526 



SCIENCE 



[N. S. Vol. XLV. No. 1169 



faint stars in the far distant outposts of the 

 G-alaxy, but we have no reason to suspect 

 that their characteristic motions are unique. 

 We can see no escape from the condition 

 that all the stars are in motion. Under 

 these circumstances we are unable to 

 explain how within a great volume of 



Fig. 15. Irregular Nebula N.G.C., 5146, ijhoto- 

 graphed by Curtis with the Crossley Eeflector of 

 the Lick Observatory. 



(The circular halos around the brighter stars are 

 unavoidable defects and are not real.) 



space that is rich in stars there can be a 

 smaller, but still enormous volume of space, 

 nearly free of stars. Their random motions 

 should distribute them more uniformly 

 than we observe to be the case. Let us 

 illustrate by the celebrated black holes, the 

 so-called "coal sacks," in the constellation 

 of Sagittarius (Fig. 17). In a I'egion 

 where the stars are especially plentiful are 

 two small areas all but empty of visible 



stars. An enlargement of a photograph by 

 Barnard, the preeminent student of this 

 subject, shows the effect still better. How 

 can such great "holes" through the stellar 

 system be surrounded by a plenitude of 

 stars moving more or less at random so as 

 to give a high density of star distribution 

 right up to the sharply defined edges of 

 the holes, and yet leave the holes empty of 

 stars? Wdth the help of all astronomical 

 experience we can not explain the phe- 

 nomena by the absence of stars. I think we 

 must assume, with Barnard and others, 

 that the stars are actually there, and that 

 they are invisible because invisible mate- 

 rials between us and the stars are absorb- 

 ing or occulting the light which the stars 

 are trying to send us. Much of the inter- 

 esting structure in the Slilky Way is prob- 

 ably due in a like manner to obstruction 

 by materials lying between us and the great 

 clouds of stars (see Fig. 18). It is char- 

 acteristic of the galactic structure that 

 where luminous nebulae seem to reduce the 

 numbers of faint stars visible, the reduc- 

 tion in numbers of stars extends also far 

 out beyond the limits of visible nebulosity 

 (see Figs. 14, 15 and 16), and we can 

 scarcely resist the conclusion that invisible 

 extensions of the luminous nebular fabric 

 exist as obstructing agents. 



There are many other lines of evidence 

 in support of the hypothesis that invisible 

 matter exists in abundance within the 

 stellar system : 



1. Newcomb and Kelvin, working inde- 

 pendently and on the assumption that the 

 motions of the stars are generated b.y gravi- 

 tational attractions originating within the 

 stellar system, were unable to account for 

 the high observed velocities of stars, except 

 on the hypothesis that the visible stars 

 contain only a small fraction of the matter 

 existing in the system; the greater part of 

 the attracting material in the stellar sys- 



