528 SCIENCE PROGRESS 



concentration towards the galactic plane seems to indicate this, 

 and both Shapley and Curtis accept the assumption. Shapley 

 then uses the distances which he found for the globular clusters 

 by various indirect methods to define the extent of the galactic 

 system, and obtains a diameter for the system of the order of 

 300,000 light-years. Curtis denies the reliability of this figure, 

 and argues in favour of a diameter of only about one-tenth 

 the amount. 



The process by which Shapley deduced the distances of the 

 globular clusters may be briefly recalled. In the Lesser 

 Magellanic Cloud a number of Cepheid variables were discovered 

 by Miss Leavitt. Variables of this type are characterised by 

 a light curve by means of which they may easily be identified, 

 and it is believed that the light variation is due to pulsations 

 in the atmosphere of the star. Miss Leavitt found that there 

 was a definite relationship between the period and apparent 

 magnitude of these stars. But this implies a direct relationship 

 between period and absolute magnitude, since the Magellanic 

 Cloud is so distant that the differences between the distances 

 of individual variables may be neglected. Under such con- 

 ditions absolute magnitude differs from apparent magnitude 

 by a constant, which cannot be determined, however, unless 

 the distance of the cloud can be determined. 



Shapley found that in many globular clusters Cepheid 

 variables occur, and investigation of their variation showed 

 that in each case their periods and magnitudes fell on a curve 

 which, by adjustment of its zero, could be fitted to the curve 

 obtained from the variables in the Magellanic Clouds. These 

 results served to confirm and extend the original curve. To 

 determine the zero, Shapley discussed the near Cepheids in the 

 Galactic System, and from their proper motions determined 

 their mean parallax and, therefore, their mean absolute magni- 

 tude. From this curve, so standardised, the absolute magnitude 

 and thence the distance of any Cepheid variable can be deduced, 

 when its period of variation has been determined. In this 

 way, and by certain extensions of the method to which it is 

 not necessary now to refer, Shapley showed that the distances 

 of the globular clusters range up to about 200,000 light- 

 years. 



Curtis accepts the relative distances determined by Shapley 

 as correct, but criticises the method by which the zero is fixed, 

 and argues in favour of a much smaller scale. The differences 

 in the consequences resulting from the two view-points may 

 be illustrated for the case of the Hercules cluster (Messier 13), 

 for which Shapley finds a distance of 36,000 light-years, but 

 for which Curtis would assign one of the order of 3,600 light- 

 years. 



