The shore of the bight is sandy and low, and trends from Point Avoid about five miles to the east; after which it takes a more southern direction and becomes higher, and the projecting parts of the waving coast line are cliffy.  Behind the shore the land rises to a moderate height, is destitute of vegetation, and of a yellow colour, but whether from the surface being of bare rock, or of sand, could not be distinguished.

In stretching off again, with the wind at east-south-east, we passed near to a small circular reef, lying nine miles from Point Avoid and six from the nearest shore.  Azimuths taken at this time with three compasses on the binnacle, and the ship’s head at south (magnetic), gave the mean variation 1 deg. 12’ east; but with the surveying compass alone it was 1 deg. 39’ east, which is what I allowed in the survey.  On the preceding day the two guns upon the quarter-deck, nearest to the binnacle, had been struck down into the after-hold, from a persuasion that the differences so often found in the variations and bearings when on different tacks must arise from some iron placed too near the compasses.  Strict search had been repeatedly made for sail needles, marline-spikes, or other implements of iron which might have been left in or about the binnacle, but I could fix on nothing unless it were the guns; for it is to be observed that, notwithstanding the constancy of the differences, the idea of any regularly acting cause to derange the needle had not yet fixed itself in my mind.  The perfection to which naval science had arrived did not allow me to suppose, that if a constant and unavoidable attraction existed in ships, it would not have been found out, and its laws ascertained; yet no longer than three days before, differences had been observed sufficient, one would think, to have convinced any man that they were produced by some regular cause.  Off Point Drummond, about fifteen leagues to the north of where the variation 1 deg. 39’ east was observed with the ship’s head at south, both azimuths and an amplitude had been taken with the same compass.  The first gave 1 deg. 33’ west, the head being south-east-by-east; and after we had tacked, and the head was south-west-by-west, the amplitude gave 3 deg. 56’ east!  I did not yet see that as the ship’s head was as much on the east side of the magnetic meridian in one case as it was to the west in the other, so was the variation as much too far west then as it was too far east afterward.  Differences like this, of 51/2 deg., which had frequently occurred, seemed to make accuracy in my survey unattainable from not knowing what variation to allow on the several bearings.  The guns were removed in the hope to do away the differences, but they still continued to exist, nearly in the same proportion as before; and almost in despair, I at length set about a close examination of all the circumstances connected with them, in order to ascertain the cause, and if possible to apply a remedy; but it was long, and not without an accumulation of facts, before I could arrive at the conclusions deduced and explained in the Appendix No.  II to the second volume.