.s/A' WILLIAM SIEMENS, I-'.K.S. 



363 



iiiu-line. It may here be remarked, however, that the indications 

 of variation of gravitation with variation in the depth of water, 

 which have been obtained by the use of the instrument, show in 

 excess of what the above calculation gives with the mean density 

 of the rock composing the crust of the earth as a factor, and agree 

 more nearly with what would result if the upper strata of the earth 

 WIT- of a density equal to the mean density of the whole earth. 

 Actual observations, as given in the Table further on, confirm, in 

 a remarkable degree, the arithmetical ratio of decrease of gravita- 

 tion by depth which results from the foregoing calculation. 



FIRST ATTEMPT TO CONSTRUCT A BATHOMETER. Several years 

 ago I constructed an instrument in which the gravitation of the 

 earth was represented by a column of mercury in a glass tube 

 closed at its upper end, and resting upon a cushion of air enclosed 

 in a large bulb, which air, when kept at a perfectly uniform 

 temperature, represented uniform elastic force unaffected by gravity 

 or atmospheric density. The principal difficulty that presented 

 itself in designing a workable instrument on this principle, con- 

 sisted in obtaining a scale sufficiently large to show stlch extremely 

 slight variations in the total gravitation of the earth as would 

 result from ordinary variation in the depth of water. From the 

 calculation given under the previous head, assuming the mercury 

 column to have a height of 760 millims., each fathom of depth of 

 water would represent a variation of potential force in that column 

 equal to a height of "0002059 millini., a quantity which it would 

 IK impossible to show on any scale. A scale would in reality not 

 even realize this quantity of decrease in the upper surface of the 

 column, because a portion of the adjustment of height would take 

 place in the air-bulb below, partly from the rise of mercury into 

 the bulb, and partly through increase of pressure of the imprisoned 

 air due to its compression. I succeeded, however, by means of an 

 arrangement of the instrument with three liquids of different 

 densities, in increasing the effect of a change of gravitation upon 

 the mercury column three hundredfold, whereby a change of 10 

 fathoms depth would be represented by a movement of *6177 

 millim. of the boundary between the two liquids in the vertical 

 tube, a quantity sufficiently large to be appreciated in the divisions 

 of a scale. This instrument is shown in Plate 30, Fig. 2. 



TESTS OF INSTRUMENT. This instrument was tested by me in 



