VARIABLE AND LOST STARS 131 



of the common people, where a practised observer was 

 sure there was no star visible an hour or two before, 

 was he to conclude that it had flared up as if it were 

 on fire, and that it would go out as the fire died down ? 

 Or, if he saw a star brightening, paling, going out, and 

 brightening again every three or four days, or weeks 

 or months, or every three or four years, was he to 

 infer that dark bodies of vast size were thrusting 

 themselves between that distant sun and our eyes, 

 eclipsing it, in fact; or that immense reaches of un- 

 lighted space, or dark regions on its surface, were 

 turned for a time towards us, as it revolved on its 

 axis? Dark spots on a sunny star's surface and a 

 rotation more or less rapid were the causes accepted 

 by Herschel from previous astronomers for this change 

 of brightness in what are called changing or variable 

 stars. He examined seven that were then known. 

 Their periods were 3, 5, 6, 7, 331, 394, and 497 days. 1 

 He felt, however, that his views were discredited by 

 the sudden bound from 7 days to 331. Unless a star 

 were found bridging the gulf between these two, he 

 would not have had confidence to give his theory to 

 the world. But the star a Herculis seemed to him 

 to bridge the gap, and satisfy the theory. Its period 

 was found to be about 60 days. These and other 

 changes on the face of the heavens, known for many 

 years and registered in books, formed Herschel's pre- 

 lude to the work he had set his heart on, The Con- 

 struction of the Heavens. That they are a building, 

 a wonderful temple consecrated to Almighty Power 

 and Wisdom, he never doubted. To discover the plan 



1 Phil. Trans. (1796), pp. 455-56. Professor Holden gives the 

 numbers as 3, 5, 6, 7, 334, 404, and 494 (p. 133). 



