ON THE ELLIPTICITY OF THE EARTH. 479 



distance. The sum of the discrepancies of the two 

 observers from each other at the three stations is 

 l s -94; the sum of the discrepancies between the 06- 

 served and computed vibrations is, as above stated, 

 13 s -38. Having been myself the person who specially 

 proposed this repetition, I may now permit myself 

 to refer to this statement, as an additional proof 

 that the discrepancies between the observed and com- 

 puted vibrations are due, in a far greater degree, to 

 local peculiarities, than to what may be more strictly 

 called errors of observation. The " probable error " at 

 a single station, calculated from the differences between 

 the observed and computed vibrations at the 22 stations 

 in p. 469, includes of course both station and observa- 

 tion errors; it is l s *7. The same computed from the 

 differences between the two observers at the three 

 stations above cited (regarding the arithmetical mean 

 between the observers as in each case the true result, 

 and half the difference between them as the error of 

 each), is about O s> 3. The latter is the probable error of 

 observation ; the former the probable error of observa- 

 tion and station combined ; the error of observation is 

 therefore not a quarter as large as the station error. 



The differences of opinion which once prevailed re- 

 garding the causes of such discrepancies have now passed 

 away ; and they are now universally admitted, I believe, 

 to be the effects of the unequal density of the superficial 

 strata of the earth. They were first recognised as such by 

 Captain Kater, in the Phil. Trans, for 1 8 1 9, pp. 424426, 

 who drew from his own pendulum experiments at the 

 principal stations of the British Trigonometrical Survey 

 the conclusion that, " in passing through a country com- 



