As regards the chemical composition of the Stars, it is, moreover, obvious that the powerful engine of investigation afforded us by the spectroscope is by no means confined to the substances which form part of our system. The incandescent body can thus be examined, no matter how great its distance, so long only as the light is strong enough. That this method was theoretically applicable to the light of the Stars is indeed obvious, but the practical difficulties are very great. Sirius, the brightest of all, is, in round numbers, a hundred millions of millions of miles from us; and, though as bright as fifty of our suns, his light when it reaches us, after a journey of sixteen years, is at most one two-thousand-millionth part as bright. Nevertheless, as long ago as 1815 Fraunhofer recognised the fixed lines in the light of four of the Stars; in 1863 Miller and Huggins in our own country, and Rutherford in America, succeeded in determining the dark lines in the spectrum of some of the brighter Stars, thus showing that these beautiful and mysterious lights contain many of the material substances with which we are familiar. In Aldebaran, for instance, we may infer the presence of hydrogen, sodium, magnesium, iron, calcium, tellurium, antimony, bismuth, and mercury. As might have been expected, the composition of the Stars is not uniform, and it would appear that they may be arranged in a few well-marked classes, indicating differences of temperature, or perhaps of age. Thus we can make the Stars teach us their own composition with light, which started from its source years ago, in many cases long before we were born. Spectrum analysis has also thrown an unexpected light on the movements of the Stars. Ordinary observation, of course, is powerless to inform us whether they are moving towards or away from us. Spectrum analysis, however, enables us to solve the problem, and we know that some are approaching, some receding. [Illustration: Fig. 55.--Displacement of the hydrogen line in the spectrum of Rigel.]