528 SCIENCE PROGRESS 



that as there is an extensive overlapping of the Hnes of the 

 different series, and as there are changes of relative intensity 

 with temperature, the cyanogen band is not suitable for a 

 definitive test of the theory. 



Perot has also used the magnesium line \5i 72. His 

 observed sun-arc shift was corrected for a centripetal move- 

 ment in the solar atmosphere of 1-57 km. per second. After 

 reducing the arc wave-length to zero pressure, he again found 

 a displacement of the amount required by theory. St. John 

 points out, however, that this magnesium line is subject to a 

 marked pole-effect, and that three independent investigations 

 at Mount Wilson have failed to show the centripetal motion 

 assumed by Perot. " Fabry and Buisson find that the differ- 

 ences between the arc wave-lengths of iron reduced to vacuum 

 and the wave-lengths at the sun's centre are of the order of 

 the Einstein effect. They conclude that the differences, sun 

 mifiiis arc, are perfectly interpreted by assuming the Einstein 

 effect to be the sole cause of the displacement of the Fraun- 

 hofer lines. They assume zero pressure in the solar atmosphere, 

 but disregard the limb-centre shifts, which they formerly 

 referred to an increase in pressure of 7 atmospheres in 

 passing from the centre to the limb, and which, if taken into 

 account, would give a displacement in excess of the Einstein 

 requirement. Perot applies a large correction, o-02 7A, for 

 centripetal motion, and finds the difference between the wave- 

 lengths at the limb and the arc in vacuum to agree with the 

 Einstein effect. Fabry and Buisson apply no correction for 

 radial movement of the solar vapours, and find the differences 

 between the wave-lengths at the centre and the arc in vacuum 

 to agree with the Einstein requirement, but not the differences 

 between the wave-lengths at the limb and the arc," 



It will be seen that the present state of the question is 

 an extremely unsatisfactory one, different and mutually in- 

 consistent corrections having been applied by various investi- 

 gators. As St. John states, the problem must be envisaged 

 as a whole and not in detached portions, and it is intended 

 to carry out at Mount Wilson an extensive programme on 

 sun-arc displacements, including observations at centre and 

 limb and covering the widest possible range in wave-length 

 and line-intensity, so that a large amount of reliable data 

 will be accumulated for statistical discussion. It is only in 

 this way that it will be possible to disentangle the Einstein 

 effect from the other disturbing factors and to determine 

 accurately its amount. The process will necessarily be a slow 

 one, but the result may be accepted with confidence. 



The Boundary of a Star. — It is well known that there is a 

 gradual dissipation of an atmosphere owing to the gradual 



