44 



NATURE 



[September 9, 1915 



400 are being measured by more than one observatory. 

 We may expect results at the rate of two hundred a 

 year, and may therefore hope for a rapid increase of 

 our knowledge of the stars within our immediate 

 neighbourhood. 



Velocities in the Line of Sight. 



The determination of radial velocities was initiated 

 by Huggins in the early 'sixties, but trust- 

 worthy results were not obtained until photographic 

 methods were introduced by Vogel in 1890. Since, that 

 time further increase in accuracy has been made, and 

 the velocity of a bright star with sharp lines is deter- 

 minable (apart from a systematic error not wholly 

 explained) with an accuracy of i kilometre per second. 

 As the average velocities of these stars are between 

 10 and 20 kilometres a second, the proportional accu- 

 racy is of a higher order than can be generally ob- 

 tained in parallax determinations or in other data 

 of sidereal astronomy. A number of observatories in 

 the United States and Europe, as well as in South 

 America, the Cape, and Canada, are engaged in this 

 work. Especially at the Lick Observatory under 

 Prof. Campbell's direction, the combination of a large 

 telescope, a well-designed spectroscope, and excellent 

 climatic conditions have been utilised to carry out a 

 bold programme. At that observatory, with an off- 

 shoot at Cerro San Christobal in Chile, for the ob- 

 servations of stars in the southern hemisphere, the 

 velocities of 1200 of the brightest stars in the sky 

 have been determined. Among the results achieved is 

 a determination of the direction and amount of the 

 solar motion. The direction serves to confirm the 

 results from proper motions, but the velocity is only 

 obtainable accurately by this method. This quantity, 

 which enters as a fundamental constant in nearly all 

 researches dealing with proper motion, is given by 

 Campbell at 195 kilometres per second, or 4-1 times 

 the distance of the earth from the sun per annum, 

 though there is some uncertainty arising from a 

 systematic error of unknown origin. 



The observations of radial velocities have shown 

 within what limits the velocities of stars lie and have 

 given a general idea of their distribution. The most 

 important result, and one of a somewhat surprising 

 character, is that the mean velocities of stars, the 

 motion of the sun being abstracted, increase with the 

 type of spectrum. Thus the stars of type B, the 

 helium stars, the stars of the highest temperature, 

 have average radial velocities of onlj' 6-5 kilometres 

 per second ; the hvdrogen stars of type A have average 

 velocities of 11 kilometres per second ; the solar stars 

 of 15 kilometres per second ; while for red stars of 

 types K and M it has increased slightly more to 

 17 kilometres per second. Further, the few planetary 

 nebulae — i.e. condensed nebulae with bright line spectra 

 — have average velocities of 25 kilometres per second. 

 There can be no question of the substantial accuracy 

 of these results, as they are closely confirmed by dis- 

 cussions of proper motions. Thev are, however, very 

 diiBcult to understand. On the face of it, there does 

 not seem any reason why stars of a high temperature 

 should have specially high velocities. A suggestion 

 has been thrown out by Dr. Halm that as the helium 

 stars have greater masses, these results are in accord- 

 ance with an equi-partition of energy. . But the dis- 

 tances of stars apart is so great that it seems impos- 

 sible that this could be brought about by their inter- 

 action. Prof. Eddington suggests that the velocities 

 may be an indication of the part of space at which 

 the stars were formed (e.g. stars of small mass in 

 outlying portions), and represents the kinetic energv 

 they have acquired in arriving at their present posi- 

 tions. 



NO. 2393, VOL. 96] 



The stars the radial velocities of which have been 

 determined are, generally speaking, brighter than the 

 fifth magnitude. Fainter stars are now being observed. 

 At the Mount Wilson Observatory, Prof. Adams has 

 determined the velocities of stars of known parallaxes, 

 as there are great advantages in obtaining complete 

 data for stars where possible. Extension of line-of- 

 sight determinations to fainter stars is sure to bring 

 a harvest of useful results, and a number of great 

 telescopes are engaged, and others will shortly join 

 in this important work. 



Proper Motions. 



As proper motions are determined by the comparison 

 of the positions of stars at two different epochs, they 

 get to be known with constantly increasing accuracy 

 as the time interval increases. The stars visible to 

 the naked eye in the northern hemisphere were 

 accurately observed by Bradley in 1755. Many 

 thousands of observations of faint stars down to about 

 9-om. were made in the first half of the nineteenth 

 century. An extensive scheme of re-observation was 

 carried out about 1875 under the auspices of the 

 Astronomische Gesellschaft. A great deal of re- 

 observation of stars brighter than the ninth magnitude 

 has been made this century in connection with the 

 photographic survey of the heavens. For the bright 

 stars all available material has been utilised and their 

 proper motions have been well determined, and for the 

 fainter stars this is being gradually accomplished. 



Proper motions differ widely and irregularly in 

 amount and direction. Herschel observed a tendency 

 of a few stars to move towards one point of the sky, 

 and attributed this sign of regularity to a movement 

 of the solar system in the opposite direction. But 

 puzzling differences given by different methods re- 

 mained unexplained until the difficulty was resolved 

 by Prof. Kapteyn in a paper read before this section 

 of the British Association at its meeting in South 

 Africa ten years ago. He showed that the proper 

 motions had a general tendency towards two different 

 points of the sky and not towards one only, as would 

 be expected if the motions of the stars themselves 

 were haphazard, but viewed from a point in rapid 

 motion. He concluded from this that there was a 

 general tendency of the stars to stream in two opposite 

 directions. It is interesting to notice that this great 

 discovery was made by a simple graphical examina- 

 tion of the proper motions of stars in different regions 

 of the sky, after the author had spent much time in 

 examining and criticising the different methods which 

 had been adopted for the determination of the direction 

 of the solar motion. The subject was brought into a 

 clearer and more exact shape bv the analytical formu- 

 lation given to It by Prof. Eddington, and after him 

 by Prof. Schwarzschild. 



This star-streaming is corroborated bv observations 

 of velocities in the line of sight. It applies— with the 

 exception of the helium stars — to all stars which are 

 near enough for their proper motions to be deter- 

 minable. We may say with certainty that it extends 

 to stars at distances of two or three hundred parsecs ; 

 it may extend much further, but I do not think we 

 have at present much evidence of this. Prof. Turner 

 pointed out that the convergence of proper motions 

 did not necessarily imply movements In parallel direc- 

 tions, and suggested that the star-streams were move- 

 ments of stars to and from a centre. The agreement 

 of the radial velocities with the proper motions seems 

 to me to be opposed to this suggestion, and to show 

 that star-streaming Indicates approximate parallelism 

 in two opposite directions in the motions of the stars 

 examined. As the great majority of these stars are 

 comparatively near to us, it is possible that this 



