298 PROCEEDINGS OF THE AMERICAN ACADEMY 



ous observers in measuring wave-lengths of metallic spectra in the 

 ultra violet can be summed up as follows. 



1. "We believe that the method of photographing images of the slit 

 upon the photographic plate, due to Cornu, in order to determine 

 positions, leads to unavoidable errors. 



2. The best method of determining wave-lengths of metallic spec- 

 tra is by the use of concave gratings ; for linear measurements are 

 substituted for angular ones ; underlying spectra are brought to the 

 same focus as overlying spectra ; and, since a great number of lines 

 are in focus on the same plate, the conditions are the same for all, 

 viz. breadth of slit, length of exposure, and source of light. 



3. Hypotheses in regard to coincidences of gaseous and metallic 

 spectra cannot be safely based upon existing measurements of spectra 

 in the ultra violet. 



4. The limit of the copper lines is extended by our investigation. 



