CONSTITUTION OF THE GLOBE. 23 



instead of 5, as Cavendish had made it 5,48, and 

 allows 2j for the superficial. It has been said that 

 this latter computation is too low, as there are no 

 rocks which are so low in specific gracity as 2i ; and 

 it has been proposed to make it 2f, as Playfair has 

 done; so as to allow 5,48, as Mr. Cavendish had done 

 before, for the mean density, or, taking a medium, 

 5,4. But it must be remembered that the density of 

 the ocean, which is little more than 1, is to be taken 

 largely into the estimate for the superficial parts; so 

 that La Place's computation is perhaps, after all, the 

 nearest to the truth. 



It now follows, that the matter which forms the 

 earth increases in density from the circumference to- 

 wards the centre. The law deduced from the experi- 

 ments on the specific gravity of the globe, coincides 

 with the results which have already been shown to 

 follow from comparing the ellipsoid of the earth on 

 the principle of its homogeneity, with observations 

 on the vibrations of the pendulum. That ellipsoid is 

 less unequal than it should be were it homogeneous 

 throughout. 



The irregularity of the terrestrial spheroid has been 

 already argued from the measurement of degrees on 

 different parts of its surface: it will not be useless to 

 show how this conclusion is thought to be confirmed 

 by observations on the variations in the force of the 

 pendulum in different places. 



If the earth be supposed an ellipsoid, the force of 

 gravity at its different points may be determined by 

 calculation; and hence the pendulum may be applied 

 to verify this regularity, just as it was shown capable 

 of ascertaining whether the matter of the earth was 

 homogeneous or not. The figure of the earth may 

 consequently be determined, cither by a direct measure- 



