1 12 CAUSES which affeEl the ACCURACY 



= o, and |3 = b, fo that 



o 



_ g ( * "0. + _f_( 1_ + _*^ \ . if t hi s va j ue be fubftituted for 



7 2 O O 



T 



?, and if all the terms be divided by (i m) , we {hall have 



n in n 



q (s+x~) a n I i i 



-f 



THE approximation which has been ufed here for finding 







the fluent / - - , is fufficiently exadl, becaufe no 

 ' 



terms have been rejecled but fuch as are divided by s 2 , and 

 which, of confequence, are extremely imall in refpedl of the 

 reft. 



27. WE are now to fuppofe, that x becomes equal to z, or 

 to the whole height that is to be meafured ; then alfo, 



Hb 

 = r h, x = - } an d T = r H, a s in 14. ', 



z 



n in n 



and fo by fubftitution, iL J_ _ 



21 n n 



h} (i TO) 



