THE SEMI-CENTENNIAL ANNIVERSARY 51 



the results must be given with some reserve, though they appear 

 to leave no doubt as to the reality of the effect. Observations in 

 the second order spectrum failed to give satisfactory indications 

 of the field. But with the higher dispersion of the third order, 

 eleven independent determinations, made with every possible 

 precaution to eliminate bias, show opposite displacements in 

 the northern and southern hemispheres, decreasing in magnitude 

 from about 45 north and south latitude to the equator. Three 

 of these determinations were pushed as close to the poles as con- 

 ditions would permit, and the observed displacements may be 

 compared with the theoretical curve (Fig. 15). In view of the 

 very small magnitude of the displacements, which never surpass 

 0.002 Angstroms, the agreement is quite as satisfactory as one 

 could expect for a first approximation. 



The full details of the investigation are given in a paper 

 recently published.* The reader will find an account of the 

 precautions taken to eliminate error, and, I trust, no tendency to 

 underestimate the possible adverse bearing of certain negative 

 results. It must remain for the future to confirm or to over- 

 throw the apparently strong evidence in favor of the existence 

 of a true Zeeman effect, due to the general magnetic field of the 

 sun. If this evidence can be accepted, we may draw certain con- 

 clusions of present interest. 



Taking the measures at their face value, they indicate that the 

 north magnetic pole of the sun lies at or near the north pole of 

 rotation, while the south magnetic pole lies at or near the south 

 pole of rotation. In other words, if a compass needle could 

 withstand the solar temperature, it would point approximately 

 as it does on the earth, since the polarity of the two bodies ap- 

 pears to be the same. Thus, since the earth and sun rotate in 

 the same direction, the preponderant influence of a negative 

 charge distributed through their mass (in combination with an 

 equal positive charge of slightly different volume density) would 

 account in each case for the observed magnetic polarity. 



* Contributions from the Mount Wilson Solar Observatory, No. 71. 



