622 SCIENCE PROGRESS 



measurement of the wave-lengths of the Fraunhofer lines by 

 Rowland and others, when compared with the measures of 

 the wave-lengths of the same lines as obtained from terrestrial 

 sources, shows that there is a general shift of the Fraunhofer lines 

 towards the red. Various reasons have been assigned for this 

 shift and several recent communications have dealt with the 

 subject. The original explanation was that the effect was 

 one due to pressure, the requisite pressure in the sun's atmo- 

 sphere being about 5 atmospheres. This explanation has 

 several strong arguments against it. Julius attempted an 

 explanation based on his theory of anomalous dispersion in 

 the Sun's atmosphere. If this theory is true there should be 

 a mutual influence of the Fraunhofer lines upon one another ; 

 the theory requires that the effect should have opposite signs 

 for a given line according as the companion line is towards 

 its blue or its red side, and that it should be larger in the 

 latter case than the former. Julius claimed that his theory 

 was supported by the measures made by St. John of the 

 displacement of the Fraunhofer lines at the edges of sun-spots. 

 St. John (Ap. J. xli. 191 5, p. 28) denied this, stating that 

 systematic errors were introduced by the method of discus- 

 sion and the selection of lines adopted by Julius. The 

 question was taken up again by Sebastian Albrecht (Ap. J. 

 xli. 1915, p. 333), who compared the wave-lengths of the 

 iron lines in the solar spectrum, as given by Rowland, with 

 the laboratory wave-lengths of the same lines on the Inter- 

 national system. This made it possible to eliminate systematic 

 pressure shifts. Albrecht then found that lines with close 

 companions towards the red were shifted towards the violet 

 and those with close companions towards the violet were 

 shifted towards the red, the former effect being the larger. 

 These results are in accordance with the deduction of Julius's 

 theory, and Albrecht accordingly concluded that his results 

 seemed definitely to establish the operation of anomalous 

 dispersion in the Sun. J. Evershed (Observatory, xxxix, 1916, 

 P- 59) regards these results as fictitious. He remarks that on 

 Julius's theory all close double lines should show a wider 

 separation in the solar than in an arc spectrum, where there 

 is no question of anomalous dispersion. Photographing the 

 Sun and arc spectra together, and selecting at random two 

 pairs of iron lines well shown in both $un and arc spectra, 



