474 Scientific Proceedings, Royal Dublin Society. 



a steady thinning-out of the star-system sets in as early as the 

 average distance of a star of the fourth magnitude ; but I do not 

 see on what other principle the results embraced in the foregoing 

 Table can be explained unless it be the absorption of light by some 

 interstellar medium. In the case of a star of the ninth magnitude 

 I cannot estimate this absorption at less than one-third of its total 

 light — probably more. At all events the aspect of star- distribution 

 on which I have dwelt deserves more attention than it has hitherto 

 received; and now that photographic measures of the intensity of 

 the light of stars are being superadded to the photometric measures 

 previously employed, more numerous and reliable results may ere 

 long be expected. 



There are two other modes of regarding the problem which I 

 may briefly notice. The first is : — We have seen that on the 

 hypothesis of uniformity the 20 inner stars whose average light 

 is 1 are succeeded by a set of 140 stars whose average light is -f J 

 these by a third set of 380 whose average light is -Jg, and so on. 

 Now, instead of dividing the stars into successive sets of 20, 140, 

 380, &c., in order of brilliancy, and ascertaining whether their 

 total light exceeds or falls short of the constant quantity supposed 

 by our hypothesis, we may divide them into successive sets whose 

 average light is 1, j-, ^g-, &c., and examine whether the number of 

 stars in each set exceeds or falls short of that given by the theory. 

 I endeavoured to apply this method, and the results were rather 

 singular. The first set of course contains the 20 first-magnitude 

 stars as before, but the second set, instead of containing 140, was 

 found to contain no less than 508. After such a figure a deficiency 

 in the succeeding set was perhaps to be expected. At all events 

 it appeared, for the set only contained 245 stars while theory -gave 

 380. But when I came to the fourth set I found that it vanished 

 altogether. Its average light is 3-^ ; but the third set, constructed 

 on the present principle, terminated among the fifth magnitude 

 stars, whose average brightness is between -^-g- and ^i^-. It thus 

 appeared as if all the stars whose light averaged -^y were required 

 to balance the brighter stars of the third set, and bring down the 

 average of that set to -j-^g-. Probably, however, the complete vanish- 

 ing of the fourth set was illusory. The set commences in the 

 first hundred of the 1100 stars of the fifth magnitude, and there 

 can be little doubt that the brilliancy at this stage exceeds -^j. 



