140 TIME 



so-called circumpolar star, or one that remains all day and 

 night above the horizon, it will appear, like the end of 

 his astronomical clock-hand, to describe a circle; the star 

 ought to appear to the observer to describe equal parts of 

 its circle in equal times by his clock, or while the end of the 

 clock-hand describes equal parts of its circle. In this 

 manner the hours on the Greenwich astronomical clock, 

 and ultimately on all ordinary watches and clocks regulated 

 by it, will correspond to the earth turning through equal 

 angles on its axis. We thus throw back our measurement 

 of time on the earth which is taken as a time-keeper. We 

 admit that equal rotations correspond to equal intervals of 

 conscience. But all clocks being set by the earth, how shall 

 we be certain that the earth itself is a regular time-keeper? 

 If the earth were gradually to turn more slowly upon its 

 axis, how should we know it was losing time, and how 

 measure the amount? It might be replied that we should 

 find that the year had fewer days in it. But then, how could 

 we settle that it was the day that was growing longer and 

 not the year that was growing shorter. Again, it may be 

 objected that we know a great number of astronomical 

 periods relating to the motion of the planets expressed in 

 terms of days, and that we should be able to tell by com- 

 parison with these periods. To this we must answer that 

 the relation of these periods expressed in days and in terms 

 of each other, appears now indeed invariable. But what 

 if all these relations are found to have slightly changed a 

 thousand years hence? Which body shall we say has been 

 moving uniformly? Which bodies have been gaining or 

 losing? Or, what if the ratios of their periods remaining 

 the same, they were all to have lost or gained? How shall 

 we, with such a possibility in view, assert that the hour 

 to-day is the "same interval" as it was a thousand or 

 perhaps a miUion years back? Now, certain investigations 

 with regard to the frictional action of the tides make it 

 highly probable that the earth is not a perfect time-keeper 

 nor are we able to postulate that regularity of motion by 



