1215 



JC 



o 



-10 



-20 

 -30 

 -W 

 -50 

 -60 



show the same behaviour as to (he results for the distauce-correction. 

 Table VIII, mentioned above, seems, however, to point to a different 

 state of affairs. It shows not only that the relative density decreases 

 from the centre to ihe margin, but also that this decrease is larger 

 for [)lates with a large mean density than for those with a small 

 one. This phenomenon makes it desirable that, in the further i-esearch, 

 not all the [)lates should be treated in the same way ; and so I 

 divided them into groups, according to the mean density. Following 

 Henie I formed three groups; in the first group I took the 27 plates, 

 which showed a mean density less than 20 stars per square degree, 

 the second group was formed by 17 plates with a mean density 

 between 20 and 35, while the 11 lemaining plates, which had a 

 mean density exceeding 35, formed the third group. The change in 

 tlie relati\'e density from the centre towards the margin being 

 different for each of these three groups, they led to three different 

 curves showing the variation of the distance-correction. These curves 

 are shown in Pig. 1. The abscissae represent the distance from the 



Fig. 1. 



20 éO 60 80 W 120 MÜ 



20 w 60 80 m 120 m 





 -10 



-20 

 -30 



-w 



-50 

 -60 



-70 



-80 



-90 

 -100 



20 iO 60 80 100 120 m 



\-z-z 



1 II III 



centre, in milli-metres ; the ordinates give the difference between 

 the limiting magnitude at this distance and that at the centre, 

 expressed in hundredths of a magnitude. 



There is a striking difference between the curves I and III, and 

 it seemed important to investigate its cause. I have pointed out in 

 my first paper that such a difference may be due to two causes. 

 In the first place the colours of the stars might play a part. As 

 said before, the decrease of the limiting magnitude towards the 

 margin is an effect of spherical aberration, which, again, depends 

 on the refractive index. Now, since I he percentage of blue stars 



