268 L.Bell — Absolute Wave-length of Light. 



justment with a high degree of accuracy, and keeping it 

 through a series of measurements, the gain is by no means 

 considerable. Aside from this question, Mascart's spectrom- 

 eter read only to five seconds, and while his results with differ- 

 ent gratings agree very well individually they are certainly col- 

 lectively in error by quite a large amount, very possibly owing 

 to bad standards of length. 



It is a fact to be noted in discussing all these earlier wave- 

 length determinations, that sufficient attention was not paid to 

 the measurement and study of the gratings — by all odds the 

 most difficult part of the problem. The angular measures of 

 any one of the above investigators were good enough to have 

 given very exact results had they been combined with proper 

 investigations of the grating spaces. As most of Nobert's 

 gratings were small and by no means accurately ruled there 

 was peculiar need of care in measuring them, and when one 

 considers that the defining lines on most standards of length 

 are far from being good, it is clear that the chances for error 

 were numerous. In Angstrom's first paper he even relied on 

 the grating space assigned by the maker. Ditscheiner em- 

 ployed a grating which had belonged to Fraunhofer himself, 

 but the number of spaces was uncertain and this led to a large 

 error which he corrected, in part, in a supplementary paper 

 some years later. Ditscheiner' s principal paper was published 

 in 1866, and was followed in 1868 by an elaborate discussion 

 of the whole problem by van der "Willigen, whose paper is 

 valuable mainly for a particularly elaborate review of sources 

 of error. Like his predecessors he used ISTobert's gratings, but 

 as the construction of his spectrometer confined his angular 

 measurements to the deviation on one side of the normal, their 

 accuracy may be open to some question, while his standard of 

 length was anything but reliable, as it was a glass scale only 

 three centimeters long and the only assurance of its accuracy 

 was the certificate of the maker that it was " tres exacte " at 50° 

 Centigrade. For one or both of the above reasons van der 

 "Willigen' s results were larger than any which have been ob- 

 tained, before or since his time. 



In the same year appeared Angstrom's great research which 

 has so long served as the standard in all questions of wave 

 length. It is hard to say too much of the conscientious and 

 painstaking experiments on which his results were based, and 

 any want of accuracy in the final result was due to no lack of 

 skill or care on his part but rather to the imperfect instruments 

 with which he was obliged to • work. Like every one before 

 him he used Robert's gratings and in spite of the fact that like 

 all ISTobert's gratings they gave very imperfect definition and 

 showed numerous " ghosts," his results were more than usually 



