MOUNT WILSON OBSERVATORY. 233 



Bracket! in a number of the best spot triplets. The results indicate 

 that while there may be very appreciable differences of behavior in 

 different triplets, the shift of the p-component is in general smallest 

 near the center of the spot, and approaches a maximum near the outer 

 edge of the penumbra, where the field is very weak. Thus, both in 

 sign and magnitude of displacement, the p- and n-components behave 

 in opposite ways. The investigation is being followed up, both on the 

 solar and laboratory sides. 



In this connection we may recall the existence of other anomalies 

 in the Zeeman effect, as observed in the sun-spot spectrum. A notable 

 instance is that of the vanadium line X6 11 1.872, which Mr. Babcock's 

 photographs show to be a quadruplet in the spectrum of a spark be- 

 tween the poles of a magnet, with the p- and n-components exactly 

 superposed. In sun-spots this line appears to have a more complex 

 structure, but the separation of the components is not sufficiently com- 

 plete to permit their character to be determined with certainty. Other 

 interesting cases, soon to be described in detail, might also be mentioned 

 as valuable guides to further laboratory research. When viewed in 

 contrast with the hne-for-line agreement, in the vast majority of 

 instances, of the Zeeman effect in sun-spots and in the laboratory, 

 these few exceptions assume special interest and will warrant careful 

 study by physicists. 



INCLINATION OF THE SUN'S MAGNETIC AXIS. 



The first series of observations for the location of the sun's magnetic 

 axis, made on 74 days between June 8 and September 25, 1914, have 

 now been completely reduced for the determination of mean values of 

 the magnetic elements. The results are based on the assumption 

 that the sun is a uniformly magnetized sphere. The measures, as in 

 previous years, have been made by Mr. van Maanen. The reductions 

 by Miss Wolfe, carried out under the direction of Mr. Scares, have 

 given the following results : 



i = 6.0±0?4 P = 3L52±0.28days. 



f = 1914, June 25.38±0.42 days, G. M. T. 



fc = 0.99 (a constant inversely proportional to the polar field-strength). 



The probable error of the period of revolution of the magnetic axis 

 about the sun's axis of rotation is naturally large because of the short 

 interval covered by the observations. A comparison of these results 

 with a short series of observations in September 1916 indicates, how- 

 ever, that the period itself is near the true value. The complete re- 

 duction of the 1916 series, which is now in progress, should reduce the 

 uncertainty in the period well below a tenth of a day and make it 

 possible to carry the longitude of the pole forward, without ambiguity 

 as to the number of revolutions, to the coming sun-spot minimum, 

 when further observations can be undertaken without risk of inter- 

 ference from the magnetic fields of the spots. 



