tion before us more in detail than would have been possible if only general principles were discussed.

The committee beg to state that the comments made regarding this system may be applied to nearly all other systems proposed, bearing in mind, however, that changes in particulars may influence the result to which we have come, but only in an unimportant degree.

The light-system of Lieutenant Fletcher consists in making use of a range of three lights, which have been placed as follows:

Forward of the foremast, at a distance apart not less than the beam of the ship, two bright white range-lights of the same character as the present masthead-light, visible from ahead to two points abaft the beam on either side. These lights shall be in the vertical plane of the keel, with the lower one so placed forward of the upper one that an imaginary line through them will make an angle of 55 degrees with the vertical. The third light is placed near the stern on the taffrail and screened so as to throw a light from right astern to two points forward of the beam.

When right ahead of a steam ship an imaginary line through her two range-lights will appear vertical if the ship be on an even keel.

On the starboard bow of a steam-ship an imaginary line through her two range-lights will appear inclined from the vertical to the right, and on the port bow to the left from the vertical.

When more than four points on the bow of a steam-ship her rangelights will appear inclined more than 45 degrees, and at six points her stern-light is visible.

If we imagine that an observer is stationed at a great distance and a ship, lighted in the proposed manner, turned around the perpendicular through her middle range-light, then it will be seen that the horizontal distance between the masthead-light and the forward range-light does not increase in the same ratio as the angle between the visual line of the observer and the keel of the ship, but only in the ratio of the sines of this angle. (See Fig. 1, in which the distance of the observer is assumed to be indefinite, and, in consequence, the visual lines parallel to each other.)

By a glance at the accompanying diagram (Fig. 1) it is easily seen that the lights at an angle of six points appear at the same distance as at an angle of ten points. The same is true with regard to the angles of seven and nine points, and others which may be symmetrically placed to the line, Ll, up to two points of this line.

The horizontal distance at which the two forward range-lights at different angles appear has been calculated, and will be found in column b of the following table for a perpendicular distance between the lights of 21.2 feet, which makes the horizontal distance between the lights for an angle of 55 degrees = 30 feet.

We see from column b that the horizontal distance under which the two range-lights are seen, after the ship has turned four points, has become 21.2 feet, and at this point is equal to the perpendicular distance between the lights; the triangle formed by the horizontal line through the forward range-light, the perpendicular through the foremastheadlight, and the line connecting the lights, has become equilateral, both angles at the connecting line (Ll, Fig. 2, being 45°).

Now, order a look-out to watch carefully, and give him the following instructions: That when he sees the two white lights, one below the other, the ship which shows them is heading directly for him; that if a plumb line from the masthead-light appears to be separated from the forward light one-quarter of that distance, her course makes an angle of two points; and that if the distance is equal to three-quarters, or equal to the whole of the perpendicular, then that ship's course will be three or four points inclined to his eye-line.

Such instruction may be understood by a sailor and may not. But if he should understand it and is accustomed to judge distances somewhat correctly, he will be able to determine the other ship's course pretty accurately, for, as column d shows, there is only a small difference between columns b and c, and it appears probable that, as Commander Chad. wick found to be the case during the trials he made, an ordinary sailor may very well judge by this method whether a ship's course is inclined to his eye-line one, two, three or four points, without probably making a greater mistake than half a point. (See report, page 7, ¶ 4.)

In Sector B, Fig. 3 (that is, from four to six points), it would become more difficult to judge the course correctly; because here the horizontal alteration of distances between the lights becomes much smaller for the same alteration of the course of the ship, but the error will probably not become more than one point. At six points the after rangelight would come in view, but it will probably not be possible to estimate the bearing in sector C more closely than to say, that an observer is

"abeam" of the other vessel. Errors up to two points may, on this account, be expected.

Further aft than ten points from right ahead, both the forward rangelights, as well as the side-lights, will be shut in, and by the stern-light alone it will only be possible to say that an observer is within six points on either quarter.

The advantages of Lieutenant Fletcher's system of range-lights, and all others similar to this, consist, therefore, in that it affords the means by which the course of a ship can be found by simple observation and comparison alone, by an observer stationed in a sector from right ahead to six points on either side (Sectors A and B, Fig. 3), with tolerable accuracy, say within one point, provided, however, that the ship observed has no perceptible heel.

If this should be the case, then the indication of the course will become less correct, and be changed in a ratio of about three-quarters of the amount of heel. This will be seen by reference to Fig. 4, in which S is the ship seen by an observer stationed right ahead; L and I the two range-lights and a the angle of heel. The two lights appear not in a vertical line, one below the other, but inclined, and the lower light to the right of the perpendicular through L, at a distance llLl sin a.

If we assume, as before, Ll=21.2 feet, and a=120, then ll, is 4.4 feet; if we compare this with the table page 73, column b, we find that this would correspond to a change of course of about three-quarters of a point (more exactly: 8.5°), and that the observations of the ship's lanterns lead in this case to an error of this amount, whereas if the ship had been on an even keel the indication of the ship's course would have been more correct in this position than in any other.

True, in the present case, the side-lights would have been visible and could have been used to correct this error.

But this could not have been done if the ship's head had been turned from one to two points, where the error would have been about the

same.

From the foregoing it follows that the use of range-lights on the high seas does not offer the same advantages as have been found to be derived from it by vessels navigating inland waters, and this should not be lost sight of in coming to a conclusion regarding the introduction of such a system.

A practical difficulty will also be found in the placing of the forward range-light according to the demands of this system. For if a ship's foremast is placed 60 feet from the bow, and the foremasthead-light at a height of 30 feet above the hull, the forward range-light, with a minimum vertical distance of 15 feet, would be placed 15 feet above the hull and 34 from the bow; that is, in a place where it is presumably difficult to keep it, if a fast steam-ship is going at full speed against a head sea.

If such a ship be provided with a turtle deck it would probably be

difficult to hoist or lower the lantern without slackening the speed and to control the burning of the light.

It has also to be kept in view that a white light, accidentally shown somewhere about the stern, for instance, a bright binnacle light, could be mistaken for the stern-light, and thus lead to the belief that the observer is abeam of a vessel, when he is actually in a sector within six points from ahead.

It has to be remembered, too, that what a seaman especially wants to find out when he watches another ship is whether she is proceeding in such a manner as to involve risk of collision.

But the knowledge of the course of the ship alone is not sufficient to determine this. It would remain necessary to watch the compass-bearing of the ship in sight, whether she be fitted with range-lights or not. The greatest advantage of the above system, in the opinion of the committee, is that a small change in the course of a steamer approaching end on, or nearly so, is at once and unmistakably indicated.

Lieutenant Fletcher also proposes that the foremast-light be occulted every half minute for steamers going at a greater speed than 13 knots. This proposition can not be recommended by the committee, as the construction of such a lantern is too complicated. It would, moreover, give only a very approximate idea of the speed of steam-ships, as these are known to proceed at the present day at any speed up to 20 knots and more.

Conclusions.-Range-lights, if properly placed and fitted, indicate, under certain circumstances, a ship's course in a more accurate way than at present. This is of some advantage, and therefore, in the opinion of the committee, the optional introduction of some such system may be proposed for the favorable consideration of the Conference, adding, however, that no definite system should be decided upon for adoption before this subject has been more fully studied, and before a larger number of experiments under different circumstances have been carried out.

B.-SYSTEM IN WHICH TWO OR MORE SIDE LIGHTS ARE EMPLOYED ON THE SAME SIDE.

There are certain general objections to the introduction of a system which, in some cases, would require double the number of lanterns carried at present, and more, and it is considered advisable to state these before going into further particulars.

(1) It has been found very difficult to compel certain classes of ships to carry the lights prescribed for them by the regulations actually in force. It has to be expected that had they to carry double the number of lanterns there would result presumably still more cases of non-compliance with the law.

(2) The care of a larger number of lanterns would require more men than could be spared for such service on board of small vessels.

(3) The cost of the outfit of every vessel, and the cost of providing the illuminant would be considerably increased. This would probably not affect the owners of large steam-ships, but would not easily be borne by the man who owns only a portion of a small sailing vessel. Of the proposals submitted a number were so definite that diagrams showing the number and arrangement of lights proposed could be drawn (see Figs. 5 to 12).

In these A is to be the sector in which two colored side lights are visible; B the sector in which one colored side light; C the sector in which only the white stern light (Article 11) or a white side light (Fig. 12) is visible.

It will be seen by reference to the figures that an observer has only to count the number of side lights visible from his position in order to know in which of the sectors A, B, or C he is. It is evident then that the course of a vessel by the adoption of any one of the proposals submitted could only be ascertained very approximately, and this advantage does not seem to the committee important enough to warrant the recommendation of any of the proposals submitted to the favorable consideration of the Conference.

It is true, that by placing the lights on each side at a certain defined distance from each other, it would be possible to estimate the course of a vessel more approximately than at present, at least when two lights

are seen.

Practically it would be most difficult to comply with such a proposition, and it therefore can not be recommended. In addition to this it ought to be stated that, as far as known to the committee, only a few experiments have been made with systems employing double side lights at sea. The Royal Spanish Navy has tried it already in a squadron, and has found that the advantage gained by double side lights were not so great as to recommend their adoption. Neither have they given satisfaction to the officer in charge of the trials made in the United States (Report on trials quoted above, p. 7, 4), and some of the advantages claimed for these systems remain to be proved before they can be considered as having more than theoretical value.

C.-SYSTEMS IN WHICH THE DIRECTION OF A STEAM-VESSEL IS SHOWN BY PLACING THE SIDE LIGHTS IN A CERTAIN POSITION RELATIVE TO THE MAST-HEAD LIGHT.

It is evident that if the side lights of all steam-vessels were so placed that a vertical plane through the line drawn from them to the masthead light would form a certain known angle with the keel, it would be possible to estimate the direction in which a ship is heading more approximately than at present.

If, for instance, a ship had her lights so placed that the vertical plane through the line connecting the foremast-head light with the side lights