162 COSMOS. 



of neighbouring strata. Of these three methods,* the last is 

 the most certain, since it is independent of the difficult deter- 

 mination of the density of the mineral masses of which the 

 spherical segment of the mountain consists, near which the 

 observations are made. According to the most recent experi- 

 ments of Reich, the result obtained is 5 '44 ; that is to say, the 

 mean density of the whole Earth is 5 '44 times greater than that 

 of pure water. As according to the nature of the mineralogical 

 strata constituting the dry continental part of the Earth's sur- 

 face, the mean density of this portion scarcely amounts to 2 '7, 

 and the density of the dry and liquid surface conjointly to 



* The three methods of observation give the following results : (1) by 

 the deflection of the plumb-line in the proximity of the Shehallien 

 Mountain, (Gaelic, Thichallin,) in Perthshire, 4*713, as determined by 

 Maskelyne, Button, and Playfair, (1774-1776 and 1810,) according to a 

 method that had been proposed by Newton ; (2) by pendulum vibrations 

 on mountains, 4 '8 37, (Caiiini's observations on Mount Cenis compared 

 with Biot's observations at Bordeaux, Effemer. Astron. di Milano, 1824, 

 p. 184) ; (3) by the torsion-balance used by Cavendish, with an appa- 

 ratus originally devised by Mitchell, 5'48, (according to Button's revision, 

 of the calculation, 5'32, and according to that of Eduard Schmidt, 5'52; 

 Lehrbuch der math. Geographic, bd. i. s. 487) ; by the torsion-balance, 

 according to Reich, 5'44. In the calculation of these experiments of 

 Professor Reich, which have been made with masterly accuracy, the 

 original mean result was 5'43, (with a probable error of only 0'0233,) a 

 result which being increased by the quantity by which the Earth's centri- 

 fugal force diminishes the force of gravity for the latitude of Freiberg, 

 (50 55') becomes changed to 5'44. The employment of masses of cast- 

 iron instead of lead, has not presented any sensible difference, or none 

 exceeding the limits of errors of observation; hence disclosing no 

 traces of magnetic influences. (Reich, Versuche iiber die mittlere Dich- 

 tigheit der Erde, 1838, s. 60, 62, and 66.) By the assumption of too 

 slight a degree of ellipticity of the Earth, and by the uncertainty of the 

 estimations regarding the density of rocks on its surface, the mean 

 density of the Earth, as deduced from experiments on and near moun- 

 tains, was found about one-sixth smaller than it really is namely, 4*761, 

 (Laplace, Mecan. celeste, t. v. p. 46,) or 4'785. (Eduard Schmidt, Lehrb. 

 ' der math. Geogr., bd. i. 387 und 418.) On Halley's hypothesis of the 

 Earth being a hollow sphere, (noticed in page 163 ; ) which was the germ 

 o Franklin's ideas concerning earthquakes, see Philos. Trans, for the 

 year 1693, vol. xvii. p. 563, (On the Structure of the Internal Parts of 

 the Earth, and the concave habited Arch of the Shell.) Halley regarded 

 it as more worthy of the Creator, " that the Earth, like a house of several 

 stories, should be inhabited both without and within. For light in the 

 hollow sphere (p. 576) provision might in some manner be contrived." 



