ge,
log, currents, half-minute-glasses; and sometimes want of care, as in so
long a run cause often a difference of many leagues in the whole account.

Most of my men that kept journals imputed it to the half-minute-glasses:
and indeed we had not a good glass in the ship beside the half-watch or
two-hour-glasses. As for our half-minute-glasses we tried them all at
several times, and we found those that we had used from Brazil as much
too short, as others we had used before were too long; which might well
make great errors in those several reckonings. A ship ought therefore to
have its glasses very exact; and besides, an extraordinary care ought to
be used in heaving the log, for fear of giving too much stray line in a
moderate gale; and also to stop quickly in a brisk gale, for when a ship
runs 8, 9 or 10 knots, half a knot or a knot is soon run out, and not
heeded: but to prevent danger, when a man thinks himself near land, the
best way is to look out betimes, and lie by in the night, for a commander
may err easily himself; beside the errors of those under him, though
never so carefully eyed.

Another thing that stumbled me here was the variation, which, at this
time, by the last amplitude I had found to be but 7 degrees 58 minutes
west, whereas the variation at the Cape (from which I found myself not 30
leagues distant) was then computed, and truly, about 11 degrees or more:
and yet a while after this, when I was got 10 leagues to the eastward of
the Cape, I found the variation but 10 degrees 40 minutes west, whereas
it should have been rather more than at the Cape. These things, I
confess, did puzzle me: neither was I fully satisfied as to the exactness
of the taking the variation at sea: for in a great sea, which we often
meet with, the compass will traverse with the motion of the ship; besides
the ship may and will deviate somewhat in steering, even by the best
helmsmen: and then when you come to take an azimuth there is often some
difference between him that looks at the compass and the man that takes
the altitude height of the sun; and a small error in each, if the error
of both should be one way, will make it wide of any great exactness. But
what was most shocking to me, I found that the variation did not always
increase or decrease in proportion to the degrees of longitude east or
west; as I had a notion they might do to a certain number of degrees of
variation east or west, at such or such particular meridians