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



to that point any careful study Of spectra had been impossible 

 for lack of definite standards of reference, and because the 

 apparatus was as yet very defective. Fraunhofer's research, 

 " Bestimmung des Brechungs- und Farbenzerstreunngs- 

 Verniogens verschiedener Glasarten," was presented to the 

 Munich Academy of Sciences in 1814, and was published in 

 the fifth volume of the " Denksckrif ten." It then became 

 possible to study in detail the properties of rays of definite 

 position and the work was taken up almost immediately. 

 Almost the first step was to determine the wave-lengths of 

 prominent points in the solar spectrum, and, as is well known, 

 Fraunhofer himself took it, determining the wave-lengths 

 corresponding to his lines B, C, D, E, F, G, H. As there 

 seems to have been — noticeably in Yerdet's papers — some 

 confusion concerning his papers on this subject, it may be well 

 here to clear the matter up. 



Fraunhofer's first paper dealing with the subject was 

 presented to the Munich Academy in 1821. It is entitled: 

 "Neue Modification des Lichtes durch gegenseitige Einwir- 

 kung und Beugung der Strahlen, und Gesetze derselben," and 

 was printed in the eighth volume of the " Denkschriften." 

 It is of considerable length and deals with various diffraction 

 phenomena, but its chief interest lies in the wave length 

 measurements made with wire gratings. The experiments 

 made with ten of these are given in detail and are remarkably 

 careful and consistent. The gratings were quite various, the 

 wires being from O04 mm to 0-6 mm in thickness and the grating 

 space as ordinarily measured, from 0-0528 to 06866 mm . From 

 these proportions it is evident enough that the spectra must 

 have been very imperfect, but in spite of this, Fraunhofer 

 obtained results which agreed remarkably well with each 

 other. The wave lengths of D as obtained from the above 

 mentioned ten gratings were as follows : reduced to millimeters. 



mm . mm 



(1) 0-0005891 (6) 0-0005888 



(2) 0-0005894 (7) '0-0005885 



(3) 0-0005891 (8) 0-0005885 



(4) 0-0005897 (9) 0-0005882 



(5) 0-0005885 (10) 0*0005882 



The mean value adopted was 0-0005888 mm , which considering 

 the gratings and the fact that most of the angles of deviation 

 were less than 1°, is certainly remarkably accurate. It should 

 be noted too, that the finer gratings (1) to (4) gave even better 

 results. 



A brief discussion of this paper appeared in the seventy- 

 third volume of Gilbert's Annalen and a French reprint in 

 Schumacher's Astronomische Abhandlungen (ii, 46). 



