Physics. 415 
of burning sodium, which in itself gives a continuous spectrum, passing 
through the portion of the flame ignited at the edges in contact with the 
air and of lower temperature, or still more probably through vaporized 
metallic sodium which has escaped combustion, will reverse the brilliant 
line D, and thus give a dark line upon a continuous bright spectrum, 
this explanation, upon further knowledge of the facts, should prove 
correct, it will not be necessary, with Kirchhoff, to suppose that the solid 
y of the sun is ignited or luminous. For the temperature of the pho- 
tosphere may reasonably be supposed to be highest nearest the surface of 
the body of the sun, since there the condensation is greatest. Those 
layers or strata nearest the sun will then give continuous spectra, and 
the rays from there passing through the outer strata will give spectra 
tion by dividing the heat of combustion by the specific heat. According 
to Favre and Silbermann, we have for the heat of combustion the num- 
ber 3195, which, divided by 0°5, gives for the temperature of combustion 
0° C., w i 
under a constant pressure. But if the specific heat of soda, NaO, in the 
form of vapor, be taken as one-fourth of that of water, we shall have 
for the temperature of combustion 12780° C. It appears by no means 
. 
AB, he. ip. Ag. Ar. ha. dn. 
6-878 6-564 5'888 5°260 4°843 4°291 3°928 
_ €xpressed in hundred-thousandths of a centimeter. The employment of 
_ the three hydrogen lines, «, 8, 7, permits the observations to be made at 
I times and with the greatest facility and accuracy. Landolt furthe 
determined the indices of each subsiance for a series of temperatures, 
C 
ter was in each case plunged directly into the liquid; the prism and 
: liquid were heated to 30° C, and allowed to cool slowly, the observations 
made from degree to degree. The method of measurement em- 
