In the course of a survey of the Arafura Sea by H.M.A.S. Barcoo, many astronomical observations were made under good, and controlled, conditions; through the courtesy of the Admiralty these observations have been made available for analysis. This analysis has been regarded as a pilot investigation for the working party on the accuracy of astronomical observations at sea. Its main, crude, result gives (from these observations) a standard deviation of a single observation as 1.1.

During the latter half of 1952, H.M.A.S Barcoo was engaged in surveying an area, 200 miles by 30, in the Arafura Sea some 50 miles north-west of Cape Wessel. Seven fleets of beacons were laid out, each consisting of about 14 beacons at intervals of about 5 miles. The first beacon of each fleet was used as a datum beacon and its position determined by astronomical observations by three observers, both morning and evening on two days; each observer took as many stars as possible on each occasion and the mean of three or four shots of each star. The datum beacons were connected by taut-wire runs and it was found that the taut-wire distances were not in agreement with the mean positions determined from the astronomical observations. It was primarily for this reason that the observations were carefully collated by Lieutenant J. M. Child, R.N., and forwarded to the Admiralty. The observations were in due course sent to H.M. Nautical Almanac Office, where it was decided to subject them to a detailed statistical analysis. They are not suitable for inclusion in the (deliberately) heterogeneous observations being collected for the working party of the Institute on the accuracy of astronomical observations at sea; but they provide the ideal material for a pilot investigation into methods of analysis.