Wave-length of Light. 247 



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

 Schumacher's Astronomisehe Abhandlungen (ii. p. 46). 



Fraunhofer's second and more complete paper appeared in 

 1823 in Gilbert's Annalen (lxxiv. p. 337). Its title is, " Kurzer 

 Bericht von den Eesultaten neuerer Versuche tiber die Gesetze 

 des Lichts, und die Theorie derselben." This paper gives a 

 detailed account of his experiments with two glass gratings. 

 Of these, the grating-spaces were respectively 0*0033 and 

 0"0160 millim. The former was apparently much the better, 

 and upon it Fraunhofer based his final result, which for D 

 was 0*0005886 millim., while the experiments with the coarser 

 grating gave 0'0005890 millim. These values apply quite 

 certainly to the mean of the two D lines, and not, as has been 

 sometimes supposed, to one of them alone. 



The experimental work with these glass gratings was much 

 better than with the previous wire ones, since the angular de- 

 flections were very much larger, and the gratings themselves 

 were susceptible of far more exact measurement. But at best 

 they were but indifferent instruments, and the terminal lines 

 were so bad that they had to be retraced before the grating- 

 space could be determined. So, between poor gratings and 

 indifferent standards of length, Fraunhofer's determination of 

 absolute wave-length left very much to be desired. However, 

 nothing much better could be accomplished until the art of 

 making gratings was very much improved ; and it was not 

 until Robert's gratings became tolerably well known that any 

 serious attempt was made to improve on Fraunhofer's results. 

 From time to time various investigators worked at the 

 problem, both with Robert's earlier gratings and by utilizing 

 various interference phenomena. When, however, the great 

 investigations of Bunsen and Kirchhoff revolutionized spec- 

 troscopic work and emphasized its great importance, the 

 attention of scientific men was called to the need for accurate 

 measurements; and for half a dozen years investigators were 

 active, and Mascart, Ditscheiner, and Angstrom appeared on 

 the field almost simultaneous^. Each published a paper in 

 1864, and of these that of Mascart is probably the most accu- 

 rate and painstaking, though now it is quite certain that the 

 values he obtained were considerably too small. He employed 

 four or five of Nobert's gratings, and instead of placing the 

 grating perpendicular to either the collimator or the observing- 

 telescope, used it in the position of minimum deviation, that 

 is to say, so that the plane of the grating should bisect the 

 angle formed by the incident and diffracted rays. This 

 position has certain advantages ; but as the experimentation 

 is rather more difficult than in the ordinary position, the 

 S2 



