20 
The Astronomical Survey of the Universe [oh. i 
main structure of the universe. In rushing in imagination to the depths 
of space, travelling with millions of millions of times the velocity of light, 
we have not paused to discuss, or even to mention, except perhaps incidentally, 
the various minor classes of objects which are found scattered through the 
universe. Retracing our steps, let us discuss astronomical objects no longer 
primarily in respect of their distances, but rather in respect of their frequency 
of occurrence in the sky. 
Binary Stars. 
19 . The commonest object of all is the simple star, which appears merely 
as a point of light and shines with a perfectly steady light. If we set a 
telescope on the sky at random, we shall find that the vast majority of the 
objects visible in it are simple stars of this kind. We are likely to find, how 
ever, that these points of light are not scattered at random in the field. In 
too large a number of cases to be attributed to mere chance, pairs of stars are 
found to lie very close together in the sky. Some such pairs undoubtedly appear 
close as the result of mere chance, the line joining them happening to pass 
near to the earth at the present moment. If such a pair of stars were watched 
for several centuries, they would be seen to move steadily apart, and the 
chance cause of their appearing together would be disclosed ; they had merely 
happened to lie almost one behind the other when we first noticed them. 
Other pairs of contiguous stars shew no tendency to separate when con 
tinuously watched; on the contrary, they are observed perpetually to describe 
orbits one about the other like partners in a never-ending waltz. Such pairs 
clearly do not appear close together in the sky as the result of mere chance, 
but because they actually are close together in space; they are bound 
together by the force of gravitation and describe regular orbits about one 
another just as the earth describes an orbit about the sun, and for the same 
reason. Such systems are spoken of as “ binary systems.” 
The periods in which the orbits of binary stars are described may be too 
long to be determined by observation at all. Many must be reckoned in 
thousands of years, and periods of hundreds of years are common. At the 
other end of the scale, when the period is less than about a year, the stars 
may be so close as to appear merely as a single point of light in the telescope, 
so that visual observation cannot detect their binary nature. Methods other 
than direct visual observation, however, reveal the existence of binary stars of 
periods extending down to only a few hours. 
20 . Spectroscopic Binaries. The spectrum of a star generally shews a 
number of sharply-defined dark lines, and if the star is in rapid motion, these 
lines are observed to be displaced from their normal places by an amount 
which, by Doppler’s principle (cf. p. 50 below), is proportional to the star’s 
velocity relative to the earth. This extremely valuable circumstance makes it 
possible, within limits set by the power of optical instruments, to determine