present chapter we propose to give an account of what the spectroscope
tells us about the physical constitution of the fixed stars.

Quite a new phase of astronomy is here opened up. Every improvement in
telescopes revealed fainter and fainter objects, but all the telescopes
in the world could not answer the question as to whether iron and other
elements are to be found in the stars. The ordinary star is a mighty
glowing globe, hotter than a Bessemer converter or a Siemens furnace; if
iron is in the star, it must be not only white-hot and molten, but
actually converted into vapour. But the vapour of iron is not visible in
the telescope. How would you recognise it? How would you know if it
commingled with the vapour of many other metals or other substances? It
is, in truth, a delicate piece of analysis to discriminate iron in the
glowing atmosphere of a star. But the spectroscope is adequate to the
task, and it renders its analysis with an amount of evidence that is
absolutely convincing.

That the spectra of the moon and planets are practically nothing but
faint reproductions of the spectrum of the sun was discovered by the
great German optician Fraunhofer about the year 1816. By placing a prism
in front of the object glass of a small theodolite (an instrument used
for geodetic measurements) he was able to ascertain that Venus and Mars
showed the same spectrum as the sun, while Sirius gave a very different
one. This important observation encouraged him to procure better
instrumental means with which to continue the work, and he succeeded in
distinguishing the chief characteristics of the various types of stellar
spectra. The form of instrument which Fraunhofer adopted for this work,
in which the prism was placed outside the object glass of the telescope,
has not been much used until within the last few years, owing to the
difficulty of obtaining prisms of large dimensions (for it is obvious
that the prism ought to be as large as the object glass if the full
power of the latter is to be made use of), but this is the simplest form
of spectroscope for observing spectra of objects of no sensible angular
diameter, like the fixed stars. The parallel rays from the stars are
dispersed by the prism into a spectrum, and this is viewed by means of
the telescope. But as the image of the star in the telescope is nothing
but a luminous point, its spectrum will be merely a line in which it
would not be possible to distinguish any lin