704 
PEOFESSOR J. N, LOCKTER OR THE PHOTOGRAPHIC 
important iron-absorbing region in a Orionis is nearer that of the oxy-coal gas flame 
than that of the electric arc. Still, as many of the arc lines appear to coincide with 
lines in the spectrum of the star, though they have diflerent relative intensities, it is 
probable that the average temperature is intermediate between that of the arc and 
that of the flame, but nearer to the latter. 
It is evident then that the flame spectra of other substances should also be 
compared with the spectrum of the star. Those which show any coincidences are 
indicated on the map. They include manganese, chromimn, calcium, magnesium, 
strontium, cobalt, and lithium. Here, again, the arc spectra, especially in the case of 
manganese, show many inversions, and hence the conclusion that the temperature of 
the absorbing regions is lower than that of the electric arc is strengthened. 
The results of the investigations into the origins of the lines are shown in Plate 28. 
Elements identified in a Orionis. 
H 
Mg 
Fe 
Co 
Hn 
Sr 
Cr 
Li 
Ca 
C 
Many of the lines remain for the present unidentified, although careful comparisons 
have been made with the flame spectra of all the principal elements. 
Amongst the strongest unknown lines are those at wave-lengths 4393, 4406, and 
4763. There are also three very strong lines at X 4038, 4053, and 4164. These are 
either entirely absent from the solar spectrum or only appear as very faint lines. 
That at X 4406 forms a close double with the slightly more refrangible iron line 
X 4405. It is strongly marked in the spectrum of a Tauri, as photographed at 
Kensington, and also aj:)pears in the spectra of ^ Andromedas, a Herculis, /3 Pegasi, 
and other stars of Table C. 
Bright Flutings. 
In the spectra of a Orionis and a Herculis which have been photographed at 
Kensington on isochromatic plates there is the bright edge of a fluting at X 5165 as 
near as can be determined, and this very closely agrees with the wave-length of the 
brightest edge of the carbon fluting X 5164'8, according to measures made from my 
large scale photographic solar comparisons. If this were the only criterion for bright 
carbon, the evidence might be considered conclusive. There are, however, two 
other bright maxima in this compound carbon fluting, at wave-lengths 5128'5 and 
5098'3, and we should also expect to find these in the spectrum. In a Herculis, two 
secondary bright flutings are clearly visible, and the following table will show their 
positions relatively to the carbon flutings. 
