A variety of observations with the compass had shown the magnetic needle
to differ from itself sometimes as much as six, and even seven degrees,
in or very near the same place, and the differences appeared to be
subject to regular laws; but it was so extraordinary in the present
advanced state of navigation, that they should not have been before
discovered and a mode of preventing or correcting them ascertained, that
my deductions, and almost the facts were distrusted; and in the first
construction of the charts I had feared to deviate much from the usual
practice. Application was now made to the Admiralty for experiments to be
tried with the compass on board different ships; and the results in five
cases being conformable to one of the three laws before deduced, which
alone was susceptible of proof in England, the whole were adopted without
reserve, and the variations and bearings taken throughout the voyage
underwent a systematic correction. From these causes the reconstruction
of the charts could not be commenced before 1813, which, when the extent
of them is considered, will explain why the publication did not take
place sooner; but it is hoped that the advantage in point of accuracy
will amply compensate the delay.

Besides correcting the lunar distances and the variations and bearings,
there are some other particulars, both in the account of the voyage and
in the Atlas, where the practice of former navigators has not been
strictly followed. Latitudes, longitudes, and bearings, so important to
the seaman and _un_interesting to the general reader, have hitherto been
interwoven in the text; they are here commonly separated from it, by
which the one will be enabled to find them more readily, and the other
perceive at a glance what may be passed. I heard it declared that a man
who published a quarto volume without an index ought to be set in the
pillory, and being unwilling to incur the full rigour of this sentence, a