720 
PROFESSOR J. N. LOCKrER ON THE PHOTOGRAPHIC 
involved in uncondensed nebulous matter. Hence, another possible condition in 
tlie hottest stars is satisfied l)y the photographic facts at our disposal. 
It will be seen then, that these considerations of the conditions of increasing tempe¬ 
rature demanded by the hypothesis, have enabled us to determine that the series of 
spectra represented in Plates 26 and 27 is in all probability a series in ascending order 
of temperature. All the phenomena we should expect, on the hypothesis, are met 
with among the spectra of celestial bodies. 
We have next to consider the phenomena connected with stars of decreasing 
temperature. 
Stars of Decreasing Temperature. 
(Stage 1.)—We have seen that with the failure of the supply of meteorites falling 
into the centre of the now vaporized mass, cooling will commence, and the longest 
lines in the spectra of the various chemical elements should make their appearance. 
This condition is met with in the stars of Table A, Sub-division 8. The following 
Table will show that this is true in the case of the iron lines, if we take the spectrum 
of Sirius as a type. 
Some of the Iron Lines in the Spectrum of Sirius. 
4045 
4063 
4071 
4215 
4298 
4307 
4383 
4404 
4414 
(Stage 2.)—The conditions at this stage of cooling are satisfied by the stars of 
Table A, Sub-division In these stars, as already pointed out (p. 696), we get, in 
addition to fairly broad lines of hydrogen, nearly all the lines which appear in the 
solar spectrum, and these it is well known agree in the main with the arc spectra ot 
the various chemical elements. 
(Stage 3.)—The stars of Table B, Sub-division /3, represent the conditions which 
are required by the liypothesis at this stage of cooling. The metallic line absorption 
is again at a maximum, and we find the lines of the various chemical elements 
similar to those seen at Stage 2 of the ascending series, but with difierent intensities 
and with different amounts of continuous absorption at the violet end of tlie spectrum. 
This difference, so far as the known lines are concerned, I have already pointed out 
will be due to a different percentage composition of the absorbing mass of vapour 
(p. 713). 
