820 



SCIENCE 



[N'. S. Vol. XXIX. No. 751 



dust are distributed with some rough approach 

 to uniformity in general, but with a marked 

 tendency to local aggregations and clusterings. 

 Through this chaos moves the second part, a 

 cluster of stars of great but measurable di- 

 mensions, long, broad, but comparatively thin 

 and including the sun as one of its central 

 members. The diameter of this cluster is to 

 be measured in hundreds of light years and 

 throughout at least a large part of its central 

 regions the stars are much more densely clus- 

 tered than in the external chaos. 



In accordance with well-known dynamical 

 laws this moving star stream would operate 

 much after the fashion of a snow plow, sweep- 

 ing away the cosmic dust from its path and 

 piling it on either hand, above and below the 

 plane of the cluster. The transparent rift 

 thus formed is the milky way through which 

 we see farther and command a view of more 

 stars than through the intensified dust clouds 

 on either side. The dust ejected toward the 

 poles of the milky way constitutes the sub- 

 stance of the nebulae which there abound. 



The narrow half of the milky way is that 

 which lies behind the moving swarm, since 

 here we see the accomplished work of ages and 

 look between dust clouds, long since produced, 

 that converge to a vanishing point. Ahead, 

 in the direction of motion, the work of clear- 

 ing a path is in progress and relatively near 

 at hand, so that the partially cleared space 

 subtends for us a broader angl^ in the midst 

 of which there are collected considerable 

 quantities of primordial dust, the slower- 

 moving particles, which have been captured 

 and permanently annexed by the vanguard of 

 the swarm. The annexed dust cloud consti- 

 tutes the long cleft in the milky way, while its 

 attendant holes and rifts mark unfinished 

 work on the inchoate side of the galaxy. 



The interpretation of the milky way here 

 offered must not be judged solely by its con- 

 sonance with the phenomena above suggested. 

 It must harmonize with every other fact 

 known about the milky way and it has been 

 the task of the author to seek widely for dis- 

 cord as well as harmony between such facts 

 and the theory above outlined, making the test, 

 wherever possible, a quantitative numerical 



one. No serious discordance has been found, 

 but the agreement with knowledge hitherto 

 vaguely (or . not at all) correlated with the 

 milky way is in some instances striking. Thus 

 the researches of Kapteyn, Eddington and 

 Dyson upon the proper motions of the stars 

 have shown that in the main these bodies be- 

 long to two groups in relative motion along 

 the line and in the direction above suggested 

 in explanation of the milky way. Again 

 Pickering has recently announced certain well- 

 marked differences in the distribution of stars 

 of different spectral types, viz., the fainter 

 stars of the galaxy are almost wholly of the 

 first (Sirian) spectral type. The number of 

 stars of the first type increases with diminish- 

 ing brightness in a four-fold ratio for each 

 successive magnitude, as theoretically it should 

 increase if these stars were uniformly distrib- 

 uted through infinite space. Per contra, stars 

 of the second type (solar) increase from mag- 

 nitude to magnitude only in a ratio repre- 

 sented by the number 3.25, thus implying a 

 distribution very unlike that of the first type 

 stars. All three of Pickering's propositions, 

 together with similar ones earlier formulated 

 by Seeliger for the totality of stars, without 

 distinction of spectral type, are nimierically 

 accounted for by the supposition that the stars 

 composing the primitive chaos are mainly of 

 the sirian type while those of the solar group 

 are predominantly of the solar type. By di- 

 rect enumeration Eddington finds that this 

 relative predominance of spectral types is 

 shown in the two groups into which the stars 

 are divided with reference to their systematic 

 proper motions. 



It is easily seen to be a necessary corollary 

 from this explanation of the milky way that 

 stars in or near the galaxy should on the 

 whole appear to move more slowly across the 

 sky than do similar stars remote from the 

 milky way, and that this predicted result is in 

 fact true has been shown by at least two as- 

 tronomers. Another interesting corollary may 

 be derived through the supposition that a star 

 belonging to the solar group may by virtue of 

 its motion be made to pass so near to one of 

 the sirian stars as to produce disturbances, 

 one upon the other, that will long remain 



