196 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 3 3 



distances of the 80-odd globular clusters, curiously distributed over 

 the sky, being nearly all in one hemisphere, which he found to vary 

 from 20,000 to 200,000 light-years. When their positions with re- 

 spect to the plane of the Milky Way were plotted, the remarkable 

 fact emerged that they were distributed symmetrically with respect 

 to the central plane with the same number on each side, as seen in 

 figure 5, the next step in the scale of the universe. There seems to 

 be no escape from the conclusion that the clusters and the stars 

 belong to the same dynamical organization, that each forms part 

 of the galaxy, and that they must be concentric and coterminous with 

 one another. The central position of our sun in the galaxy forever 

 disappeared when the distribution of the clusters showed that it was 

 about half-way between the center and the edge of the great dis- 

 coidal system of stars, which Shapley estimated as having a diametei- 

 of 300,000 light-years and a thickness of 12,000. 



Shapley's conception of the galaxy has been generally accepted 

 with the exception that more recent investigations with more ex- 

 tensive material have shown that Shapley's distances of the clusters, 

 and hence the scale of the galaxy, should be reduced by approxi- 

 mately 40 percent. But even such a reduction leaves the galactic 

 system nearly 200,000 light-years in diameter, of almost inconceiv- 

 able dimensions, requiring for example 50,000,000,000 tons of spider 

 web to span. 



The modern conception of the sidereal system is then not a single 

 watch-shaped cluster of stars surrounding the sun and gradually 

 thinning out toward the edges, but rather a great aggregation com- 

 posed of this cluster and numerous others, represented by the Milky 

 Way clouds gradually merging into one another at the edges where 

 the stars are thinner, the whole forming one great disklike system 

 roughly circular in outline, but with its thickness only about one- 

 twentieth of its diameter. The globular clusters, though part of the 

 galaxy, are, however, nowhere near the disk (see fig. 5), but form 

 a roughly spherical or spheroidal shell concentric w^ith the disk, a 

 cluster of guardian attendants on each side of the main disklike 

 system — the latter estimated to contain some 200,000,000,000 s<"ars, of 

 which our solar system is but one average member. 



A very natural question arises as to the structure and arrangement 

 of the stars in this system. We have seen that, while the stars in 

 the neighborhood of the sun are fairly uniformly distributed in a 

 watch-shaped cluster, gradually thinning out toward the edges, the 

 photographs of the Milky Way clouds show that the structure of 

 adjacent clusters or star clouds as viewed from our position in the 

 disk itself is complex and apparently irregular, without orderly ar- 

 rangement. But may this not be due to our situation within the 



