MAGNETIC SURVEY OF GREAT BRITAIN. Ill 



course of every observation. From some defect of workman- 

 ship, however, the balance of the needle was much deranged 

 in some positions of the axle ; and it was accordingly employed 

 by Major Sabine as an ordinary dipping needle, the axle being 

 permanently fixed in one position in which the needle was to- 

 lerably balanced. This was the case during the observations 

 made with it in August, September, and October, 1834 (Fifth 

 Report, p. 139) ; the axle being undisturbed during the whole 

 of the series. In 1835, when Captain Ross used this needle 

 at Westbourne Green, the axle had been repolished, and was, 

 moreover, fixed by the artist in a different position from that 

 which it had occupied during the observations of the preceding 

 year. So far, therefore, as axle error is concerned, the needle 

 must, then and thenceforward, be regarded as a different nee- 

 dle. 



In order to deduce the amount of the axle error, previously 

 to the alteration just alluded to, we may compare the result 

 obtained with this needle at Limerick, in August 1834, with 

 the mean dip of the place as given by other needles. The dif- 

 ference (4''2) is probably not greater than the probable error 

 of observation, which, owing to the imperfect polish of the 

 axle, was in this needle considerable. Under these circum- 

 stances, we are not justified in assigning to it any correction. 



The needles employed by Mr. Fox appear to give results 

 extremely consistent with one another, and with those of other 

 needles. In their case, therefore, no correction is required. 



We are now prepared to exhibit in one view the mean* 

 values of the dip, as deduced from these various needles. The 

 following table contains the results of observations arranged 

 chronologically, and corrected as has been above explained. 



* Where the needles L. 1 and L. 4 have been employed together, double 

 weight has been allowed to the results of the former in taking the mean, in 

 accordance with the conclusion of page 98. 



