DENSITY OF THE EARTH. 29 



Islands, and in New Holland, compared with New York, 

 Dunkirk, and Barcelona,, have proved the contrary. 



From what has been said, it appears to follow, that the 

 pendulum, (no unimportant instrument of geological in- 

 vestigation, being a kind of sounding-line by which the 

 deep unseen terrestrial strata may be reached), affords us 

 less -well-assured knowledge respecting the figure of our 

 planet than that which may be gained from measurements 

 of degrees and the lunar movements. Strata concentric, 

 elliptic, each taken separately homogeneous, but increasing 

 in density according to certain functions of the distance in 

 proceeding from the surface towards the centre of the earth 

 (which are the supposed theoretical conditions), may, by 

 their actual diversity of constitution, position, and order of 

 density in different parts of the earth, occasion at the sur- 

 face local deviations in the force of gravity. If the circum- 

 stances which produce these deviations are much more 

 recent than the hardening of the outer terrestrial crust, we 

 may conceive the figure of the earth's surface as not having 

 been locally modified by the internal movement of the 

 molten masses. At all events, the diversities of the results 

 of measurements by means of the pendulum are far 

 too great to be now ascribed to errors of observation. 

 When pendulum .stations are grouped or combined in 

 various ways, so as to give accordance or systematic regu- 

 larity in the general results derivable from such com- 

 binations, they are still always found to give a greater 

 amount of ellipticity than measurements of degrees have 

 hitherto done. 



If we take the result of these latter as now most generally 

 received, according to Bessel's last determination, viz. : 



