22 OLDHAM: THE STRUCTURE OF THE HIMALAYAS, ETC. 



is obvious that this constant correction does not affect the differ- 

 ences between the deflections, but it is convenient as bringing the 

 deflections into closer approximation to their true values. 



We may now pass on to the consideration of the variations in 

 the attraction in a vertical direction. These are measured by 

 comparing the period of a free-swinging pendulum at different 

 stations ; in practice many precautions have to be taken and cor- 

 rections applied for temperature, pressure of the atmosphere, 

 flexure of the support, etc., with which we are not here concerned, 

 and in the result it is now possible to determine the vertical force 

 of gravity at any particular station with a very high degree of 

 accuracy. This result has been expressed in several different forms ; 

 at one time it was commonly expressed by the number of swings 

 in twenty-four hours of a pendulum which would beat exact seconds 

 at sea level on the equator, or it might be expressed by the accele- 

 ration which would be produced in a free falling body ; more recently, 

 however, it has become customary to express it in dynes per 

 gramme of mass, the dyne being the unit of force which, acting 

 on a mass of one gramme for one second, would produce a velocity 

 of one centimetre per second. Numerically, the value in dynes 

 is identical with the acceleration, expressed in centimetres and 

 seconds, but it is sometimes more convenient to express the result 

 as a force than as an acceleration. 



Having obtained a local measure of the force of gravity, it is 

 compared with the theoretical value of gravity at the station, and 

 the difference expressed as an " anomaly " which is positive if the 

 former is in excess, and negative if it is in defect, of the theoreti- 

 cal value ; but before this can be done it is necessary to reduce 

 the observed value to what it should be at sea level immediately 

 under the station, and to reduce the accepted equatorial value of 

 gravitation to the latitude of the station. 



To take the latter question first, the mean value of the force 

 of gravitation at the equator is not far from 978*03 dynes with 

 an error of not more than '01 ; the formula for the reduction from 

 this to the latitude of the station depends on the form of the earth, 

 which is not yet known with exactitude, but any error introduced 

 by this cause would not vary largely within the limits of the groups 

 of observations to be considered. The position is very similar 

 to that of the deflections of the plumb-line, in neither case can 



[ 170 ] 



