THE MOVEMENTS OF THE STARS. 



139 



side by side for detailed examination. Without going into 

 technicalities we may say that the appearances of these 

 <: spectra" are sufficiently distinctive to enable us to group the 

 stars into different classes according to the quality of their light. 

 Those who have studied these questions are pretty well agreed 

 that these classes or " types " represent different stages in the 

 life of a star, and we may class the stars in this way as young, 

 middle-aged or old. Now the remarkable result has been found 

 that on the average the older a star is the faster it moves. 

 I refer now not merely to its apparent displacement across the 

 sky but to its actual speed in miles per second. There is a 

 steady increase in the speed from about 6 km. per second for 

 the youngest stars to 17 km. per second for the oldest.* That 

 is true provided that you take the average of a considerable 

 number of stars; of course, speeds of individual stars may 

 differ widely from the average of their type. 



I have only described to you our studies of the movements 

 of the stars. In another branch of the subject parallel investi- 

 gations are being made of the distribution and magnitudes of 

 the stars, which are also extending our knowledge of the 

 stellar universe. It appears from these that we are in the 

 midst of a great mass of stars arranged in an oblate or 

 flattened shape, something like a bun, or perhaps a lens. We 

 are somewhere towards the middle of this distribution. In its 

 thinnest direction our telescopes penetrate quite easily to the 

 limits, or rather to the place where the stars thin out very 

 much, for we cannot suppose there is a definite edge. Eound 

 the circumference of this mass, and continuing its plane, are 

 coiled a great series of star-clouds which appear to us as the 

 Milky Way. The whole structure would probably bear some 

 resemblance to one of the flat spiral nebulas which form such 

 remarkable objects in the sky. 



We know scarcely anything about these spiral nebulas, but 

 the question suggests itself, may they not be replicas of our 

 own stellar system ? That is to say, island universes, not con- 

 tained among the stars but separated from them at a much 

 greater distance than any we have yet spoken of ? There are 

 some hundreds of thousands of these spiral nebulas, so that the 

 conception is almost appalling in its vastness. Suppose that 



* I have given, in accordance with the usual practice, the average 

 value of one component of the motion, e.g., the velocity of recession or 

 approach. The three-dimensional speeds are twice the values given. 



