45 



middle line is away from the true centre of the telescope, 

 by observing some pretty slow moving star near the pole 

 as it crosses the first two wires of the group, we will say. 

 Now we know the distance which these two wires are 

 from this middle one, and we can therefore determine the 

 time at which the star would cross the middle wire. We 

 left our star slowly moving, and after it had passed the 

 second wire, now suppose that we skilfully and rapidly 

 lift our instrument out of its bearings and replace it with 

 its horizontal axis end for end, and point the telescope 

 upon the same star which has not yet crossed these last 

 two wires. Now if you reflect a moment you will see 

 that these are the same two wires across which we have 

 first observed the star's transit, only they are now on the 

 other side of the field of view, and if we now compute at 

 what time the star crossed the middle wire from these last 

 two transits we shall find that it differs from the time 

 derived from the first two transits by twice the amount 

 the middle line is from the true centre. Second, we can 

 determine how nearly level this horizontal axis is by 

 means of this very delicate spirit-level, which you see 

 rests only on the axis itself. And third, we can determine 

 how much our instrument (due allowance being made for 

 the two preceding sources of error) still deviates from the 

 meridian, by combining star transits taken near the pole 

 and the equator. It is the habit of astronomers to con- 

 sider a series of stars observed in this manner as consti- 

 tuting one time determination. And if our clock shows 

 by its face that the sidereal day just measured has ended 

 ten minutes too soon, then we say that our clock is ten 

 minutes slow. 



I think now you will understand me when I say that 

 the first objection the astronomer has to observing the 

 sun for time, is, that it is difficult to get enough stars in 



