OF THE COMPASSES. 73 



between north-west and west, it amounted 1824. 

 to eight points, while with the head to the August. 

 southward, the compasses would generally rest 

 wherever they were directed by the finger, 

 and sometimes each persisted in maintaining 

 a direction of its own. Barlow's plate now be- 

 came useless, and its want of effect was decided 

 by finding Gilbert's compass, while under its 

 immediate influence, the dullest in the ship. 

 Ellis, in his account of the expedition of the 

 Dobbs and California^ 1746, says, " I cannot 

 help taking notice in this place" (while off 

 Chesterfield inlet,) '* of an accident that hap- 

 pened to us, and which as it was the object of 

 our astonishment then, has often been the sub- 

 ject of my serious thoughts. In short, amongst 

 these islands, and in sailing through the ice, 

 the needles of our compasses lost their mag- 

 netical qualities, one seeming to act from this 

 direction, and another under that, and yet they 

 were not for any considerable time constant to 

 any. We laboured to remedy this evil by 

 touching them with an artificial magnet, but 

 to very little purpose, for if they recovered 

 their powers by this means they presently lost 

 them again." P. 220. London edit. 1748. 



With a light wind, but heavy sea, from the 

 south-west, we made a N.w.b.N. course, over 



