32 



C. V. L. Charlier 



crowd round the square GD^., (Sagittarius), just as the globulars, and are rare in 

 the opposite direction. Assuming now with Campbell (P. A. P. Vol. 26, 1915) the 

 association of the planetary nebulae with the Galaxy — a conclusion based on the 

 observed radial velocities — we have reason to accept a similar association regarding 

 the globular clusters. 



How is, under this supposition, our scale a to be determined? It may be done 

 either with the help of the ^-coordinate of the centre of the system of globulars or 

 usiug the dispersion in the same coordinate. In both cases we use the corresponding 

 values for the J5-stars as a standard of comparison. The result is somewhat de- 

 pendent on whether, and to what extent, globulars of small apparent diameter are 

 taken into account or not. I have calculated e under three different suppositions: 

 1) considering all clusters (in fact only 70), 2) excluding those having a diameter 

 smaller than 2' (it remains G5 clusters), 3) excluding moreover all clusters having 

 a diameter = 2' (it remains 55 clusters). In all cases I have computed the scale 

 from the mean value of the e- coordinate [M{z)) as well as from its dispersion a{z). 



Value of the scale (-.) of the globular clusters. 



From M(z) 



From z{z) 



1) 1.G3 



1.53 



2) 2.2t 



1.05 



3) 2.41 



1.62 



The order of magnitude of the scale is the same in both cases. The value 

 of £ derived from a(,?) is more independent of the number of globulars excluded than 

 that obtained from M{.<), depending on the smaller value of the mean error in the 

 former case. I have therefore given those values (that are obtainerl from q{z)) greater 

 weight, and, in the computation of the absolute coordinates used the value 



= - 1.60. 



With this value of the scale the values of x\ y' , z' given in table 2 are ob- 

 tained as well as the distances given in col. 16. 



18. Absolute magnitudes. Knowing the distance we are able to determine 

 the ahsohde magnitude of the stars which are members of a globular cluster. Be- 

 fore treating this cjuestion I premise some words on the detinition of the term 

 »absolute magnitude». 



There are used in the astronomical literature three different scales for the 

 absolute magnitudes dilïering by a constant. All define absolute magnitude as the 

 apparent magnitude at a certain distance, but differ in fixing this distance. The 

 three distances used are: 



1 Siriometer corresponding to :r = 0".206 



0.206 » » » TT = 1". 



2.06 » » » jr = 0".l. 



