158 India's Contribution to Geodesy. [May 16, 



the mountain. The geodetic operations of the survey show the same 

 thing in the Southern Peninsula, which is the part of India closest 

 to the ocean. Here the latitude stations almost invariably show 

 deflection of the plumb-line towards the ocean, instead of away from 

 it, as might naturally be expected. Here, too, the astronomical 

 amplitudes of the longitudinal arcs are all too small, with one excep- 

 tion, which can be readily explained, thus also showing deflection 

 of the plumb-line towards the ocean at the stations on the coast- 

 line. 



It is not difficult to draw a line beyond which Himalayan 

 attraction need not be feared. Excluding all stations obviously 

 under Himalayan influence, 148 latitudes and 50 longitudinal arcs 

 remain for employment with similar data obtained in other parts of 

 the world in determining the figure of the earth. But on com- 

 paring the astronomical with the geodetic determinations it is seen 

 that the former are often grossly influenced by deflections of the 

 plumb-line at points where there is nothing visible to suggest 

 disturbance. Thus, at the Colaba Observatory, Bombay, there 

 is probably 8" deflection to the north, while at the Madras 

 Observatory there is probably as great deflection to the south, neither 

 place having any apparent source of meridional attraction; here the 

 astronomical amplitude is 16" less than the geodetic, a difference 

 which is fully twenty times greater than any error that can have been 

 made in the triangulation. There are also discrepancies of corres- 

 ponding magnitude in the longitudinal arcs at places where there is 

 no apparent source of attraction on the prime vertical. 



Clearly, therefore, no single astronomical determination can be 

 regarded as sufficiently free from deflection of the plumb-line to be 

 safely employed in an investigation of the figure of the earth. But 

 all investigations hitherto made have rested on single determinations. 

 This, however, must be almost certainly due to the paucity of data ; 

 had a sufficiency of astronomical results been forthcoming to indicate 

 the presence of large local disturbances of the plumb-line and 

 emphasise the necessity to eliminate these disturbances as far as 

 possible before proceeding to employ the astronomical results, there 

 can be no doubt that some attempts at elimination would have been 

 made. 



There is only one way of eliminating these errors, and that is to 

 combine the astronomical observations together in groups, each con- 

 taining determinations at as many places in a given belt of latitude 

 as possible, and to take the mean of the group ; for the mean astro- 

 nomical latitude of a number of points may certainly be assumed to 

 be far more free from deflection than that of any single point. The 

 geodetic latitudes of the same points, as obtained from the triangu- 

 lation, would be similarly combined ; and thus the most accurate 



