214
of a material corresponding to the mean density of the Earth must be very nearly one-lialf of that of lead ; or, that theweight of the Earth is very nearly one-half of that of a massof lead of the same size as itself. It is surely an exceedinglyinteresting and important fact that so small and delicate alittle instrument as a torsion-balance, by that very delicacy,enables us to solve the problem of weighing so huge a massas the Earth .
The final result of the best determination yet made by theabove method is, that the Earth ’s mean density is almostexactly one-half of that of lead, or about 51 times that ofwater.
The original investigation by Cavendish, made in a.d. 1797and 1798, is recorded in the volume of “Philosophical Transac-tions ” for the year 1798. It gave a value less than the aboveby about ^th part, which was again somewhat decreasedwhen it was repeated, with certain modifications, in 1836, byProfessor Reich, of Freiburg . The value which we have statedis, however, more trustworthy, and was obtained by Mr. Baily,by an exceedingly careful series of experiments which heperformed between a.d. 1838 and 1842. Diagrams of theapparatus used by Cavendish may be found in Lardner’s“ Handbook of Astronomy,” and in other well-known books.A full description of Mr. Baily’s improved apparatus (withdiagrams), and of the elaborate precautions used by him toavoid the disturbing effects of currents of air, or of the heat ofthe observer’s body, is given in the fourteenth volume of the“ Memoirs of the Royal Astronomical Society,” the whole ofwhich is occupied with an account of his proceedings. Historsion-balance was very carefully enclosed in a glass case,while the large and small balls were separated by partitions,impervious to heat, the necessary movements being effected byropes worked from a distance, and the observations made bytelescopes through holes in a wall.
The especial value of the Cavendish experiment consists inthe simple character of the observations involved, and in itsbeing easy to make a torsion-balance which will oscillate soslowly (as, for instance, in 200, or more, seconds), that we are