722 



SCIENCE 



LN. S. Vol. XXXVni. No. 980 



that they are exactly opposite, whereas the 

 streams observed in the sky make an angle 

 of about a hundred degrees. For opposite 

 streams, viewed from a self-moving body, 

 as in our earth, will appear to make an 

 angle and we can readily determine the 

 earth's motion in such a way as to bring us 

 in perfect harmony with observation. Thus 

 far no objection. But there are further 

 consequences. 



. In an elongated universe, as here sup- 

 posed, 'ioth the mean longitudinal motion 

 (what in this lecture was called the stream 

 motion) and the deviations therefrom (the 

 internal motion) must gradually increase, 

 beginning with velocity zero. 



Now as to the internal velocities, this is 

 exactly what we find by observation. Do 

 we find the same for the stream motion? 

 By no means. 



Eecent Mt. Wilson observations have 

 enabled us to derive at least a pretty re- 

 liable value of the relative stream velocity 

 for the first type stars. For the helium 

 stars we can as yet only assign a limit 

 which the relative velocity of the two 

 streams must exceed. For the older stars 

 we have had reliable information for some 

 time. 



All these determinations show, contrary 

 to what takes place with the internal mo- 

 tion, that the relative velocity of the two 

 streams or clouds does not change, or does 

 not change very much, with age. It cer- 

 tainly is not nearly vanishing for the 

 helium stars. It seems to me that this con- 

 sideration is fatal to the present explana- 

 tion. 



Professor Schwarzschild has developed 

 a different theory, which also leaves the 

 universe a unit ; but this theory too, elegant 

 though it be, can not, I think, be main- 

 tained. Among other things, we have, as a 

 main objection, the fact — which was not 

 known at the time Professor Schwarzschild 



proposed his theory — ^that the richness of 

 the two streams is not the same for stars of 

 different age. The tacit assumption is 

 made, and must be made, in Schwarzschild 's 

 theory, that the two streams have the same 

 number of stars. Now, this may be more 

 or less approximately true of the stars of 

 the second and third types, for the first type 

 the number of stars in the second stream 

 can not be much different from one third 

 of that in the first stream. For the helium 

 stars it must not be a tenth. The second 

 stream is so poor here that it has been alto- 

 gether overlooked till quite recently. 



The conclusion to be dravni from all 

 this seems obvious. It would seem that we 

 are driven to the theory assumed here, from 

 the first, the theory of the two-star clouds, 

 which, owing to their initial velocity, have 

 come to meet and intermingle in space. It 

 must be confessed, however, that in this 

 theory also there remain some hard nuts 

 to crack. Until we succeed in this it seema 

 unsafe to claim any great certainty for the 

 theory, and it seems preferable to put it 

 forward as the hypothesis which, for the 

 time being, best fits the observed facts. 



There remains to be considered the ques- 

 tion how to explain that the second stream 

 or cloud hardly contains any helium stars. 



There is something in the small local 

 star-groups which may help us. Every- 

 body knows the group of the Pleiades. 

 There can be no doubt that the bright and 

 many of the faint stars that we see in this 

 part of the sky are really near together in 

 space and not merely near the same visual 

 line, the one far behind the other. They 

 undoubtedly form a physical system, and 

 must have had a common origin. At pres- 

 ent we know several of such local- groups, 

 among them the Hyades, the Ursa Major 

 group, and we may perhaps add the great 

 Scorpius-Centaur group. 



Now, in these local groups, we find, 



