44 



lines parallel to each other, so that the astronomer may 

 be able to tell more exactly than he could by noting the 

 transit over one line, exactly when the star crosses his 

 approximate meridian. I say approximate meridian, be- 

 cause this middle line only marks the meridian within 

 certain limits which are set by the principle which under- 

 lies all practical astronomy, that after human art is ex- 

 hausted in rendering an instrument as perfect as possible, 

 there are still measurable deviations from perfection itself. 

 In the case of our transit we are unable to have it perfect 

 in a single point ; we cannot put the middle line of this 

 group exactly in the centre of the true telescopic field of 

 view. We cannot make this axis about which the tele- 

 scope turns, perfectly horizontal, we cannot bring the 

 telescope so that it shall move exactly in the meridian, 

 and even if we could accomplish either of these three 

 things, we should hardly have time to announce it, before 

 that unseen motion which is constantly going on among 

 the molecules of every body in the visible universe, would 

 have disturbed it to an extent appreciable by the applica- 

 tion of some refined method of investigation. 



In order to measure our day, therefore, we must assure 

 ourselves of the amount of change our instrument has 

 undergone in that interval. This is accomplished by the 

 observation of a number of stars distributed over the 

 northern and southern heavens. We can compute from 

 the discrepancies among such a series of stars the correc- 

 tion we should apply to each one of them, to get the same 

 result we should have obtained in our measure of the 

 day's length, if we had used a perfect instrument moving 

 precisely in the meridian. Let us briefly consider the 

 principles we should adopt to free our observation of a 

 star of the errors springing from the three sources I have 

 mentioned. First, we can ' determine how much that 



