530 SCIENCE PROGRESS 



positive tendency in the residual line-of-sight velocities of the 

 B-stars. This has tentatively been suggested by Campbell as 

 due to a pressure-effect in the atmospheres of these stars. 

 Perrine assumes that it represents a real outward motion, and 

 attempts to find an explanation for it. Separating the stars 

 into two classes of large and small proper-motion, which 

 separation gives in the case of the B-stars a fairly complete 

 sorting into near and distant stars, he finds that the outward 

 motion occurs principally with the brightest and nearest stars. 

 For the fainter stars, the proportion with velocities of approach 

 increases for fainter magnitudes and smaller proper-motions. 

 The preponderance of outward motion is not found with the 

 late type stars. Perrine accounts for these results on his theory 

 that the great brightness and peculiar spectral condition of 

 the B-stars result from the action of cosmical matter in the 

 galactic region in which the B-stars are concentrated by sup- 

 posing that a process of selection has been at work. He 

 imagines a ring of cosmical matter, the stars which are moving 

 outwards pass into this ring of matter and become B-type 

 stars. Some stars also enter it from outside, but such stars 

 are approaching the observer and are fainter. In such a theory 

 we would naturally expect the brighter and nearer stars of 

 Class B to show almost entirely outward motions. These 

 conjectures are interesting, although, as previously pointed out, 

 their acceptance involves difficulties. Further facts need to 

 be accumulated before the theory can be viewed in its proper 

 perspective. 



The Distribution and Dimensions of Stellar Clusters. — Results 

 of considerable importance are obtained by Harlow Shapley in 

 the seventh of his series of studies based on the colours and 

 magnitudes in stellar clusters (Astroph. Journ. 48, 154, 191 8). 

 It has been previously explained in these notes (Science 

 Progress, 12, 552, 191 8) how several collateral lines of in- 

 vestigation, based upon the properties of Cepheid variables, 

 have enabled estimates of the distances of stellar clusters to 

 be made with an accuracy far exceeding what would be possible 

 by direct measurement. Shapley has applied these methods 

 to determine the distances and distribution in space of sixty- 

 nine globular clusters. One important result which is thus 

 obtained is that there is a definite relationship between the 



