No. II.
On the errors of the compass arising from attractions within the ship, and others from the magnetism of land; with precautions for obviating their effects in marine surveying.
Several instances have been mentioned in the course of this voyage, where the compass showed a different variation on being removed from one part of the ship to another; thus observations on the binnacle gave 29½° off the Start, where the true variation was about 25½° west, whilst others taken upon the booms before the main mast, sixty-eight miles lower down Channel, gave only 24°; and in the experiments made with five compasses, Vol. I. p. 18, the mean variation at the binnacle was 4° 37′ greater than on the booms. Finding that the situation of the compass was an object of importance, I determined very early in the voyage to place it always upon the binnacle; both when taking bearings for the survey, and when observing azimuths or amplitudes; nor in any observations taken by myself, was it ever displaced except by way of experiment; but the officers occasionally observed from different parts of the ship, when the sun could not be seen from the binnacle, until they were convinced that such observations were of no utility, either to the survey, or for ascertaining the true variation.
It soon became evident, however, that keeping the compass to one spot was not sufficient alone to insure accuracy; a change in the direction of the ship's head was also found to make a difference in the needle, and it was necessary to ascertain the nature and proportional quantity of this difference before a remedy could be applied. This inquiry was attended with many difficulties, and no satisfactory conclusion could be drawn until a great variety of observations were collected; it then appeared, that when the ship's head was on the east side of the meridian the differences were mostly one way, and when on the west side they were the contrary, whence I judged that the iron in the ship had an attraction on the needle, and drew