MOUNT WILSON OBSERVATORY. 289 



cised simultaneously with that of the magnetic field. For this reason, 

 and in view of the important bearing of the combined effect on theories 

 of the structure of the atom, an extensive study of this problem seems 

 to be warranted. 



A powerful solenoid magnet, designed by Mr. Anderson for use with 

 currents up to 4,000 amperes, offered within its hollow core precisely 

 such conditions as our needs demanded. In the uniform field of 33,000 

 gausses a vacuum, tube of pjTex glass, giving a strong electric field 

 just in front of the cathode, within the limits of the Crookes dark 

 space, was inserted. The discharge was produced in hydrogen by a 

 small transformer, used in conjunction with a mechanical rectifier, 

 which gave a high-voltage current sufficiently steady for preliminary 

 purposes. Under these conditions (lines of electric and magnetic 

 force parallel) the appearance of the Hy line was nearlj^ the same in 

 the electric field as in the combined field, though in the latter case the 

 components were more diffuse. It remains to be seen whether the 

 components will show further resolution under higher dispersion and 

 in a steadier electric field. Some of the lines of the secondarj^ spectrum 

 indicate distinct differences in intensity in the electric and the com- 

 bined fields. The investigation is being continued with a special form 

 of quartz tube, designed for heavy discharges. Improved apparatus 

 for high-potential direct current has also been provided. 



Another mode of approach to the problem of electric fields in sun- 

 spots is afforded by Saha's ionization theory, which should permit an 

 upper limit to be set for any field that may be present. Available 

 data, such as the weakening of the enhanced lines, indicate that if an 

 electric field exists there it must be weak. 



WAVE-LENGTHS OF IRON-ARC LINES FROM X3370 TO X6759 MEASURED WITH 

 GRATING AND INTERFEROMETER. 



A list of 1,026 iron-arc lines has been published by Messrs. St. John 

 and Babcock, of which 976 were measured on from 1 to 62 grating 

 spectrograms and 576 on from 1 to 39 interferometer plates. The 

 agreement shown indicates that for most of the lines the weighted 

 mean wave-length is accurate to 0.001 a. To avoid errors due to 

 pole effect and to obtain sharp lines, a 12-nmi., 5-ampere Pfund arc 

 was used. To eliminate instrumental errors, spectrograms were made 

 with two Michelson and three Anderson gratings and with four pairs 

 of interferometer plates. 



In the case of stable lines the agreement with the results obtained at 

 the Bureau of Standards is good, but for Imes of groups c5 and dl the 

 Bureau of Standards measurements are systematically greater. The 

 mean difference for 46 lines is +0.007 a, due to the large pole effect in 

 the center of the 6-mm., 6-amp. arc. Of 78 International Secondary 

 Standards measured with the interferometer, the 62 stable lines came 



