In 1953 an Institute working party was set up to investigate the accuracy of dead reckoning in air navigation and to attempt to produce error data covering a wide range of operational conditions. Dead reckoning errors fall into two categories, those due to wind finding and forecasting, and those due to instruments. Mr. Durst's paper surveys the whole problem, laying particular emphasis on the meteorological aspects, and represents the main progress in this subject to date.

The error of a dead reckoning position in the air will depend on the accuracy of the initial position from which the departure is made, the accuracy of the course steered (compass error), the accuracy of distance flown (airspeed-indicator or engine-revolution error), and the accuracy of the wind used (forecast error, found-wind error, &c). These quantities can be expressed mathematically, as is done in Appendix I, and it will be seen that if a pair of fixes is available it is possible to derive the dead reckoning error. The standard error that will arise when a large number of D.R. positions have been calculated on different occasions using a large number of aircraft and observers can be written symbolically; the expressions are fairly simple provided that the errors made in the different classes listed above (e.g. the errors in compass and distance run) are not closely correlated.

It is not possible at this stage to do more than hint at the possible alterations in navigational procedures which the new knowledge contained in Mr. Durst's paper on the accuracy of dead reckoning in the air logically leads to. This is so for two reasons. First, existing navigational procedures differ so widely between civil and military users (and amongst civil users themselves) that detailed modifications to existing procedures require separate consideration in the light of the different current navigational drills at present used. Secondly, only two particular areas of the world, the North Atlantic and Central Africa, were considered in the paper, though these are fairly representative of extra tropical and tropical latitudes (there will, of course, be variations from place to place owing to the known variability of the general circulation of the atmosphere). Nevertheless, the data obtained by the meteorological statisticians is of such interest and importance that certain broad applications suggest themselves.

Waves, tides and ocean currents might not be expected to produce electric currents, but by moving water, as a conductor, through the Earth's magnetic field they do in fact produce measurable voltages and flows of electricity. Young, Gerrard and Jevons, who made the first detailed study of this phenomenon, thought it conceivable that measurements of electric potential gradient might be used to estimate the drift of a ship at sea, and Von Arx showed how indications from electrodes towed behind the Research Vessel Atlantis were used to keep her on a direct course from Bermuda to Woods Hole, across the Gulf Stream. In recent years the method has been used extensively in studies of ocean currents and more ships might use it if the necessary equipment were readily available.

Scandinavian interest in transarctic flying dates back to the successful transarctic flight of the airship Norge under the joint leadership of Amundsen and Ellsworth. The geographical position of Scandinavian countries, of course, makes such interest natural as they have everything to gain from commercial flying across the Arctic. In 1946 when Scandinavian Airlines System was formed, with the primary object of establishing intercontinental routes to the Scandinavian countries, four Stratocruisers were ordered from the Boeing factory in Seattle and plans were made for a non-stop delivery flight from Seattle to Stockholm following the great-circle route across sub-arctic Canada and central Greenland. The object was to gain experience over this route, which might become of strategic importance in the near future. However, the Stratocruisers were sold before S.A.S. took delivery and the plans for the flight were abandoned. The groundwork carried out in connection with navigation was not wasted however.

This paper sets out to describe some of the ways in which merchant seamen may make useful sketch surveys of anchorages, coastlines or even whole islands, so as to improve the official charts, plans and sailing directions in areas which lack complete surveys by naval surveyors. These areas in fact include hundreds of miles of coastline, and hundreds of square miles of water where the depths are dangerously uncertain. The paper also describes some unorthodox uses of the official charts and, finally, attempts to show the usefulness of these improvements. All the work described has been done in the south-west Pacific, among the islands of Melanesia, where good surveys are few, and the gaps are filled in with scraps of early surveys.

The discussion following Captain C. V. Sølver's claim to have discovered an early Norse bearing-dial incidentally raised the question of what was the ‘dyoll’ referred to in English ship inventories of the fifteenth century. There are various practical reasons for doubting the theory that it was a compass-dial, that is to say a sun-dial with an inset compass, which for a particular latitude would, when correctly orientated, indicate the time of day. However, the idea that the ‘dyoll’ indicated time rather than direction seems to fit the evidence, and G. P. B. Naish's note on fifteenth- to seventeenth-century ship inventories leaves little doubt that it was an instrument for measuring equal intervals of time, and probably a sand-glass. This conclusion, which to some extent casts the early history of the compass card once more into the melting-pot, prompts a number of questions. Some are stated here and are answered either wholly or in part. Others remain to be answered.

The note in the Forum ‘The Accuracy of Astronomical Observations at Sea’ by G. A. Harding (this Journal, Vol. VII, p. 195) and the paper ‘The Barcoo Observations’ by J. B. Parker and D. H. Sadler (this Journal, Vol. VII, p. 317) were interesting, but although some practical navigators have dabbled with the theory of errors one is apt to feel a little bit blinded by the science of the mathematicians on the subject of ‘errors’, when such things as ‘constant bias’, ‘quasi systematic error’, ‘internal standard deviations’, and a ‘ square root law’ (p. 320, Table I, Col. 7) are bandied about. One hopes one knows what they mean, but there is a lurking doubt, though one might not admit it, and they do not quite seem to click with the sort of phrases taught to navigators.

I think the Institute is doing a grand job in this matter of getting to the bottom of what actually happens at sea. I give a personal ‘shabash’ to D. H. Sadler & Co. for the new Abridged Nautical Almanac, and to S. M. Burton for the first-ever really practical manual and tables, but I am now completely ‘sold’ on H.D. 486 [Tables of Computed Altitude and Azimuth] and have very little to look forward to unless it be my dream of ‘a small hopper, into which one throws two or more altitudes, turns a handle, and tears off a neatly printed position’. I can see no loss of caste to seamen in this. All I ever did know about a haversine, for example, was that it contained a conglomeration of figures which when conjoined with others, cosines and suchlike, led by tortuous and unseamanlike stages to a position line on a chart. I never felt the slightest bump when crossing over to H.D. 486, where all those abominable spherical trigonometrical triangles have already been worked out by experts who, I understand, really like 'em. So roll out that hopper.

There is a fast growing realization that the Steering Rules in the Regulations for Preventing Collisions at Sea should be attuned to the use of radar. The principle underlying the suggested rules which follow is to stick to the idea of Rule 18—that in case of risk of collision both vessels should take avoiding action. In order to make these rules practicable, a universal system has been thought out. Thus, the rules are meant to apply to vessels in sight of one another, as well as to vessels in radar-contact, or to a vessel in radar-contact with another vessel which is unaware of her presence. Secondly, there will be one and the same rule for power-driven vessels and sailing vessels.