62 ROUGH WAYS MADE SMOOTH. 



the number of yards really intervening between that globe 

 and their home. If we rightly picture these conditions, 

 which fairly represent those under which the astronomer has 

 to determine the distance of the sun from the earth, we shall 

 perceive that the wonder rather is that any idea of the 

 sun's distance should be obtained at all, than that the 

 estimates obtained should differ from each other, and that 

 the best of them should err in measurable degree from the 

 true distance. 



Anything like a full explanation of the way in which 

 transits of Venus across the sun's face are utilised in the 

 solution of the problem of determining the sun's distance 

 would be out of place in these pages. But perhaps the 

 following illustration may serve sufficiently, yet simply, to 

 indicate the qualities of the two leading methods of using a 

 transit. Imagine a bird flying in a circle round a distant 

 globe in such a way that, as seen from a certain window (a 

 circular window suppose), the bird will seem to cross the 

 face of the globe once in each circuit. Suppose that though 

 the distance of the globe is not known, the window is known 

 to be exactly half as far again from the globe as the bird's 

 path is, and that the window is exactly a yard in diameter. 

 Now in the first place, suppose two observers watch the bird, 

 one (A) from the extreme right side, and the other (B) from 

 the extreme left side of the window, the bird flying across 

 from right to left. A sees the bird begin to cross the face 

 of the globe before B does, say they find that A sees this 

 exactly one second before B does. But A's eye and B's 

 being 3 feet apart, and the bird two-thirds as far from the 

 'globe as the window is, the line traversed by the bird in this 

 interval is of course only 2 feet in length. The bird then 

 flies 2 feet in a second (this is rather slow for a bird, but the 

 principle of the explanation is not affected on that account). 

 Say it is further observed that he completes a circuit in 

 exactly ten minutes or six hundred seconds. Thus the entire 

 length of a circuit is 1,200 feet, whence by the well-known 

 relation between the circumference and the diameter of a 



