Ze"eman Effect in Satellites of Mercury Lines. 335 



examined by adjusting the positions of interference points 

 with respect to the principal line. Moreover, it was easy to 

 detect overlapped lines, which with a single apparatus would 

 have been impossible. On this account, the Fabry-Perot 

 interferometer was preferable to the Lummer-Gehrcke plate 

 in the present experiment. The method was, however, inap- 

 plicable to weak lines, owing to loss of light in silver films, 

 so that some may be missing in the crossed spectra. 



The mercury lamp consisted of a vacuum capillary tube, 

 whose ends were connected with wide tubes partly filled 

 with mercury. The electrodes were sealed in them and 

 came in contact with mercury. The tube was placed ver- 

 tically between the poles of an electromagnet, and was 

 excited by means of an induction-coil, whose primary 

 was run by a single-phase current of 60 cycles per second, 

 the mean strength varying from 3 to 6 amperes. In order 

 to obtain a good image of the satellites, it was necessary to 

 place in parallel a small air-condenser, which was connected 

 with a coil of adjustable self-inductance. The electromagnet 

 was specially designed to suit the purpose of investigating 

 the Zeeman effect, having soft iron cores of 12 cm. diam. 

 and coils of 3200 turns. For examining the transverse effect 

 in uniform field up to 22 kilogauss, conical pole-pieces having 

 vertical angle of about 80° were used, the air-gap being 

 6 mm. and the diameter of the end faces 20 mm. Special 

 arrangement was provided for preventing the air-gap 

 becoming wedge-shaped in strong fields, as is often observed 

 in some electromagnets. For fields higher than 22 kilo- 

 gauss, pole-pieces having vertical angle of 112° and end- 

 faces of 4 mm. diam. were used, with air-gap of 4 mm. 



The light from the mercury lamp in the air-gap was 

 concentrated by two lenses on the slit of a collimator, 

 and after prismatic resolution was projected on a slit 

 placed in the focus of another collimator. It then traversed 

 the echelon, and after passing through another lens was 

 made parallel by a Zeiss microplanar of 5 cm. focal length, 

 and then photographed by an anastigmatic objective bv 

 Goerz. A Wollaston prism was interposed between the 

 microplanar and the photographic objective, so that the 

 components parallel and perpendicular to the field could be 

 photographed side by side, or one above the other, Im- 

 proper adjustment. Following German authors, we shall 

 hereafter denote these components by the letters p and s. 



For photographing the green line, we used Agfa chromo- 

 isolar plates and spectrum-panchromatic plates of Wratten 

 and Wainwright ; and for the violet line, process-plates. 



