374 THE WORLD MACHINE 



alteration in the apparent speed, and by a definite periodicity of 

 that alteration. For example, if such a binary were approaching 

 or receding from our sun at, say, a speed of a hundred kilometres 

 per second, and at the same time were whirling about a centre 

 in an orbital velocity of about the same rate, the lines of the 

 spectrum would shift according as the one star or the other 

 was coming towards us or receding from us in its own orbit, 

 with a regularly varying rate, at one time standing still, 

 then moving at a rate indicating twice the speed of the actual 

 translatory motion of the double system. 



Precisely these conditions were disclosed in the same year 

 of 1889, by Vogel of Beilin and Pickering of Harvard. Since 

 this time the number of these " spectroscopic binaries " has 

 grown rapidly. At the present time more than sixty such 

 systems are listed, and their periods fixed. It is easy to see 

 that it would be difficult, if not impossible, to discover by means 

 of the shifting of the spectral lines a system of long period. 

 It follows, therefore, that there is a very marked contrast in 

 the period of the spectroscopic binaries and those observed by 

 means of actual change of position. With possibly a single ex- 

 ception, no double system has been revealed by means of actual 

 shift of position of less period than eleven years. On the other 

 hand, no spectroscopic binary has been noted with a longer 

 period than about three years. They range from this down to 

 about a single earthly day. 



We cannot imagine the arrangement of these short-period 

 binaries save in one of two ways : either that their orbital 

 velocity is enormous, or else that they are relatively very close 

 together. But their orbital velocity is known by the same 

 means which disclosed the fact that they are binaries. The 

 greatest speeds computed for any of the short-period systems 

 is from 250 to 300 kilometres per second. This is eight or ten 

 times the speed of any planet in our system. It is still quite 

 insufficient to explain the extreme rapidity of their revolutions. 

 Moreover, the most of them are very much less than this. The 

 Pole-star, for example, is a double star, with a period of four 

 days, and the computed orbital velocity of its components is 

 only three kilometres per second. This is very slow. 



We are forced to conclude, therefore, that in many, if not 

 in most, of these double systems, the two suns are very much 

 closer to each other than any of the planets to our sun. For 



