290 CARNEGIE INSTITUTION OF WASHINGTON. 



out as follows: 53 within ±0.001 a, 8 within =t 0.002a, 2 within 

 ±0.003 A, and one within 0.004a of the adopted values, while in the 

 case of the 16 standards belonging to groups c5 and d, the international 

 values are systematically greater in the mean by 0.007 a, due to the 

 large pole effect in the arc used in the original determination. Though 

 the wave-lengths given by the 6-mm. 6-ampere arc may be employed 

 as standards in cases where the highest accuracy is not required, neither 

 these nor the standard wave-lengths of similar lines can be used for 

 determining the differences between their solar and terrestrial wave- 

 lengths when both are expressed in the international system, since solar 

 lines are produced under conditions in which pole-effect is not involved. 

 With the exception of further work in the violet and red, and 

 determinations for certain iron lines corresponding to faint solar lines, 

 the program on iron-arc wave-lengths has been concluded. A wide 

 range in wave-length and in line-intensity is necessary for our investi- 

 gation of the Einstein effect and for separating the causes contributing 

 to the shift of solar lines. 



PRESSURE DISPLACEMENTS OF IRON LINES. 



The investigation of the pressure effect for certain iron lines by 

 Mr. Babcock promises results of considerable interest. Previous 

 studies of this subject have been concerned for the most part with 

 large differences of pressure and have yielded data of great value. But 

 the influence of pole-effect on such observations and the fact that with 

 increasing pressure the spectral lines are not only displaced but are 

 widened make it highly desirable to attack the problem anew, using a 

 small difference of pressure and taking care to avoid the presence of 

 pole effect as far as possible. The minuteness of the change in wave- 

 length to be observed under these conditions requires a method quite 

 free from the effects of thermal and mechanical disturbances. The 

 Fabry-Perot interferometer has been found to meet these exact- 

 ing requirements when used by what may be called the differ- 

 ential method. Simultaneously with each exposure to the iron 

 arc the light from a Cooper-Hewitt mercury lamp is passed through 

 the interferometer and photographed upon the same plate. The 

 optical arrangements are chosen so that full illumination is secured 

 for from 15 to 30 interference rings upon each spectral line. For two 

 plates taken at different pressures the difference in wave-length for 

 any spectral line is given by each pair of homologous rings, the rings 

 on a single mercury line serving to eliminate all instrumental changes. 

 On account of the large number of rings available on each line, a single 

 exposure acquires very high weight. The accuracy of the method is 

 well illustrated by numerous groups of observations taken in sets of 

 three instead of in pairs. When the three photographs of such a set 

 are intercompared, three differences of wave-length are obtained 



