RECENT ADVANCES IN SCIENCE 99 



would place the star in a much earlier type. This leads to a 

 division of the late type stars into two distinct classes, re- 

 spectively of high and low luminosity. The usual method of 

 classification failed to make this distinction. This division is 

 required by the new theory of stellar evolution above referred 

 to, and is a further strong argument in its favour. Another 

 investigation pointing in the same direction was mentioned in 

 these notes, Science Progress, vol. x. p. 439, January 1916. 

 It is interesting to note that one of the strongest arguments 

 against the new theory was that of the spectroscopists who 

 could not believe that a very condensed cold star could show 

 the same type of spectrum as a very diffuse hot gaseous star. 

 It now appears that there are differences in the spectra which, 

 however, were not immediately apparent. 



The Absorption of Light in Space. — In these notes in the last 

 number of Science Progress, p. 620, it was mentioned that 

 recent estimates of the amount of absorption or scattering of 

 light in space led to the surprising result that the total mass 

 of dust to which this is due must be about 300,000 times that 

 of the lucid stars. An important paper by Harlow Shapley in 

 Proc. Nat. Acad. Set. ii. p. 12, January 191 6, may lead to a 

 revision of this conclusion. It is obvious that, ceteris paribus, 

 the reliability of a determination of this scattering is pro- 

 portional to the distances of the stars upon which it is based. 

 The estimates referred to were based upon stars which are 

 comparatively near the Sun. Shapley has approached the 

 matter in a different way ; he has measured the colour-indices 

 of the stars in the cluster M 13 and compared their distribution 

 with those of stars of similar type near the Sun. Now the 

 actual distance of this cluster is not known, but it is certain 

 that it must be at least 5,000 light-years and probably much 

 greater, and the estimates previously referred to would require 

 the smallest colour indices to be at least 2*5 magnitudes, whereas 

 many are negative and there are none greater than 2. There 

 seems to be no escape from his conclusion that the absorption 

 of light in space must be inappreciable. It may be that there 

 is cosmic dust scattered in space comparatively near the centre 

 of our Universe, but that away from the centre its density 

 rapidly decreases. This would reconcile the results deduced 

 from the near and the distant stars. It remains for further 

 investigations to settle this point. In any case, the difficulty 



