2 W. Gibbs on a normal Map of the Solar Spectrum. 
: e diffraction spectrum as obtained by a glass plate ruled 
= with fort one to ten thousand lines to the inch gives a nearer 
approach to a perfectly normal spectrum than any which has 
_ hitherto been proposed, and has been employed by Fraunhofer,* 
_ Mossottit and Draper.{ But even in this case there is a position 
Of least deviation for each line depending upon the angle which 
the incident beam of light makes with a normal to the plane of 
the ruled surface, and in addition, it is necessary either to fix 
_ meter of glass surface, or to reduce the positions of the spectral 
oy 2: foo which would have been obtained with a standard 
We pee all these difficulties by constructing a normal spec- 
trum n such a manner that each spectral line shall be entered 
: Recordings to its wave length, a method first suggested, I believe, 
by Billet.§ 
The materials which the neat condition of science furnishes 
for the construction of a chart of the spectrum upon the princi- 
ple laid down are not very copious. They consist essentially of 
the older measurements of Fraunhofer|| oF the wave lengths of 
is the lines B, C, D, E, b, F, Gand H; of recent and very accurate 
measurements by An ostrém ; io a still more extended series by 
Ditscheiner ;** of the measurements of Pliicker ++ for certain 
lines in the spectra of gaseous media, and of those us rips 
for the ultra-violet rays. In addition we have a small number 
se eae measurements by Miiller,$§ Mascart,|||| Stefan, Pant “ed 
ette 
For the construction of the chart which I present herewith to 
the Academy I have selected the wave lengths determined by | 
Angstrém as standards. This has been done because both the — 
ruled glass and the measuring apparatus employed by him ap- 
pear to have been much more > perfec cet than those of other observ- 
ers—Fraunhofer not excepted. On the other hand, however, all 
of Angstréi’s measures are not available for the reason that ‘the 
identification of the lines as produced by particular elements 
does not appear to be in all cases absolutely certain. As Ang- 
strom’s measurements are in ten-millionths of a Paris inch they 
have been reduced to millionths of a millimeter by multiplying 
them byt 2 constant 27-07. 
Ditscheiner’s measurements agree closely with those of Ang- 
* Deakschr. ag oot ead, Bd. viii, 1821-22. 
¢ Annali delle Uni er. Toscane, t. i, 1845; also Pose. Ap. lxii, p. 509. 
i L. & E. P til, Mage. Xxvi, ans 
Traite phe > PSs p- mc exxiii, p. 489. x 
*#* Sitzungsberich' alae: lee Wiss, Bat e s . 
_ +t Pogg. Ann., oe den ¢ Pogg. Ann., xeviii, p. 513. 
§§ Pogg. Ann., czviii, YD. anh a exxix, p. 637. i Pgs. Ann., exviii, 367. 
[% Pogg. Ann., exxii, p. 6 Pogg. Ann, exxiv, p. 390. 
