It is to be understood of course that the variation of any of the chronometers will be but slight if they are good instruments. Moreover the tendency to vary in one direction or the other of each individual instrument will be known from previous tests. Such tests are constantly made at the Royal Observatory in England and elsewhere, and the best chronometers bear certificates as to their accuracy and as to their rate of variation. It may be added that a chronometer or other timepiece is technically said to be a perfect instrument, not when it has no variation at all--since this has proved an unattainable ideal--but when its variation is slight, is always in one direction, and is perfectly or almost perfectly uniform. FINDING THE TIME WITHOUT A CHRONOMETER In the reference made above to the testing of Harrison's watch, it was stated that that instrument varied by only a certain number of seconds in the course of the westerly voyage across the Atlantic, and that its variation was somewhat greater on the return voyage. This implies, clearly, that some method was available to test the watch in the West Indies, without waiting for the return to England. At first thought this may cause no surprise, since the local time can of course be known anywhere through meridian observations; but on reflection it may seem less and less obvious as to just what test was available through which the exact difference in time between Greenwich, at which the watch was originally tested, and local time at the station in the West Indies could be determined. There are, however, several astronomical observations through which this could be accomplished, and in point of fact the comparative times and hence the precise longitudes at many points on the Western Hemisphere--and indeed of all portions of the civilized globe--were accurately known before the day of the chronometer.