navigator requires only to know what is the hour at Greenwich at the time he makes his observation. This he discovers in the following manner: He observes with the sextant the distance of the moon from the sun, or from some of the most conspicuous stars; he then, after certain preliminary calculations not necessary to detail here, examines in the Nautical Almanac, where he learns what the hour is at Greenwich, when it has these particular distances from the sun or the stars. Knowing this, and knowing the hour where he is, the difference of the longitude of a ship and the observatory at Greenwich is known to him.

Although the moon be of all the celestial objects the best adapted for this observation, it is not the only one which has been resorted to. It may be in a position so near the sun that it cannot be conveniently observed; in its absence, the navigator may resort to planets which may happen to be visible. These may be used in the same manner and according to the same principles as the moon, but they do not afford a result susceptible of the same accuracy, inasmuch as their motions being slower, he cannot be so certain of their exact positions.

The advantage which the lunar method of determining the longitude has for the purpose of the mariner is, that it is always available, when the sky is unclouded. There are, however, other methods which are applicable occasionally, both by sea and by land, which ought not to be omitted here; among these the most frequent, and consequently the most generally available, is the eclipses of Jupiter's satellites. Whenever that planet is sufficiently removed from the sun to be visible after night-fall, his moons may be seen with an ordinary telescope; indeed, they were discovered at so early a period in the progressive improvement of the telescope, that they must have been first observed with a very inferior instrument of that kind. The periodic time of the first of these satellites, or that which is nearest to Jupiter, being only about 42 hours, and its position and motion being such that it cannot pass behind Jupiter without going through his shadow, its eclipse must regularly recur every 42 hours. The times of the eclipses at Greenwich are registered in the Nautical Almanac, and if they are observed at a distant place, the time at which they occur may be compared with the time at which they would be seen at Greenwich, and the longitude of the place consequently known. In fact these eclipses may be regarded as signals which can be seen at the same time from the two places; the only difference between them and common signals being that their occurrence can be certainly and accurately predicted. It is proper however to observe, that although this method is eminently useful to the geographical traveller, it can scarcely be said to be available in navigation.

There are other celestial phenomena of occasional occurrence which may also be used for determination of longitudes. Such are solar eclipses, but more especially the occultation of stars by the dark edge of the moon. This latter phenomena is one which admits of very great precision.

In connexion, with the subject of this discourse, it may not be uninteresting or unprofitable to explain the expedient by which the British government enable all navigators leaving the Thames to take with them the precise Greenwich time, which, as we have shown, is necessary for the determination of the longitude of the ship in the absence of the opportunity or ability of practising the lunar method. For a great number of years, the establishment of an easy and certain method of accomplishing this was regarded as an object of great national importance by the English public. At length the object was accomplished by the expedient now in use, and which we are about to explain.

The Royal Observatory of England is built on the summit of an elevated ridge that overhangs the town of Greenwich, on the right bank of the Thames,

and forms a conspicuous object from the river. The towers of the observatory are at all times visible from ships sailing down the river. It was, therefore, decided that a signal should be given at the instant of one o'clock in the afternoon of each day; by observing which, navigators within view of the observatory could correct their chronometers. The signal adopted for this purpose was the sudden fall of a large black ball, placed upon a pole raised from the top of one of the towers of the observatory.

Before elevating the ball, at five minutes before one o'clock, a signal is made of the intention to do so by raising it half-mast high. Observers are then instructed to prepare their chronometers; and as the descent of the ball occupies several seconds, they should confine their attention to observing the moment when the ball leaves the top, as it is that alone which indicates the hour.

The use of this signal is not merely confined to the indication of the mean time at Greenwich for navigators going down the river. By observing the drop of the ball, repeated day after day, mariners who are in the river will be enabled to ascertain the daily rate of their chronometers. Thus, if a clock were found to show the time of 3 min. 5 sec. after 1 o'clock at the moment of dropping the ball one day, it will appear that the clock is 3 min. 5 sec. faster than the mean Greenwich solar time. On the following day, if you again observe the descent of the ball, and find that the clock shows 3 min. 7 sec. after 1 o'clock, you find that it gains 2 seconds per day. Thus you are enabled, not only to ascertain the actual error of the chronometer, but also predict the manner in which that error will be augmented or diminished for the future.

In noticing the different methods which have been proposed for determining the longitude, I ought not to omit one which has been recently resorted to with considerable advantage, and which is called the method of determining the longitude by moon-culminating stars. In the practice of this method a star is chosen which culminates or passes the meridian nearly at the same time with the moon, and which differs so little in declination with the moon, that it may be seen at the same time in the field of view of the telescope. The transit of the star and that of the moon's limb, is observed at both stations, and the difference of the time at the two stations noted. This difference being dependant on the moon's change of position on the firmament, in passing from the meridian of one station to the meridian of the other, will enable the observers to determine the time which the centre of the moon takes to pass from the one meridian to the other, which will give the difference of the longitudes.

The spirit of this method is derived from the great accuracy of the knowledge we have acquired of the moon's motions, and the precision with which we can observe its transits over the meridians. In the practice of this method, it is indispensable that the moon and star should differ so little in declination that the position of the telescope will not require to be changed to observe their respective transits. Although the method has been called that of moonculminating stars, the only reason why the moon and star should be required to pass the meridian nearly together is, that the same errors may, as far as possible, affect both transits, and if so no effect would be produced on the ultimate result.

THEORY OF COLORS.

Refraction of a Ray of Light.-At plane Surfaces.-By a Prism.-The Prismatic Spectrum.-The Decomposition of Light.-Newton's Discoveries.-Colors of the Spectrum.-Brewster's Discovery of three Colors-How three Colors can produce the Spectrum.-Colors of natural Bodies.-How they are produced.

THEORY OF COLORS.

WHEN a ray of light meets the surface of a transparent medium, such as water or glass, in a line perpendicular to that surface, it will pass through without changing its course; but, if it meet the surface at any oblique angle, it will be bent into another direction, which will depend on the direction of the incident ray, and the relative densities of the media, between which the ray passes. Generally, when it passes from a less dense into a more dense medium, it is bent toward the perpendicular drawn to the surface of the medium at the point of incidence of the ray. In this deflection it does not leave the plane passing through the incident ray, and that perpendicular.

To render this more clear, let c, fig. 1, be any visible object placed on the bottom of a vessel of water. Let c n be a ray of light passing from that object to the surface of the water, that ray after leaving the surface of the water and passing into the air will not continue in the direction en, but will take