KTJNGL. SV. VET. AKADEMIENS HANDLINGAR. BAND 60. N:0 8. 13 



data for the photo-visual magnitudes are used as a basis, and if only those stars 

 are used that are at a distance >2' from the centre of the clusters: 



M 



'Ma 



14,99 



0", 000290 





"Jt/a 



15,99 



0". 000154 



Owing to the fact that it has perhaps not been possible for Shapley to survey 

 all the stars down to the limiting magnitnde 15,8 and 17,o respectively, I have chosen 

 the above given limits for m Q . For m 13,99 and 14,99, one would obtain the values 

 0",oooioo and 0",ooo2oo for the parallaxes. The limits 15,79 and 16,99 give higher 

 parallax-values. 



The most certain values of the parallax in conformity to the above process are 

 obtained if one has determined the luminosity curves for different spectral classes, 

 and knows the photometrical magnitude of the stars within a certain spectral class 

 up to the magnitude m . Generally it has not yet been possible to determine the 

 luminosity -law for the different spectral classes, but for B — B h Kapteyn 100 has ob- 

 tained a no doubt good determination of the luminosity-curve. He has also deter- 

 mined the luminosity-curve for the B b — B and A — A 9 stars 101 . 



Kapteyn's frequency-curve for B — B h has the following form: 



0,409 fl — ö^Ö9»(lf— O,*»»)" in\ 



1/3,14 • 



If we suppose that Shapley's colour-classes correspond to spectral classes, and 

 that, consequently, his stars designed b Q — b 6 belong to the spectral types B — J9 8 , 

 we obtain in accordance with (3) and (7): 



Ä Jfi = 0",000043, 

 71 4f»=0" 5 000106. 



This determination is very likely the most certain of the determinations made 

 by means of the luminosity-law, and the double weight ascribed to these values 

 seems to be well justified. We are now going to use the frequency-curve for 

 B h — B 9 , which has the form: 



g -0,^2(^-2,»)^ 



1/3,14- 



From Shapley's data is obtained: 



^Ma = 0",000210, 

 z M n =Q"> 000190- 



