of making a deviation table for the various points of one or more compasses, or of correcting the deviation by permanent magnets and soft iron.

But it should be well understood that, for use on board ship, no instrument is preferable to the dumb card, or compass card without a needle, from the facilities it offers in the comparison of compass readings, and the rapidity with which the correct magnetic direction of the ship's head may be determined.

In placing a compass have it well raised above the iron beams, as far as convenient from an iron bulkhead, and, if possible, not within six or seven feet of any vertical iron. Take care to have the lubber's point in the midship fore-and-aft line.

The card of a compass, to be properly adjusted, should not have less than two parallel needles of great directive force, placed edgeways, fixed to the card, and 60° apart. The pivot on which it moves should be of hard steel, well pointed but not needle-like, and perfectly smooth. The central stone should be a hard gem, with the concavity well rounded and polished; an agate is not to be trusted, though some are very hard. The card resting on its pivot should be truly horizontal. The ship should be perfectly upright.

These conditions being fulfilled we may proceed to adjust compasses; and if you choose, and cannot trust your eye, you may draw two chalk lines-one fore-and-aft, and the other athwartship -crossing each other under the centre of the compass.

With the ship's head on correct magnetic North or South, if the compass by its indication shows any deviation, the compensating bar-magnet must be placed athwartship, either before or abaft the compass (it matters not which), the middle of the magnet on the midship fore-and-aft line with the centre of the compass, and the N-marked end (red pole) of the magnet, directed to that side of the ship towards which the N end of the compass-needle appears to be drawn: thus, the red (N) pole of the bar magnet must be to starboard if the deviation is Easterly with the ship's head North, or Westerly with the ship's head South (this is + C); but the red (N) pole of the magnet goes to port if the deviation is Westerly with the ship's head North, or Easterly with the ship's head South (this is - C). Begin

by placing the magnet at a distance from the compass, and then gradually approach it until the compass shows, by the lubber's point, that it indicates (within 1°) the correct magnetic point (N. or S.) to which the ship's head has been brought by the bearings. You must rather under-adjust than over-adjust; this can be remedied, if need be, when coming to the opposite point. Finally, with a small magnet, draw the compass about 8° or 10° from its lubber's point indication, and note if it returns to, and steadies at, its former apparent direction: does it do so, you are sure that the correction is good. Throughout this operation, be careful to observe that the ship's head is on its N. or S. azimuth.

Having commenced at North or South, next proceed to East or West; it matters not which. The method is similar to that already described, except in so far as regards the disposition of the compensating magnet.

Having brought the ship's head correct magnetic East or West, then, if there is any deviation, the compensating bar-magnet must be placed in a fore-and-aft direction, either on the starboard or port side of the compass (it matters not which), the middle of the magnet on the 'thwartship line with the centre of the compass, and the N-marked end (red pole) of the magnet directed to that part of the ship towards which the N. end of the compass needle appears to be drawn: thus, the red (N) pole of the bar-magnet must be aft if the deviation is Westerly with the ship's head East, or Easterly with the ship's head West (this is B); but the red (N) pole of the magnet goes forward if the deviation is Easterly with the ship's head East, or Westerly with the ship's head West (this is + B). For the rest proceed exactly as before indicated for the correction of the N. or S. point, not forgetting to under-adjust rather than over-adjust; seeing that the ship is steadied to the azimuth; and testing the accuracy of the adjustment.

-

A few diagrams will illustrate, in respect to the compass, the disposition of the compensating magnets.

Fig. 34 shows a ship, built head East (giving co-efficient + C), brought with her head correct magnetic North, and the consequent attraction of the N. end of the needle to the starboard side compensated by a 'thwartship bar-magnet, the N.-marked (red) pole of

which lies to starboard, and thus repels the N. end of the compass needle to its normal correct magnetic direction.

Fig. 35 shows a ship, built head West (giving co-efficient - C), brought with her head correct magnetic North, and the consequent attraction of the N. end of the needle to the port side compensated by a 'thwartship bar-magnet, the N.-marked (red) pole of which lies to port, repelling the N. end of the compass needle to its normal correct magnetic direction.

Fig. 36 shows a ship, built head South (giving co-efficient + B), brought with her head correct magnetic East, and the consequent attraction of the N. end of the needle towards the bow compensated by a fore-and-aft magnet, the N.-marked (red) pole of which lies forward, and thus repels the N. end of the compass needle to its normal correct magnetic direction.

Fig. 37 shows a ship, built head North (giving co-efficient - B), brought with her head correct magnetic East, and the consequent attraction of the N. end of the needle towards the stern compensated by a fore-and-aft magnet, the N.-marked (red) pole of which lies aft, and thus repels the N. end of the compass needle to its normal correct magnetic direction.

The bar-magnets in these illustrations have been shown to

correct the deviation of ships built heading on one or other of the four cardinal points; but it is only occasionally that a magnet in one position (athwartship or fore-and-aft) is sufficient for the compensation, since the majority of ships must be built on one or other of the remaining 28 intercardinal points; and thus a magnet in each of the two positions will most probably be required. Hence, besides the disposition of the magnets in Figs. 34, 35, 36, and 37 there must be four others,

viz., for the combinations of B and C as given at the foot of p. 804; and Fig. 38 may be taken to represent a ship, after adjustment, lying head East, but having been built heading in a N.Wly. direction, and hence with-B and -C; the disposition of the com

pensating magnets is a combination of Figs. 35 and 37.

A ship built head N.Ely. (B and C) would require a combination of Figs. 37 and 84; built S.Ely. (+ B and + C), a combination of Figs. 36 and 34; and built S.Wly. (+ B and C), a combination of Figs. 36 and 35.

The bar-magnets compensate only the semicircular deviation.

Fig. 38 also illustrates the position of the correctors on the starboard and port sides of the compass for a positive quadrantal deviation. The ship having been swung for the correction of two adjacent cardinal points, and the magnets placed in their respective positions, she is then brought to one of the quadrantal points, N.E., S. W., N.W., or S.E.; the first two points have Easterly, and the last two Westerly, deviation, for which soft iron in some form, and to starboard and port, is applied on the same level as the compass needles. A minus quadrantal deviation, which would require the correctors to be before and abaft the compass, is so rare, and even then of such small amount, that it may be dismissed by just mentioning its possible existence. The correctors may be soft iron chain in brass boxes; or elongated or spherical cast iron. I am surprised that no enterprising optician has tried a large globular receptacle of brass into which a suffi

cient quantity of soft iron chain could be put, for the purpose of this compensation: I believe it would answer better than cast iron. To suppose, as I have seen it stated, that it would want tons of that metal for the correction, is simply an absurdity.

Reverting to the compensation, by bar-magnets placed parallel with the deck, of the semicircular deviation represented by the coefficients B and C, it will be seen, from Figs. 34, 35, 36, and 37, that in each case the correcting magnet trends in the direction of the line of force, i.e., of the magnetic meridian, as it passed through the ship at the time she was on the stocks; hence for ships built N. and S. the magnet naturally lies fore-and-aft, and for ships built E. and W. it lies athwartship, since the line of force would at the time trend athwartship. But a ship built in the direction of any of the intercardinal points would have the line of force making an oblique angle with the keel. Now, though two magnets, laid in different directions, one fore-and-aft, and the other athwartship, are used to correct B and C, nevertheless a single magnet, so placed that it shall make the same oblique angle with the keel that did the line of force while the ship was building, would be sufficient for the purpose of correcting the deviation due to the ship's subpermanent magnetism. The respective values of B and C, plus and minus, are the data whence are derived the angle of direction for the magnet, and the total force to be corrected; thus, the maximum semicircular deviation; and

C

B

=

tan a = the angular direction of the magnet with the keel; and these, having due regard to the signs of the co-efficients, indicate whether the magnet that is placed horizontally beneath the compass, and at an oblique angle with the keel, must have its N.-marked (red) pole forward or aft, to starboard or port.

But the ordinary Traverse Table suffices for the computation. Suppose+B 23° 6 and C 11°; then, with 23-6 in Lat. column. and 11 in Dep. column, we get in the Dist. column 26° as the semicircular deviation to be corrected, and the course of 25° is the angle that the magnet should make with the direction of the keel, the N-marked (red) pole pointing to the port bow.