VAklATION Ot THE COMPASs, '^7 



Dipping Neadle. 



The dipping-needle with which the observations in this The dipphig- 

 comixmnicaLion were made, being the same instrument used in "eedlc. 

 former observations of the dip, and it having also been de- 

 scribed by Mr. Cavendish in the paper before alluded to, it 

 will not be necessary to say any thing of its construction here. 

 Its situation in the house was in the easletn window in, the 

 meeting-room, next the door. 



As the observations made with the dipping-needle were not 

 affected by any other source of error than that of the iron-work 

 of the room, in order to ascertain the quantity of error, the 

 instrument was taken out of doors at two different times, 

 after an interval often years, diflbrently situated each time, and 

 the observations made at both these times out of doors, com- 

 pared with the observations made in the room, giving for the 

 error 20-' more than the dip was found to be in the room, and 

 both agreeing to one minute; that quantity has been added to 

 all the observations made with the dipping-needle in the room 

 for its error, as affected by the iron-work of the room. 



Although a valuable paper on the diurnal variation of the Canton's ob- 

 , . , • I 11 1 -Kir T J r> servations on- 



honzontal magnetic 'needle, by the late Mr. John Canton, duirnal varia- 



F.R.S. was published in the first part of the List volume of'^on. Those of 

 the PhiL Trans, for the year 1759, containing a great number ^g^g carried 

 of observations made at different and irregular times of the day through six- 

 throughout the year, yet, it appeared to me, that if the varia- ^^'^'^ months, 

 tion were to be observed at short but stated intervals of the day 

 ior one year, the results would perhaps not only prove more 

 satisfactory in determining the times of the needle becoming 

 stationary, but would show its progressive and regressive mo- 

 tions better than If observed at irregular intervals. To effect 

 which, I imposed this laborious task upon myself for the space 

 of sixteen months. 



The observations contained in Table L in sixteen pages, viz. Tabulated re- 

 from September 1786, to December 1787, both inclusive, are *"^^^* 

 the results made at many but stated times of the day, and so 

 disposed, that the progress, or regress, of the varialion, may 

 be readily seen by mere inspection.*' 



Table 



* For these, on accoxmt of their length, reference naust be had to the 

 Transactions. 



