J 



8S . Bowditch on'the Oblateness ofihe Earth, 



was the case ^vith the first observations made in France, by 

 Picard, Cassiai, &c. for the purpose of determining the form of 

 the earth. The errors of the observations aflfected the result so 

 much that it was supposed by many Mathematicians that its fig- 

 ure was prolate, or lengthened in the direction of the polar axis. 

 Even the late very accurate measures made in France by Messrs. 

 Delambre and Mechain, and in England by General Roy, make 

 the oblateness of that portion of the earth nearly double what its 

 general value is found to be by more distant observations. The 



■ 



i-eater part of this difference no doubt arises from a real irregu- 



larity in the figures of the meridians of the earth ; but the method 

 of computation itself labours under a similar defect to that which 

 exists in finding the distance of any terrestrial object by means of 

 its azimuths observed from the extremities of a base line, whose 

 length is very small in comparison with the observed ais,tancc of 

 the object. The second method of determining the earth's figure 

 by means of distant observations is much more accurate than the 

 preceding, but the various errors of the observations, and the ir- 



These irres: 



regularities of the surface of the earth, have a very perceptible ef- 

 fect on the oblateness computed by this method 

 ities ought not to be so sensible in the results of the third method 

 by means of the observed lengths of pendulums, as La Place has 

 proved in Book II, § 33, of his ^^ Mecanique CHeste,^^ and with 

 respect to ih^ fourth method, it is evident without any calculation 

 that it must be almost wholly independent of this same error. 

 For, on account of the great distance of the moon from the earth 



tf 



ffect of all the little irregularities of the earth's form, will be 

 hardly perceptible in the attraction on the moon, and the result of 

 the general figure only will prevail. Therefore by using the last 



i^ethod and taking a great number of observations, it would seem 



