82 A SHORT HISTORY OF SCIENCE 



reasonable that all should be moved in the same motion, and yet 

 among the stars the sun only seems to do so at its rising or setting, 

 and even this one not in itself but only owing to the distance of our 

 sight, as this when turned on a very distant object from weakness 

 becomes shaky. This is perhaps also the reason why the fixed stars 

 seem to twinkle, while the planets do not twinkle. For the planets 

 are so near that the eyesight reaches them in its full power, but when 

 turned to the fixed stars it shakes on account of the distance, be- 

 cause it is aimed at too distant a goal ; now its shaking makes the 

 motion seem to belong to the star, for it makes no difference whether 

 one lets the sight or the seen object be in motion. But that the 

 stars have not a rolling motion is evident ; for whatever is rolling 

 must of necessity be turning, while of the moon only what we call its 

 face is visible.' Dreyer. 



Aristotle adopts the system of spheres of Eudoxus and 

 Calippus, but seems to suppose these spheres to be concrete, and 

 not a merely geometrical device for interpreting the phenomena 

 or determining the positions. In order however to secure what 

 he conceives to be the necessary relation between the motions of 

 the spheres, he is obliged to increase their total number from 33 

 to not less than 55. The earth is fixed at the centre of the uni- 

 verse. That the earth is a sphere is shown logically, and is also 

 evident to the senses. During eclipses of the moon, namely, the 

 boundary line, which shows the shadow of the earth, is always 

 curved. ... If we travel even a short distance south or north, 

 the stars over our heads show a great change, some being visible 

 in Egypt, but not in more northern lands, and stars are seen to 

 set in the south which never do so in the north. It seems therefore 

 not incredible that the vicinity of the pillars of Hercules is con- 

 nected with that of India, and that there is thus but one ocean. 



The bulk of the earth he considers to be " not large in compar- 

 ison with the size of the other stars." The estimated circumfer- 

 ence of 400,000 stadia about 39,000 miles is the earliest known 

 estimate of the size of the earth, and is of unknown origin, but 

 may quite likely be due to Eudoxus. While the heavens proper 

 are characterized by fixed order and circular motion, the space 



