214 ANNUAL REPORT SMITHSONIAN INSTITUTION, 193 3 



" interstellar " lines, as we prefer to call them, were generally equal 

 and opposite to the apparent solar motion, toward Hercules, deter- 

 mined by Campbell, the conclusion was hardly escapable that these 

 high temperature stars were in motion, frequently in rapid motion, 

 through widely extended clouds of ionized calcium which are, com- 

 paratively spealdng, at rest with respect to the stellar system as that 

 system is represented by Campbell's 280 stars. Further, the D lines 

 of neutral sodium atoms having the same sharpness and stationary 

 character as H and K, which had been predicted by Slipher and 

 had been observed by Miss Heger ® at the Lick Observatory in some 

 B-type stars, were also shown to be present along with the spectra of 

 some of the 0-type stars. 



Since both calcium and sodium have been found in interstellar 

 space, there seems then good grounds for the hypothesis that the 

 interstellar matter is of the same general composition as the stars 

 and contains most of the known elements. The sodium D lines and 

 the calcium H and K lines are, however, the lines most likely to be 

 observed in celestial spectra. The principal spectral lines of other 

 elements of sufficiently great relative abundance in cosmic matter 

 to be expected in this connection lie in the far ultraviolet spectrum, 

 and are cut off from our observation by the ozone of the earth's 

 atmosphere. The recognition that the interstellar lines did not gen- 

 erally appear in stars of lower temperature than the very blue ones, 

 taken in conjunction with Hubble's '^ recent theory of the excitation 

 of the gaseous nebulae to visibility by neighboring high-temperature 

 stars, led to the hypothesis, mainly due to H. H. Plaskett, that the 

 calcium and sodium in this widely extended diffuse interstellar mat- 

 ter are rendered absorbing at H and K and at the D lines by the 

 excitation of neighboring stars of very high temperature of the 

 types designated " O " and " B " of the Harvard star classification. 

 This hypothesis,® advanced in 1923, required an extensive distribu- 

 tion of this diffuse gaseous matter, nearly stationary with respect to 

 the stellar sj^stem, and extending to distances of several thousand 

 light-years throughout the space inhabited by the O- and B-type 

 stars. 



A great stimulus to, and a great advance in, the problem of the 

 cosmic diffuse matter in the galaxy was given by Eddington ^ in the 

 1926 Bakerian Lecture of the Royal Society. He discusses theoret- 

 ically the physical conditions likely to be present in diffuse matter 

 in interstellar space and showed the probability of uniform distri- 

 bution except where condensations in this matter gave rise to the 



« Licks Obs. BuU., vol. 10, no. 326, p. 59, 1919 ; no. 337, p. 141, 1921. 



^Contr. Mt. Wilson Obs., vol. 11, pp. 252, 397, 1922. 



'Monthly Not., vol. 84, p. 80, 1923. 



•Proc. Roy. Soc, ser. A, vol. Ill, p. 424, 1026. 



