THE RIGIDITY OF THE EARTH 587 



Other interference arrangements are obvious which would give 

 a displacement of a greater number of fringes, or permit the use of 

 a shorter pipe; e.g., the fringes formed between the water surface 

 and the lower mirror might be used, or the lower mirror might be 

 dispensed with and use made of the fringes formed by the light 

 reflected from the water surface and the vertical mirror. But the 

 arrangement actually used was the most satisfactory, since the 

 long pipes, 502 feet, were already installed. 



The sources of Hght were commercial alternating-current Cooper- 

 Hewitt mercury lamps. They proved very rehable and satis- 

 factory. By using as filters thin films of a saturated solution of 

 escuHn in water, all wave-lengths from the arc shorter than X4358 

 were absorbed, and the positive film used was not sensitive to the 

 longer wave-lengths. The exposed portions of the films were 

 removed and developed each week. The light was abundantly 

 strong for satisfactory negatives, and it was possible to use i . 5 mm 

 diaphragms on the projecting lenses. This gave sufficient sharpness 

 to the fringes, even when there was a considerable change in their 

 focus. It was necessary to readjust and refocus the fringes in only 

 one pit during the entire year, although the width of the fringes 

 was altered once or twice in two other pits. One of the mirrors 

 required resilvering. One of the pits ran throughout the year 

 without readjustment of the fringes or camera. The pits and 

 cameras were in charge of Mr. George Monk and Mr. Frank 

 Sullivan, of the Yerkes Observatory staff. 



A relay which moved a shutter in front of the projecting lens 

 was placed in each pit. The four relays were connected in series 

 with a clock in the observatory, so that the time could be con- 

 veniently and accurately controlled. Once an hour the clock 

 made contact, and a storage-battery circuit was closed through the 

 relays and the light was cut off by the shutters for about 20 seconds. 

 Very accurate time-marks were secured in this way. The control 

 clock was kept six minutes faster than Central Standard Time in 

 order to simpKfy the computations and bring the observations 

 into agreement with them. (The longitude of Yerkes Observatory 

 from Greenwich, is s^ 54™ 13®.) 



