MOUNT WILSON SOLAR OBSERVATORY. 249 



determination of the corresponding period of rotation. This campaign 

 was begun in June, after the close of the rainy season, and continued 

 throughout four solar rotations until September, with only an occa- 

 sional interruption in the entire period. The resulting measures, which 

 can not be completed before next year, will thus be very uniformly 

 distributed in longitude. Moreover, preliminary observations have 

 shown that occasional sun-spots which appeared during this time will 

 require the rejection of very few plates. 



Twenty-five lines have now been found to give displacements of the 

 same character: positive in the northern hemisphere, negative in the 

 southern, varying in magnitude from zero at the equator to a maximum 

 value near latitude 45°. In every case, therefore, the existence of ellip- 

 tical polarization attributable to the Zeeman effect is clearly indicated. 



The elements represented are iron (10 lines), chromium (8), nickel 

 (4), vanadium (2), unidentified (1). The weakest lines in the list are 

 of intensity 0; the strongest is the wide sun-spot triplet Fe 6173.553, 

 intensity 5. Thus the displacements due to the general field appear to 

 be confined to the lower levels of the solar atmosphere, in harmony with 

 our earliest results. 



As the levels corresponding to these lines can be found from eclipse 

 data, it should be possible to plot field-strengths as a function of level. 

 We have recently discovered, however, that the observed displacements 

 depend in some way upon the intensity and character of a line, so that 

 an obscure source of error must be eliminated before the change of field 

 strength with level can be determined. 



The hypothesis of local whirls, in which the innumerable pores 

 scattered over the sun are supposed to act like sun-spots, thus giving 

 rise to the general magnetic field, was briefly mentioned in Mount 

 Wilson Contribution No. 71. While this hypothesis might be able to 

 account for certain of the observed phenomena, the weak points named 

 in the above paper have been accentuated by the results given below, 

 which show that sun-spots are apparently not of the same polarity 

 throughout the northern hemisphere or the southern. If the pores 

 follow the same rule, their polarity in low and high latitudes would 

 presumably be opposite, and the observed Zeeman displacements could 

 not be accounted for on the local whirl hj^pothesis. 



THE POLARITY OF SUN-SPOTS. 



The magnetic polarity of a sun-spot is easily ascertained by deter- 

 mining whether the red or the violet edge of a widened line is trans- 

 mitted by a Nicol prism and quarter-wave plate. In our earlier work it 

 was found that spots of opposite polarity occurred in the same hemi- 

 sphere, and this result was confirmed by Evershed's and St. John's 

 observations of radial motion. The reasonable expectation that the 



