Ob^^cn-txfiofis on the Dircclion and In(e>i,s//// nf the [Arftit 



Necdlo No. 2 — 4 7 10 Mean of S observations, jioles at // and a 



2 22 56 a and h 



2 22 43 (; and 



2 28 4 1 rt' and b' 



The above results, being derived from one i)Osili()\i of the needle only, do 

 not represent the true dip, but furnish, by means of dilTerenees, wliat we 

 require: thus, taking the means — 



Needle No. 1 shews— that if the poles, instead of being situated at c, e» 

 should take up a position such a,s e and a, (wliich in 

 practice has occasionally occurred) then, the reading 

 would be erroneous to the amount + 0'',49',0" ; and 

 that, in an extreme case, when both poles are astray, 

 then an error in the reading to nearly double of this 

 amount (or + 1*,35',23") results — or, we obtain on the 

 whole. 



The Poles hevig situated at 



Needle. 



e I 





a 





«' Si 



' b' 



^ * [i 



a ^ b' 

 a' S^h 





. error. 



error. 



error. 



error. 



error. 





o 





o 





o 



/ 



o 





o 



No. 1 



0 



0 











+0 



46 



4- 1 35 



— 2 



0 



0 



0 



0 





6 







1 44 



^ 3 



0 



0 















1 37 



— 4 



0 



0 











+ 0. 



37 



1 37 



Mean. 



0 



0 



0 



0 



+ 0 



6 



+0 



42 



4- I 38 



A greater degree of consistency would no doubt exist between these 



numbers, were the ends of the needles symmetrical, and of exactly the 



same size ; then we should expect to find ° ' 



Poles 6 and e' the error of observation =00 



a and b or a' and b' — • =00 



e and h ox e and b' > , ^ 



q' and a or e' and a' \ — - 



a and b' or a! and b — — = it ^ 38 



The greatest disagreement from these being in the case of needle 



No. 4, (poles at e and I have, since writing the above, turned to 



