ON STELLAR SPECTROMETRY. 167 
lines occasionally visible, but which are not characteristic of the type-form. 
Tn this type the red rays are very faint in proportion to the blue, violet, and 
green, so that the colour of the star tends to the blue hue, and occasionally 
to the green. Of this last kind is the group of the large constellation Orion 
and its neighbourhood. 
The second type is that of the yellow stars, as Capella, Pollux, Arcturus, 
Aldebaran, a Urse Majoris, &c. These stars have a spectrum exactly like 
that of our sun—that is, distinguished by very fine and numerous lines. 
These stars give occasionally a continuous spectrum, when the state of the 
atmosphere is not good; but in general the lines may be distinguished very 
easily. A fuller description is unnecessary, since the spectrum of the sun 
is very well known. The only thing which deserves particular attention is 
that in this class occasionally the magnesium lines are very strong, so as to 
produce very strong bands, and the iron lines in the green are in some very 
distinct. These stars can be distinguished even without the prism by the 
difference of colour, a rich yellow, which contrasts strongly with that of the 
first type. Stars of this second type are yery numerous, and embrace almost 
the other half of the stars. 
The third and very remarkable type is that of orange or reddish stars. 
These have as a prototype the stars a Herculis, a Orionis, Antares, o Ceti, 
 Pegasi. The spectra of these stars show a row of columns at least eight in 
number, which are formed by strong luminous bands alternating with darker 
ones, so arranged as to represent apparently a series of round pillars, closely 
resembling a colonnade. a Herculis is exceedingly remarkable in this respect ; 
the other stars are more or less clearly divided into pillars; but it is quite 
impossible to describe the beauty of the appearance which is visible in a 
telescope on a fine night. 
All the pillars are generally resolved more or less completely in different 
stars into smaller and finer lines, very sharp and clear. I have carefully 
drawn, after actual measurements, the spectrum of a Orionis and a Hereulis ; 
and in my memoir those of Antares and Aldebaran are given. In these stars 
some of the divisions of the pillars correspond to some principal lines of Fraun- 
hofer, as D and 0; but others, although very near, do not coincide with them, 
as Cand F. The presence of hydrogen, however, is certain, the lines C and 
F haying been found in the principal of them. 
The divisions of the pillars after many measurements have been found to 
agree perfectly in all these stars ; so that this type is very constant and well 
marked. In my catalogue 25 of these most interesting objects are registered ; 
and I do not imagine that I have exhausted the number. 
A very interesting feature connects this type with the preceding one. 
Here I must remark that we have to distinguish between lines and bands 
of shadow. The lines are strips narrow and sharp, the bands are shaded; 
although perhaps each band may be composed of very small lines, the 
aspect with our instruments (as at present constructed) is that of a more or less 
continuous shade. This shade is analogous to that which is produced by the 
vapour of our atmosphere in the spectrum of the sun when it is near the 
horizon. 
Now it is a very remarkable fact that these types seem to differ from one 
another not in the metallic lines, but in the nebulous bands. Thus, for in- 
stance, the spectrum of Arcturus and Aldebaran represent the same metallic 
lines as a Orionis, but this has bands in addition; the feature, however, is 
altogether so peculiar that a different type must be constituted. It is to be 
remarked also that all the pillars haye their luminous sides toward the red, 
