DEPARTMENT OF MERIDIAN ASTROMETRY. 179 



STELLAR WAVE-LENGTHS. 



The relation between the K-term in the B-type stars and wave-lengths is 

 btill under investigation by Dr. Albrecht. Although the wave-lengths of 

 silicon by Crooks would entirely remove their proportionate share of the 

 K-term, they are unsatisfactory because of the very large internal discordance 

 between results from the individual lines. In response to an appeal for 

 additional laboratory study, new wave-lengths of silicon have been deter- 

 mined by Mr. Barrell from measures by Professor Fowler and himself. The 

 new values, which are in fairly good internal agreement with stellar wave- 

 lengths, give radial velocities which are systematically 1.9 km. smaller than 

 those obtained with the wave-lengths generally employed in radial-velocity 

 determinations. 



For the accurate determination of correlations between stellar wave- 

 lengths, spectral types, and absolute magnitudes it was found highly desirable 

 to reduce the basic data to a much more nearly homogeneous system than 

 was available in the preliminary study of the problem. For this purpose 

 the original reduction sheets for the measures of Palmer, Wright, and Albrecht 

 in Publications of the Lick Observatory (vol. ix, part iv) were kindly made 

 available by Professor Wright and Director Campbell. All of the reduction 

 tables and original measures are being rediscussed practically anew, due to 

 the presence in them of numerous inaccuracies which, although affecting the 

 mean radial velocities for the plates very little, do affect the observed wave- 

 lengths of the individual lines quite appreciably, in both an accidental and a 

 systematic sense. The reduction tables have been entirely recomputed and 

 about three-fourths of the new reductions completed. 



