4 PROCEEDINGS OP THE AMERICAN ACADEMY 



and set to any required wave-length upon a convenient scale. The 

 photographic plate is then placed in the camera, and a shutter imme- 

 diately in front is set to expose the upper half of the plate. Expos- 

 ure for the sun is then made ; the sunlight is then cut out, and the 

 shutter moved to cover that part of the plate already exposed, and 

 the lower half exposed. The spark is then started and worked from 

 15 to 30 minutes. In addition to the spectrum of lines there is a con- 

 siderable continuous spectrum, which after a time causes fogging of 

 the plates ; so there does not seem to be any gain in an exposure of 

 more than half an hour. The feebleness of the air lines can be 

 judged of when we state that, with the same plate, breadth of slit, 

 etc., we get a metallic spectrum in the arc in ten seconds, strongly 

 photographed. There was sufficient iron present in the electrodes as 

 impurity to give the strongest iron lines feebly, and these have been 

 of use in determining that no displacement had happened, although 

 from the nature of the arrangements such disturbance could hardly 

 occur. 



On the negative produced as above indicated the two spectra lie 

 exactly edge to edge, like a vernier and scale, and are in the best 

 possible position for the accurate determination of the position of 

 the air lines. The original plan contemplated a determination of 

 wave-lengths of all the air lines throughout the entire spectrum ; but 

 persistently bad weather and other causes have compelled the post- 

 ponement of the completion of this work, though v?^ are now able to 

 give it complete from wave-length 3740 to wave-length 5030. 



The photographic map of the solar spectrum of Professor Rowland 

 has made easy what would otherwise have been an undertaking of 

 extreme labor and difficulty. The best of engraved maps of the violet 

 region of the spectrum to beyond F are comparatively worthless. 

 Even on the elaborate map of Vogel, the result of years of labor, it 

 is difficult certainly to recognize other than the more prominent lines, 

 and you never feel quite sure of your positions ; but we turn to the 

 map of Rowland with the certainty of finding every line in its true 

 order and magnitude, so that what was formerly most difficult has 

 now become very simple, and the position of any well-defined air or 

 metallic line can be read directly, by comparison of the photograph 

 with the map, to the tenth of a wave-length. 



We here give a table of wave-lengths as determined from our photo- 

 graph of the sun and air spectra : — 



