192 SCIENCE PROGRESS 



tude with the theoretical values that would be deduced on 

 the supposition that the sun was a uniformly magnetised 

 sphere, the magnetic poles of the sun being at or near the 

 poles of rotation. The approximate vertical intensity of the 

 sun's general field at the poles was found to be about 50 

 gausses. Evidence was also obtained that the general mag- 

 netic field decreased rapidly in intensity in the upper levels 

 of the solar atmosphere. It was a remarkable piece of work 

 which enabled such conclusions to be put forward with con- 

 fidence, based upon what at first sight appeared slender evi- 

 dence, the displacements obtained being very small and not 

 exceeding o'ooi Angstrom unit in amount. Further results 

 since obtained have, however, confirmed the earlier conclusions : 

 they are given in a paper by G. E. Hale, F. H. Seares, A. van 

 Maanen and F. Ellerman in the Astroph. Journ. 47, 206, 191 8. 

 In this paper, measures of displacements are given for 26 

 lines in the solar spectrum belonging to iron, chromium, 

 nickel, vanadium, and titanium. Eighteen ether lines, which 

 had previously been found susceptible to the influence of the 

 magnetic field in sun-spots, showed no measurable shift. The 

 explanation of this is not clear. Using the laboratory data 

 for the separation of these lines by a magnetic field of known 

 strength, the field-strengths producing the observed separa- 

 tions were calculated and it was found that the field-strength 

 decreased with increasing line-intensity. The strongest lines 

 are those which originate in the upper levels of the solar atmo- 

 sphere, and therefore this result is interpreted as indicating 

 a field-strength diminishing rapidly with increasing elevation. 

 Using Mitchell's results for the depths at which the chromo- 

 spheric lines originate, it is concluded that the part of the 

 field accessible to observation lies within the bounding sur- 

 faces of a thin shell in the solar atmosphere about 150 km. 

 thick and that definite values of the calculated field-strength 

 always correspond to definite levels in the solar atmosphere. 



The general nature of the sun's magnetic field may now 

 be said to be known with some certainty, but the underlying 

 causes of it remain obscure. The hypothesis that it is due 

 to local whirls is examined, but the evidence on the whole 

 seems to be against it. 



The Atmospheric Scattering of Light. — The atmospheric 

 transmission coefficients obtained at Mount Wilson from 19 10 



