KOTATION OF THE GALAXY EDDINGTON 249 



and spiral nebulae which are too remote to partake of it. Further, 

 since stars with a large individual velocity additive to the general 

 orbital velocity would escape from the system, we have a simple ex- 

 planation of a well-known phenomenon, viz, that high velocity stars 

 favor a direction now identified as that opposed to the orbital motion. 

 Finally, the very oblate shape of the stellar system is strongly sug- 

 gestive of rapid rotation; and in the spiral nebulae, which are be- 

 lieved to be patterns of our galaxy, the rotation can be directly 

 observed and measured. 



The evidence seems convincing ; nevertheless a thread of insecurity 

 runs through the whole fabric. It is the old story — our conclusions 

 rest mainlj^ on observations of the northern celestial hemisi^here. and 

 the southern observations make a poor counterweight. This is a com- 

 mon complaint in all discussions of stellar statistics; but I think 

 that in none is it so serious as in the determination of galactic rota- 

 tion. For this the most useful data have been piuvided by the 

 Dominion Astroph3'sical Observatory (British Columbia), which by 

 reason of its rather high latitude is less able than some of the other 

 northern observatories to poach on the southern hemisphere. In th^ 

 investigation of Plaskett and Pearce, whose results I have quoted 

 (p. 244), out of 250 stars only 4 were between 193° and 343'^ galactic 

 longitude; a stretch of one-third of the whole circuit was unrepre- 

 sented by a single star. This is the operation which Kapteyn used 

 to describe as " flying with one wing." B}'' mathematical dexterity 

 the required constants of rotation have been extracted from the lop- 

 sided data; but no mathematical dexterity can avert the possibility 

 that the neglected part of the sky may spring an unpleasant surprise. 

 As a spectator I watch the achievements of our monopterous aviators 

 with keen enthusiasm; but I confess to a feeling of nervousness when 

 my turn comes to depend on this mode of progression. 



THE DYNAMICAL PROBLEM 



The admission of galactic rotation must modify our earlier views 

 in a way which is not always sufficiently appreciated, and I think 

 that there are many who retain an incongruous mixture of the old 

 with the new ideas. The distribution of the stars is far from regular 

 and it has been customary to think of the galaxy as subdivided into 

 a number of vaguely defined aggregations or star clouds. Particu- 

 lar attention has been paid to a supposed aggregation in w'hich the 

 sun is nearly central ; this is known as " the local cluster." Charlier 

 attributed to it a diameter of 700 parsecs with a thickness about one- 

 third as great; others have attributed greater dimensions. (It is 

 sometimes hinted that investigators place the boundary of the local 

 cluster suspiciously near the distance at which their observational 



