ASTRONOMY 



529 



(a) Mean absolute photographic magnitude of blue 



stars (C.I. < 0.0) ..... 



(b) Maximum absolute photographic magnitude of 



cluster stars ...... 



(c) Median absolute photovisual magnitude of long- 



period Cepheids ..... 



(d) Hypothetical annual proper motion . 



36,000 light-years. 



Between - i-o 

 and -2'0 



-2 

 o'-oo4 



3,600 light-years 

 or less. 



-f 5 or fainter 

 -j- 3'2 or fainter 



+ 3 or fainter 

 o'-04 or greater 



As regards (a), stars whose colour-index is negative are 

 blue stars of spectral type B. Such stars are all giants, of 

 relatively large mass and high intrinsic luminosity. An 

 absolute magnitude fainter than +3** is exceptional for a 

 B-type star, and an average star of this type would only appear 

 of the fifteenth magnitude, as in the Hercules cluster, if more 

 than 30,000 light-years away. This seems decisive against the 

 smaller distance assigned by Curtis. 



As regards (b), Shapley compares the maximum absolute 

 photographic magnitude of stars in various near-by groups and 

 clusters : in no case were the brightest stars fainter than 

 -fl'^o. They are therefore giant stars. The distance 

 assigned by Curtis to the globular clusters would imply that the 

 brightest stars in clusters are dwarf stars ; this is improbable 

 when comparison is made with the near groups. But the 

 matter has been directly tested at Mount Wilson by photo- 

 graphing the clusters through a thin prism on special plates 

 sensitive in the red and blue regions, and relatively insensitive 

 in the green-yellow. In this way small spectra divided in 

 the middle are obtained, and the relative intensities of the 

 red and blue regions can be compared. A comparison with 

 spectra of known giants and dwarfs enables the conclusion to 

 be made that in absolute brightness the cluster stars equal, or 

 even exceed, the average giants. 



Curtis has criticised the distances deduced by Shapley 

 from the parallactic motions. But, as seen from (c) above, 

 Curtis would make the Cepheids in Messier 1 3 five magnitudes 

 fainter than Shapley ; this would correspond to a mean annual 

 parallactic motion of the order of o^-ieo, which is much too 

 great to be admissible. 



As far as (d) is concerned, it may be deduced from the 

 average radial velocity of the clusters that the mean proper 

 motion of a cluster at a distance of 3,600 light-years would be 

 of the order of o'''04 per year. No cluster is known with so large 

 a proper motion ; in fact, the motion can hardly exceed o"-oi. 



These four lines of evidence, therefore, concur in supporting 

 distances of the order found by Shapley, and there would seem 



