MOUNT WILSON OBSERVATORY. 225 



lets and quadruplets rendered visible with the aid of a Nicol prism 

 and compound quarter-wave plate. Copies of the map, in 54 sections 

 of 50 angstroms each, will be placed on sale at cost. 



In the examination of sun-spot hypotheses the most important 

 evidence based on the Zeeman effect is probably that which indicates 

 the decrease in field-strength observed at increasing elevations. This 

 is plainly shown on the map, which also reveals other phenomena that 

 will probably prove significant. Most of the triplets show maximum 

 field-strength near the center of the spot, but others, representing higher 

 levels, indicate no increase from the pemunbra across the umbra. 



A peculiarity of the central (p) component of spot triplets, mentioned 

 in the last annual report, has been the subject of preliminary studies by 

 Mr. Brackett and the Director, though the continuation of the investi- 

 gation has been postponed pending the completion of a new solenoid 

 magnet, inclosing a special electric furnace designed for the observa- 

 tion of the inverse Zeeman effect at various angles with the lines of 

 force. A negative lens was mounted near the focus of the 150-foot 

 tower telescope for the purpose of magnifjang the image of a spot 1.4 

 diameters. With this degree of enlargement, a compound quarter- 

 wave plate of 1-mm. strips, used with Nicol prism over the slit of the 

 75-foot spectrograph, permitted a large number of displacements of 

 the p-component to be measured at regular intervals across typical 

 spots. The results, in harmony with those mentioned last year, show 

 that the p-component behaves as though it were composed of two 

 members, circularly or elliptically polarized in opposite directions, and 

 more widely separated near the outer edge of the penumbra, where the 

 n-components indicate a weaker field, than in the strong field near the 

 center of the spot. This remarkable behavior of the p-component, 

 called for neither by theory nor by laboratory experience, deserves 

 special attention in future investigations of the Zeeman effect. 



PERIOD OF REVOLUTION OF THE SUN'S MAGNETIC AXIS. 



In the last annual report there was reference to a short series of ob- 

 servations of the sun's general magnetic field made in September 1916. 

 IVIeasures of the plates taken on 7 days indicated that the period of 

 revolution of the magnetic axis about the sun's axis of rotation, as given 

 by the 1914 plates, was near the true value. During the past year Mr. 

 van Maanen has measured 479 plates of the 1916 series, completing 22 

 daj^s between September 2 and September 28. The final reduction of 

 these measm-es has not yet been completed, but the present results 

 show that the period of 31.52 ± 0.28 days given in the last annual 

 report is very near to the truth. The difference for the 26 periods 

 between the 1914 series and that of 1916 is not quite 3 days, or less than 

 0.12 day per period, which is well within the limits given by the probable 

 error. 



