190 



ANNUAL EEPORT SMITHSONIAN INSTITUTION, 1916. 



by a slight relative increase of the proportion of the fainter stars. 

 That is, the statistical description of the distribution of the stars 

 when attention is diverted from their random features. 



Passing now from the distribution in zones to the question of the 

 total nimiber of stars, the table below exhibits the data before us : 



We see that as we take in successively the second, third, down to 

 the seventeenth magnitude, the proportionate increase of numbers, 

 which is at first three per magnitude, falls progressively until at the 

 seventeenth it is less than two. 



Beyond this it is almost wholly a matter of inference, but the 

 jDrogTession is so steady that Mr. Chapman and Mr. Melotte have 

 reduced it to a formula which, within the ascertained range, admits 

 of very little latitude, and shows that at about the twenty-third or 

 twenty-fourth magnitude we should have reached one-half of the 

 total, and that this total would lie between one and two thousand 

 millions. 



I say it is a matter of inference, because hardly any material was 

 available to carry on the counts. Two plates, however, were forth- 

 coming, one by that keen observer, Mr. D'Esterre, and one from 

 Mount Wilson, and these when counted confirmed the forecast num- 

 bers in reassuring fashion beyond the twentieth magnitude— that 

 is to say, down to stars 100 million times as faint as those of the 

 first magnitude. 



Numbers and equivalent light of the stars. 



There is the result, between one and two thousand millions — I sup- 

 pose somewhere about as many as the people on the globe. I confess 



