278 PHYSICAL SCIENCE BK. vn 



again on the next and following days and set fire to 

 the same spot. As presumptive proof of this, we 

 see winds return during several days at their set 

 time. Rain, too, and storms in other forms recur 

 3 according to appointment. His opinion may be 

 briefly expressed by saying that he supposes comets 

 to be formed pretty much in the same way as fires 

 excited by whirlwind. There is this one difference, 

 that those whirlwinds are pressed down to earth 

 from a higher region, while these others are raised 

 from earth to the upper regions. 



VII 



1 A GREAT deal can be urged against this view. First 

 of all, if wind were responsible, a comet would never 

 make its appearance without wind. As a matter of 

 fact, it appears when the air is perfectly still. In 

 the next place, if it were due to wind, it would 

 fall with the wind ; and if it began through wind, 

 would increase with increase of wind, and would be 

 the brighter the more furious the wind was. This 

 point, too, has to be added to the foregoing : while 

 the wind impels many parts of the atmosphere, a 

 comet appears in one spot. The wind does not 

 mount up high, but comets are seen higher up than 



2 the winds are permitted to go. Epigenes after- 

 wards goes on to speak of the comets that, he says, 

 have a more definite resemblance to stars, traversing 

 an orbit and passing through the zodiacal signs. 

 He attributes their origin to the same causes as 

 produce those that he called lower comets, the only 

 difference being that the earth's exhalations in this 

 case contain many dry elements, and therefore seek 



