VARIATION OF THE COMPASS. 301 



ham in 1748, make the annual hicrease between this year and 

 1723 only 8M nearly what its rate had been found before 

 this great difference occurred ; and from the variation of Mr. 

 Graham in 1748, and the variation observed by Dr. Heber- 

 den -in 1773, the annual increase is 8 ',4; the variation in 

 1773, compared with the variation observed by myself in 

 1787, give for the annual rate of increase 9'j3; but between 

 1787 and 1795, the annual increase was only 4'j7 ; between 

 J 795 and 1102, 1 ',2 ; and between 1802 and ] 809, only 

 0'.7. 



The mean rate of annual increase for the above mentioned 

 period of 207 years, viz. from 1580 to 1787, is 10'. 



As there appears something curious in the rate at v/hich the Changes in the 

 variation has been moving, from observations made at Lon- variation for 

 don, for a period of more than 200 years, the annual increase " y^^is. 

 of which daring that time continued nearly the samej but in 

 a subsequent period of 18 years only, the decrease of that 

 annual increase became so rapid, that the annual increase in 

 the latter part of it does not amount to quite one minute, I 

 shall subjoin the following Table, by way of elucidating what 

 i? here mentioned. 



