434 THE YARD, PENDULUM, AND METRE. 



over again, this would be as good a unit as could be 

 proposed. 



(14.) But the true objection to the choice of the pen- 

 dulum for a universal unit of measure lies, not in any 

 metaphysical and abstract considerations of this kind ; 

 but in the uncertainty which prevails, and must neces- 

 sarily always prevail as to the true length of that normal 

 or ideal pendulum which shall stand equally related to 

 the whole globe, and from which the mean length cor- 

 responding to any assigned latitude can be calculated : 

 that is to say, the length of a pendulum which would 

 swing seconds at the pole of the terrestrial spheroid 

 an uncertainty which, as I shall proceed to show, must 

 affect the result of every attempt to deduce it with the 

 precision the subject requires from experiments made on 

 the surface of our planet : however refined the methods 

 employee 1 and however numerous and diversified the 

 geographical stations at which they may be instituted. 



(15.) In practice, the mean length of the polar or 

 equatorial pendulum is concluded from an assemblage of 

 the observations of the times of oscillation of one and 

 the same invariaHe pendulum at a multitude of geo- 

 graphical stations in all accessible latitudes in both 

 hemispheres : no 'wo combinations agreeing in giving 

 the same precise length, by reason of the local deviations 

 of the intensity of gravity due to the nature of the soil, 

 and the configuration of the ground immediatety beneath 

 and around the places of observation. Now, since the 

 pendulum cannot be observed at sea, the whole sea- 

 covered surface of the globe is of necessity excluded 



