Fig. 8 will further illustrate the way in which the permanent magnetism, and the inductive magnetism of vertical iron acts upon the compass to produce semicircular deviation. Let it be supposed that the whole of the south polarity or attractive power of the above magnetism is concentrated in the point P on the port quarter of a ship built with her head near N.W. The ship is supposed to be swung round the compass, beginning at the N.W. point. The small circles represent the compass, the thick lines N' S the compass needle, the dotted lines the magnetic meridian or the direction of the needle when free from deviation. Beginning at N.W., and noting the position of the point P, it will be observed that there can be no semicircular deviation

with ship's head in that direction, because the attractive force of the ship's magnetism at the point P is in a line with the compass needle N S. As the ship's head swings round towards the west, the relative positions of the point P and the compass needle will alter, and P will exert a pulling force upon the north end of the needle, causing it to deviate to the right from N to N', shown in the figure at West. The easterly deviation will increase until the ship's head swings to S. W., where it attains its greatest or maximum amount. After passing S.W. it gradually decreases past South until the ship's head reaches S.E., the opposite direction to that in which her head was built, where it is again zero or nothing. The point P is now on the opposite side of the compass to what it was when her head was at N.W., but it will be observed that it is in a line with the needle, and can exert no deviating influence over it.

As the ship swings with her head towards the East, the needle will gradually be drawn to the left hand until the westerly deviation attains its maximum at N.E. After passing N.E. the westerly deviation will decrease past North until the ship's head again reaches N. W., at which point there is no deviation. A very slight inspection of the figure will show that in the semicircle from N.W. round by the West to S.E., the deviation is easterly; while in the semicircle, or half the compass, from S.E. round by the East, the deviation is westerly. The above is merely given for the sake of illustration, but it must be remembered that no two ships are alike in their influence on the compass, nor will the ship's magnetism have the same effect on two compasses placed on different parts of the deck.

224. Quadrantal Deviation is so named from its being easterly and westerly, alternately, in the four quadrants as the ship moves round a complete circle of azimuth. It is caused by the transient or inductive magnetism of horizontal soft iron, such as iron deck beams, the iron spindle of the wheel, &c. It is zero or nothing when a ship's head is near the North, South, East, or West points, and greatest on the quadrantal points. It is generally easterly in the N.E. and S.W. quadrants, and westerly in the N.W. and S.E. quadrants of the compass. Quadrantal deviation remains unchanged in all magnetic latitudes, and provided that the iron in the ship be of good quality, the quadrantal deviation will be little, if at all, altered by the lapse of time.

To illustrate the way in which horizontal soft iron produces quadrantal deviation, let us suppose the whole of the induced magnetism in a ship to be represented by the soft iron bar B in Fig. 9. This cannot be so in actual practice, because the athwartship horizontal iron produces quadrantal deviation as well as the fore-and-aft iron, but we may suppose it may for the sake of clearness. The small circles represent the compass, the thick lines within the small circles the compass needle, the dotted lines within the compass the magnetic meridian. Beginning at north, it will be observed that the bar B is parallel with the magnetic meridian, and will therefore be an inductive magnet while it is in or near that position (215), its after end, marked S, being a south pole; but as the bar B is in a line with the compass needle N, it cannot exert any deviating power upon the needle, either to the right or left. As the ship's head swings towards the N.W., the relative positions of

the bar B and the needle N are altered, and the south end of the bar draws the north end of the needle to the left from N to N'. As the ship's head approaches the west, the bar B loses its polarity, and at west it is at rightangles to the magnetic meridian, and ceases to exert any influence on the compass. The ships head now swings towards the S.W., and the bar B, as it turns towards the south pole, again becomes an inductive magnet; its after end being a north pole, and drawing the south end of the compass needle from S to S'. When the ship's head reaches south there is no quadrantal deviation, because the bar B is in a line with the compass needle. As her head swings towards the S.E., the needle is drawn from S to S', causing westerly deviation. At east there is no deviation, for the same reason that there was none at west. After passing east, the after end of the bar B becomes a south

pole, and draws the north end of the needle to the right-hand in the N.E. quadrant. As the ship's head approaches the north, the quadrantal deviation gradually decreases until it becomes nothing at north. The reader will observe that the bar B in this case produces easterly deviation in the N.E. and S.W. quadrants, and westerly deviation in the N.W. and S.E. quadrants. Cases may arise where the deviation is westerly in the N.E. and S.W. quadrants, but they are very rare.

225. The constant part of the deviation is generally very small, and is the same for every point of the compass, it often arises from defects in the compass itself. An error in the correct magnetic bearing of a distant object used to ascertain the deviation, will give an apparent constant deviation: for example, if the correct magnetic bearing of a lighthouse be S. 46° E., and the observer assumes it to be S. 44° E., and finds the deviation by it, there will be an error of 2° in the deviation thus found on every point of the compass; or, in other words, the westerly deviation will be 2o less, and the easterly deviation will be 2° more than it ought to be. When a ship is swung hurriedly, and her head is not allowed to remain for a minute or two on any point before observations are made, there is a temporary constant deviation produced; and this temporary deviation is easterly when the ship is swung to the left, as from East to North, and is westerly when the ship is swung to the right, as from North to East.

226. Mechanical Compensation or Correction of the Compass by means of Magnets and Soft iron.-These adjustments were first proposed by Mr. AIRY, the Astronomer Royal, and are now universal in the merchant service.

227. Correction of the Semicircular Deviation.-As this error is caused by magnetism induced in vertical iron, and by subpermanent magnetism acquired in building, it is very difficult in practice to separate the semicircular deviation caused by vertical iron from that caused by subpermanent magnetism. Could this be easily effected, a natural inference would be to compensate the former by vertical soft iron, and the latter by a magnet, each compensator placed so as to produce opposite effects to those of the ship.

An arrangement of this kind was adapted to several ships by Mr. Rundell, the Secretary of the Liverpool Compass Committee, who placed a vertical iron bar before the compass, leading from the keel (when possible) to a point some height above the level of the card. The distance from the compass and height to which it must be carried to be determined by experiment, but a difficulty has been experienced in adjusting this method so as to ensure

success.

The order of proceeding for compensation is as follows:-
:-

The ship must be upright, or on an even beam, with all her iron stores on board, in the positions which they are intended to occupy while at sea.

The position of the binnacle being decided on, draw a line upon the deck, fore-and-aft, through the centre of the place where the binnacle is to stand. Draw another line across the deck, at right-angles to the former, through the same centre,

Provide two or more powerful magnets from 18 inches to 2 feet in length.* Let the ship's head be swung to the North or South, correct magneticeither of these points will do. When the ship's head is steady at one of these points, observe whether there is any deviation; if there is any, lay one of the magnets on the deck athwartship, with its centre exactly on the foreand-aft line drawn on the deck at some distance from the binnacle; move it gradually (not hurriedly) to or from the foot of the binnacle until the compass points correctly. The magnet may be placed either before or abaft the binnacle, whichever is most convenient, but its centre must always be over the fore-and-aft line drawn on the deck, and it must be kept at right-angles to the ship's keel. If the compass needle deviate to the left, the north end of the magnet must be placed to the left, and conversely.

After the compass has been made to point correctly at either the north or south points, swing her head round to the east or west correct magnetic (either will do), and steady her head on one of these points.

If there be any deviation, place the other magnet fore-and-aft, either on the port or starboard side of the binnacle, with its centre on the athwartship line drawn on the deck; move it tc or from the foot of the binnacle until the compass points correctly.

If the ship was built with her head nearly north or south, two magnets may be required. These may either be placed one on each side of the compass, or both on the same side, as may be convenient; if placed on the same side of the binnacle, lay them an inch apart, and under all circumstances parallel, but always similar ends (N. or S.) directed to the same part of the ship.

The adjuster should be careful to see that the centre of the magnet is kept on the fore-and-aft line, so that one of the poles of the magnet be no nearer the binnacle than the other.

It must be remembered that this correction will only hold good for a small range of latitude, and while the ship's magnetism continues in the same state as when the correction was made.

228. Correction of the Quadrantal Deviation.-As the quadrantal deviation is caused by the action of horizontal soft iron, a natural inference is, that soft iron should be used for compensating this error, so placed as to cause opposite effects to those of the ship. The compensations for semicircular deviation being complete, the ship's head should be swung to one of the intercardinal points, N.E., N.W., S.E., or S. W., correct magnetic; and the binnacle being fitted with two small brass boxes, one on each side of and on a level with the compass; if there is any deviation place a quantity of small iron chain in the boxes until the compass points correctly. For greater certainty swing the ship to each of the other quadrantal points.

* Compensating magnets should be from 10 to 18 or 24 inches in length, their breadth one-tenth of their length, and their thickness one-fourth their breadth.

U