Page:Proceedings of the Royal Society of London Vol 4.djvu/314

Rh supply the requisite data for deducing the numerical elements corre- sponding to the present epoch of the general theory of terrestrial mag netism. It consists of two sections ; the first comprises the observations of Captain Belcher, R.N., and the officers of H. M. S. Sulphur, at twenty-nine stations on the west coast of America, and the adjacent islands, between the latitudes of 60° 21' N. and 18° 05' S. The second contains a new determination, by the same officers, of the magnetic elements at Otaheite, made in consequence of the discrepancies in the results obtained by previous observers, and of a note in M. Gauss's Allgemeine Theorie, in which Otaheite is spoken of as a highly important station for the future improvement of the cal culations of the theory. Abstracts are given of the original obser vations which are deposited in the Hydrographic Office of the Ad miralty, as well as a full detail of the processes of reduction by which their results have been computed. The values of the horizontal and total intensities are expressed in terms by which the results of ob servation are immediately comparable with the maps of MM. Gauss and Weber in the " Atlas des Erdmagnetismus."

By an investigation into the " probable error " of a single inde pendent determination of the magnetic intensity with Hansteen's apparatus, derived from the data furnished by Captain Belcher's ob servations, the author shows the extreme improbability that the differences in the results obtained at Otaheite by Messrs. Erman, FitzRoy and Belcher, should be occasioned by instrumental or ob servational error. They are also far greater than can, with any de gree of probability, be ascribed to periodical or accidental variations in the magnetic force from its mean value. The only known cause adequate to their explanation is what may with propriety be termed Station error ; that is, local disturbing influences, in an island com posed chiefly of volcanic rocks, and where the spot of observation selected by the different observers may not have been precisely the same.

By a reference to the magnetic survey of the British Islands, the occurrence of station error is shown to be frequent in countries of far less decided igneous character than Otaheite ; and that its exist ence may always be apprehended where rocks of that nature approach to, or rise through, the superficial soil. The absolute determinations of fixed observatories are as liable to station error as those of the mag netic traveller, since no continuance or repetition of the observations can lead to an elimination of the error ; it consequently presents a practical difficulty to the proposed determination of the elements of the theory from exact observation at only a few selected positions on the globe. The remedy is to be found in the combination of fixed observatories and magnetic surveys : the observations of the survey, being made in concert with, and based on those of the fixed observatory, will be furnished thereby with corrections for the se cular, periodical, and accidental variations of the elements, and will consequently determine mean values ; and a proper combination of the mean vsiues thus determined, over a space sufficiently extensive