The problem of determining longitude without reference to a timepiece is one which even today commands the interest of a number of mathematicians, astronomers and navigators. Having recently undertaken a crossing of the Atlantic in a small boat, I was particularly interested in Mr. Chichester's article and Mr. Sadler's comments in the January edition of this Journal.

The use of the visible horizon as a reference point for measuring the lunar distance from a celestial body is described in some of the older books on Nautical Astronomy. One method consists of observing the altitudes of the Moon and second body on opposite sides of the meridian and thence computing their respective hour angles. The sum of their hour angles is then equivalent to the difference in right ascensions of the two bodies. With this information, the lunar distance can readily be found and compared with tabulated or calculated lunar distances for given instants of G.M.T. to obtain the G.M.T. of observation. Mr. Chichester's variation of the method is certainly ingenious, and the need to compare lunar distances is thereby dispensed with. I cannot, however, see any reason why intervals as large as one hour between plots should be necessary: the overall size of the plot could well be reduced to one-sixth of the size, without loss of accuracy, by using times say 5 min. on each side of the ‘guessed at’ G.M.T.