stellar clusters 



27 



We find that the number of stars in a globular cluster may be computed from 

 the integrated magnitude, if we know 



1) the mean magnitude, m^, of the stars in the cluster; 



2) the dispersion, n , in these magnitudes. 



Neither nor a is presently known for any cluster. It would not, perhaps, 

 be impossible to procure approximate values of these quantities, so that the number 

 of stars in a cluster may be computed as soon as moreover its integrated magnitude 

 is known from observations. I ob'serve that the value of o in the galaxy is scarcely 

 larger than 4. 



The formula (17) for N may more conveniently be written 



(18) lQÜ..,f»„-0...O--'-W_ 



We find that N is increasing with and decreasing with increasing values 

 of a or [JL. 



As an instance I suppose the question to be about a globular cluster having 

 an integrated magnitude = 5 and a dispersion = 2, so that 



[J. =r 5"' , 0 = 2'" . 



W^e now have 



^ _ -j Q0.4(mo — 6.84) 



and get the following table giving the value of N corresponding to different values 



of THq . 





N 



13 



300 



14 



800 



15 



2000 



16 



5O0O 



17 



12000 



18 



300UO 



19 



80000 



20 



200000 



For M 3 we have, according to Bailey (H. A. 78), [x = 6.0 and, according to 

 üiTOHEY, from mount Wilson photographs with the 60-inch reflector N > 30000, 

 hence the formula gives 



w?„ > 19"', 



taking even here a = 2 . 



15. For obtaining the relative distribution of the globular clusters in space 

 I have here confined me to the apparent diameter as measure of the distance. In 

 like manner as regarding the ordinary clusters I compute for every globular cluster 

 the spherical coordinates ^, rj, C, the galactic longitude (/) and latitude {h), furthermore 



