136 ANNUAL REPORT SMITHSONIAN INSTITUTION^ 1915. 



STAKS AND NEBUL.E. 



The observatory at Greenwich has undertaken the task during re- 

 cent years of the redetermination of the precise positions of ail the 

 bright stars of the corona borealis, stars already included by Car- 

 rington in a catalogue which is now half a century old. It has thus 

 become possible to study and classify a great number of proper mo- 

 tions. The discussion made by Dyson gives a result favorable to 

 the views held by Schwarzschild that the existence of a single pref- 

 erential direction for stellar motions is probable. In measure as 

 we consider a direction differing from this, the number of stars 

 having this different direction diminishes regularl3^ From the re- 

 lation between the brightness of a star and its apparent motion, 

 it may be deduced that the distribution of stars in space is neither 

 uniform nor fortuitous. The greatest frequency is found in the 

 constellation of Gemini at a distance which is small compared with 

 the dimensions of the Milky Way. "WTien we depart from this central 

 region the frequency of the stars diminishes without limit, so that 

 we may speak of the stars visible in meridian instruments as a 

 limited system of definite structure. 



Analogous conclusions w^ere derived by Eddington from the study 

 of the catalogue of Boss, in which are collected the most accurate 

 data concerning the bright stars in all j)arts of the sky. It is espe- 

 cially in high galactic latitudes that the density is found to decrease 

 most markedly. We must therefore regard the stars connected with 

 the Millv}^ Wa}^ — that is, the great majority of the visible stars — as 

 forming a globular cluster with a very marked flattening. 



At each point of such a cluster the Newtonian attraction must pro- 

 duce a field of force. A star, obedient to this field of force and 

 sensibly untroubled by neighboring bodies, would complete its revo- 

 lution about the center in about 300,000,000 years, and we would 

 expect a definite, dominant direction in each region of space. 



The researches of L. Boss and of Messrs. Hertzsprung and Plum- 

 mer have definitely revealed the existence of several groups or fami- 

 lies of stars, all the members of which travel with equal and parallel 

 velocities and having a yet further kinship in the character of their 

 spectra. These stars thus preserve a trace of their common origin 

 and move freely or under the action of a common field of force, and 

 resemble very little the final state of bodies intermingling with 

 diverse velocities. It is therefore necessary to abandon the com- 

 parison of the Milky Way to a gaseous mass where the velocities of 

 the molecules result from multiple collisions in every direction and 

 with velocities showing definite relationship to the masses, but 

 no regularity as to direction. Jeans, starting with the stellar density 

 calculated as existing in the neighborhood of our sun, has found 



