SECT, x.] AXIS OF ROTATION INVARIABLE. 87 



permanent position on the surface, if the earth be not dis- 

 .turbed in its rotation by a foreign cause, as the collision of 

 a comet, which might have happened in the immensity of 

 time. But, had that been the case, its effects would still have 

 been perceptible in the variations of the geographical lati- 

 tudes. If we suppose that such an event had taken place, 

 and that the disturbance had been very great, equilibrium 

 could then only have been restored with regard to a new 

 axis of rotation by the rushing of the seas to the new equa- 

 tor, which they must have continued to do till the surface 

 was everywhere perpendicular to the direction of gravity. 

 But it is probable that such an accumulation of the waters 

 would not be sufficient to restore equilibrium if the derange- 

 ment had been great, for the mean density of the sea is only 

 about a fifth part of the mean density of the earth, and the 

 mean depth of the Pacific Ocean is supposed not to be more 

 than four or five miles, whereas the equatorial diameter of 

 the earth exceeds the polar diameter by about 26^ miles. 

 Consequently the influence of the sea on the direction of 

 gravity is very small. And, as it thus appears that a great 

 change in the position of the axis is incompatible with the 

 law of equilibrium, the geological phenomena in question 

 must be ascribed to an internal cause. Indeed it is now 

 demonstrated that the strata containing marine dilu via, which 

 are in lofty situations, must have been formed at the bottom 

 of the ocean, and afterwards upheaved by the action of sub- 

 terraneous fires. Besides, it is clear, from the mensuration of 

 the arcs of the meridian and the length of the seconds' pen- 

 dulum, as well as from the lunar theory, that the internal 

 strata and also the external outline of the globe are elliptical, 

 their centres being coincident and their axes identical with 

 that of the surface a state of things which, according to 

 the distinguished author lately quoted, is incompatible with 

 a subsequent accommodation of the surface to a new and 

 different stage of rotation from that which determined the 

 original distribution of the component matter. Thus, amidst 



