414 



SCIENCE. 



[X. S. Vol. XXI. Xo. 533. 



part of the year. As an indication of the 

 performance of the instrument the latitudes 

 resulting from a preliminary reduction of 

 the observations made on this star are here 

 given. 



1904, Dec. 



G, 



= 39 58 1.83 





7, 



1.95 





8, 



1.85 





9, 



1.86 





13, 



1.81 





14, 



1.79 





16, 



1.84 





18, 



1.92 





31, 



1.79 



1905, Jan. 



1, 



2.00 



Variations of the Bright Hydrogen Lines 

 in Stellar Spectra: Annie J. Cannon. 

 Stars whose spectra are of the Orion 

 type, having also one or more bright hydro- 

 gen lines, form a most interesting peculiar 

 class whose position in the scheme of stellar 

 evolution is enigmatical. The Harvard 

 photographs show that the bright hydrogen 

 lines are variables in the following six 

 of these stars, rj Centauri, k Apodis, v 

 Sagittarii, e Capricorni, J Velorum and 27 

 Canis Majoris. So far as known, no varia- 

 tion in the light of any of these stars has 

 ever been observed, although the changes 

 in their spectra point either to great at- 

 mospheric upheavals or to movements of 

 two or more revolving bodies. 



The most important changes in the 

 spectrum of -q Centauri may be summarized 

 as follows : In 1897 all lines were dark and 

 H/? was nearly as intense as Hy. In 1898 

 and 1899 H/8 was very faint and appeared 

 as a dark line superposed on a faint bright 

 band. In 1901 a most striking change had 

 taken place, for H/? had become a strong 

 bright line, having considerable shift 

 towards the violet when compared with the 

 dark present in 1897. Hy was dark 

 with a bright band towards the violet. 

 Photographs taken in 1902 recorded the re- 

 appearance of the dark line on the edge of 



greater wave-length of bright H;8, and both 

 lines were of moderate intensity. In 1903 

 the speetnam was similar to that of 1898. 

 The period of these changes is probably 

 several years in length. 



The changes in the spectrum of k Apodis 

 are somewhat similar to those of -q Centauri. 

 It appears that both these stars are spectro- 

 scopic binaries, one component of each be- 

 ing a bright line star. The spectrum of u 

 Sagittarii presents another difficult spectro- 

 scopic problem, perhaps on the order of 

 /3 Lyrce. The spectrum of v Sagittarii 

 always appears to be composite. The prin- 

 cipal lines seem to be due to two bodies, one 

 having a spectrum like /8 Orionis and the 

 other like e Aurigce. The spectrum of j8 

 Orionis was strongly predominant on seven 

 photographs, but frequently the two spectra 

 seemed to be equally intermingled. Per- 

 haps the most curious phenomenon is that 

 on twenty-three photographs, on which the 

 helium lines were very strong, those of 

 hydrogen were unusually weak. H^ was 

 invisible, appearing neither as a line of 

 emission nor of absorption, while line 4.922 

 was clearly seen. Hy and HS were re.spect- 

 ively much less intense than the adjacent 

 helium lines at 4,387.8 and 4,120.5. It is 

 possible that a third body, having bright 

 hydrogen lines, might explain these ap- 

 pearances. 



Eleven photographs of the spectrum of 

 e Capricorni, taken in 1903, showed H)3 

 to be a faint but distinct bright line lying 

 on the edge of greater wave-length of an 

 equally faint dark line. On earlier photo- 

 graphs, up was dark and of varying in- 

 tensity. Some faint lines, including sev- 

 eral due to iron, are also subject to change 

 in this spectrum. 



It is possible that varying atmospheric 

 conditions may account for the changes in 

 the spectra of J Velorum and 27 Canis 

 Majoris. On June 2, 1893, the dark Bp 

 and Hy in J Velorum had a fine bright line 



