case the distance between 

 the maxima is only equal to the radius of the first dark ring. 

 Now a rigorous calculation shows that Mr. Nelson's results cor- 

 respond to a closeness for which the depression of intensity in the 

 middle will have rather more than disappeared. Hence, putting 



Fig. l. 



aside errors, which may amount to about 5 p.c, we may say that 

 Mr. Nelson succeeds in seeing the depression of intensity in the 

 middle until it completely disappears. Mr. Nelson is to be con- 

 gratulated in this achievement : it denotes quite exceptionally 

 keen vision ; but he is totally mistaken in thinking that from the 

 distance between his two stars in this case he can calculate the 

 radius of the first diffraction ring. 



Now in regard to this extremely successful resolution. The 

 usually accepted limit was never intended to be the ultimate 

 value ; it was fixed in a purely conventional way, so as to provide 

 a standard (in terms of which different openings and instruments 



