250 



NA TURE 



[July 



I T, 1907 



arrows from the point Q, (R.A. y^h., Decl. + 58°). 

 It will be convenient to refer to the two systems as 

 Drift I. and Drift II. ; Drift I. accordingly is a group 

 of stars in rapid motion from the point Q,, and 

 Drift II. a group in comparatively slow motion from 

 the point Q,, its velocity being, in fact, not quite 

 one-third that of Drift 1. Fig. 2 also shows that 

 the speed of Drift I. appears smaller in the regions 

 nearest Q,, and of Drift 11. in the regions nearest Q,. 

 This is because that part of the drift motion which is 

 in the line of sight has no effect on the proper motions, 

 and cannot be detected by e.vamining them. The 

 component of the drift motion across the line of sight 

 decreases as the apex of the drift is approached, and 

 vanishes at the apex itself. 



Another result from the mathematical analysis is 

 that the stars are nearly equally divided between the 

 two drifts, Drift II. having, perhaps, a slightly 

 greater proportion. It is rather remarkable that 

 although some parts of the sky are more than twice 

 as rich in stars as others, the approximately equal 

 division between the two drifts is inaintained in every 

 region. 



It is natural to inquire whether there may not be 

 some other distinction between stars of the two drifts 

 besides their motions. The fact that the sun moves 

 comparatively slowly relative to Drift II. rather sug- 

 gests that the sun belongs to this drift. In that case 

 it might be expected that the Drift II. stars would 

 be nearer on the average than those of Drift I., the 

 latter forming a sort of background. The magnitudes 

 of the proper motions (which have up to this point not 

 been considered) afford data for testing this point. 

 Due allowance having been made for the fact that 

 the true Drift II. motion has already been found to 

 be very much less than that of Drift I., the apparent 

 motions indicate the same mean distance for the two 

 drifts. In fact, a mathematical calculation showed 

 less than 5 per cent, difference between the mean 

 distances of the two drifts (though it is improbable 

 that the data can be trusted quite so far as this). 

 Remembering that the two drifts divide the stars in 

 nearly the same proportion in all the regions, it would 

 seem that they are as completely mixed as it is 

 possible to imagine. 



One point of great interest is the distribution of the 

 stars of different spectral types among the drifts. It 

 has been customarv to regard Sccchi's two tvpes as 

 forming to a certain extent two svstems, for Type II. 

 stars are very evenlv distributed throughout the sky, 

 whereas those of Type I. tend to congregate in the 

 plane of the Milkv Way. It is, however, quite certain 

 that the division into two spectral types and the 

 division into two drifts do not correspond. If we 

 examine Type I. stars alone, both drifts are evident; 

 and so also if Type II. stars are examined alone. 

 Nevertheless there is a systematic difference between 

 the distribution of the proper motions of the Type I. 

 and Type II. stars, which manifests itself in every 

 region examined (including (he Hradlev stars). It may 

 perhaps be interpreted as beine due to a higher percent- 

 age of Type II. stars in Drift II. than in Drift I., 

 but it now appears more probnble that the difference 

 consists in Type II. stars hcTvinrj larger " peculiar " 

 motions (the haphazard individual motions) than Type 

 I. stars. In addition to some satisfactory direct evidence, 

 this latter explanation is supported bv the fact that 

 nearly all the " runaway" st.?rs are of Type II., and 

 it also agrees with the difference in distribution of 

 Type I. and Type II. stars; the former mainly con. 

 gregate in one plane, whilst the latter, perhaps 

 orisrinnllv in the same plane, would hrive become more 

 uniformly distributed in consequence of their greater 

 individual velocities. 



NO. 1967, VOL. 76] 



There is no indication of any relation between 

 magnitude and drift, except possibly in the case of 

 the very bright stars (brighter than magnitude 40). 

 These latter seem to belong principally to Drift I., 

 but they are so few in number (in the area examined) 

 that the result mav very well be accidental. 



Having determined the motions of the two drifts of 

 stars relative to the sun, and knowing also that the 

 stars are nearly equally divided between them, it is 

 easy to determine the motion of the sun relative to the 

 combined drifts — in fact, to determine the solar apex. 

 In this way the solar apex is found to be at R..\. 

 i7h. 45m., Decl. + 31° ; it is naturally fairly close to 

 Qi, since the velocity of Drift I. predominates. From 

 the same stars, by Airy's method, Dyson and 

 Thackeray found the solar apex to be at R..'\. 

 i8h. 2om!, Decl. + 37°. The somewhat greater R..\. 

 of their determination (and of most other determina- 

 tions) is probably the result of using the magnitudes 

 of the proper motions. This position, deduced by 

 means of the two-drift theory from the directions of 

 the proper motions only, has the advantage of being 

 free from all assumptions as to the distances of the 

 stars, but the probable accidental error is large. The 

 two-drift theory further directs attention to the triie 

 nature of the " solar motion" so determined; it js 

 in no sense an absolute motion, and there is 

 now no justification for confounding it with the 

 motion relative to the £Ether, as has sometimes been 

 done. 



In conclusion, whilst Prof. Kapteyn's theory ac- 

 counts in a simple manner for the very anomalous 

 and unsymmetrical way in which the directions _ of 

 motion of the stars are distributed, it is still awaiting 

 the verdict of the spectroscopic determinations of line- 

 of-sight velocities. The material exists in plenty for 

 applying this test; it simply awaits examination by 

 those who have access to it. The investigation of the 

 motions of still fainter stars, and of regions of the 

 ■-U-v which have not yet been explored, offers a largi' 

 field for further research. 



A. S. Eddington. 



SEVENTH INTERNATIONAL ZOOLOGICAL 

 CONGRESS. 



THE arrangements for the seventh International 

 Zoological Congress, which will be held at 

 Boston, .August 19-23. under the presidency of Mr. 

 Alexander Agassiz, are now well advanced. The con- 

 gress will open formally on the afternoon of August 

 iq in the Harvard Medical School, and arrangements 

 will then be made for the meetings of the sections. 

 The subdivision proposed is rather elaborate, for there 

 are to be sections on general zoology, systematic 

 zoology, experimental zoology, marine zoology, evolu- 

 tion, heredity, and so on. There will be three general 

 meetings; the International Committee on Zoological 

 Nomenclature, under Prof. R. Blanchard, will con- 

 tinue its arduous labours ; and numerous addresses, 

 communications, and exhibits have already been 

 arranged for. It need hardly be said that the arrange- 

 ments for hospitality are generous. On each day of 

 the strict congress week there will be luncheon at the 

 invitation of the Boston local committee, _ and the 

 ( venings will be occupied with receptions and 

 dinners. 



On the afternoon of August 22 there will be an 

 excursion to Wellesley College ; August 24 will be 

 devoted to a visit to the museums of Harvard Uni- 

 versity; August 21; is Woods Hole Day; August 26 

 Coluinbia University Day; ./Vugust 27 'the Americ.-in 

 Museum Day. On .August 28 the members of the 



