vessels have no stability they will not roll at sea. Can you believe it? And yet it is upon that basis that Mr. White sets to work with his mathematics. Sir William Thomson has just looked at the calculations, and assumed that the statement of the case to which the mathematics were applied, was true. With respect to the Scientific Committee, it accepted Mr. Froude's principle as true, and made a report accordingly, but in the interim between the first and second report, the " Invincible " nearly capsized, and then what follows? Why, a practical repudiation of the theory; they say if it is possible to give more stability without increasing the rolling motions, such a thing would be very desirable. I say, what is giving increased stability but a practical break-down of the whole theory? The recommendation as to the "Inflexible" is to increase her breadth 10 feet and double her stability. That is a repudiation of the whole principle, which has been as little stability as possible, because increased stability produces short periods, and that is dangerous. Why, here they are producing short periods. How? By breadth; and they are producing short periods by these bilge-keels; the very thing they profess to dislike as dangerous they are bringing about, showing plainly enough they do not believe in the theory they have set up.

Admiral NOLLOTH: It appears to me that, taking the most favourable view of the theory of little stability, disadvantage preponderates.

If waves could be made to order, and the ship were constructed with a rolling period of the most suitable kind to avoid rhythm, synchronism of lee-lurch and waveimpact would still, as at present, frequently take place; and with ships of long period and concomitant large arcs of roll, I think the remedy would be worse than the disease.

In the case of huge ironclads with great momentum of sides when once in motion, and heavy wall-sided seas with no friendly slope gradually passing to leeward, and mitigating before arresting and finally reversing the direction of roll, little stability might prove of fearful importance.

In my humble opinion, decided advantage should be clearly established before consent be given to great restriction of the stability element in either ironclad or ordinary sailing-vessels.

Mr. SCOTT RUSSELL: I entirely agree with Admiral Nolloth that the theory of slender stability with long periods of roll is false and dangerous.

The CHAIRMAN: It now remains for me to give our thanks to Admiral Fishbourne, and on behalf of the Institution I may safely say we are all very glad indeed to welcome facts.

NOTE.-In justice it should be stated I was not aware how great was the esti mated stability of the "Inflexible" when I read this paper. I had supposed, from Mr. E. J. Reed's statements, that she could have little more than he was in the habit of giving, i.e., 3 feet of meta-centric height after 350 tons of ballast had been added, or less than 4 feet in the "Devastation," with an empty bottom capable of containing near 1,000 tons of water, whereas, the "Inflexible's" meta-centric height is 85 feet higher than almost any vessel's, except American monitors and the Russian circular ship, which has 8.9 meta-centric height.

Then as to her capsizing when well-nigh destroyed, it is only what sailors have to expect, as in the case of "Java" "Gueriere," and others. This chance can be reduced only by reducing the deep double bottoms introduced by Mr. Reed to "check rolling," erroneously, for it has the opposite effect, this has in some measure been done in the "Inflexible," or remedied by making ships unsinkable, as out in my lecture on the loss of the "Captain" in 1872, page 30. This has in part been attempted in the "Inflexible," by cork and other devices, while there is nothing pointed of the kind in Mr. Reed's designs. So with equal accuracy of five vessels of Mr. Reed's design with meta-centric heights of 22 or 24 or even 3 feet, without clothing, or only 4 or 5 inches thick, would fail, while the "Inflexible" would practically be still safe.

It must be a subject of congratulation that the theory that for "a vessel to be safe she must have very little initial stability," has been abandoned; for besides the damage entailed on many ships of the Navy, it had found its way into the Merchant Service, and had occasioned the foundering of several vessels and the loss of some of their crews.-E. G. F.

Evening Meeting.

Monday, February 4, 1878.

LIEUTENANT-GENERAL SIR JOHN H. LEFROY, R.A., K.C.M.G., C.B., F.R.S., in the Chair.

ON COMPASS ADJUSTMENT IN IRON SHIPS, AND ON A NEW SOUNDING APPARATUS.

By SIR WM. THOMSON, LL.D., F.R.S., Pres. R.S.E., Professor of Natural Philosophy in the University of Glasgow, and Fellow of St. Peter's College, Cambridge.

I-New Form of Marine Azimuth and Steering Compass with Adjuncts for the complete application of the Astronomer Royal's principles of Correction for Iron Ships.

THIRTY-EIGHT years ago, the Astronomer Royal showed how the errors of the compass, depending on the influence experienced from the iron of the ship, may be perfectly corrected by magnets and soft iron placed in the neighbourhood of the binnacle. Partial applications of his method came immediately into use in merchant steamers; and, within the last ten years, have become universal not only in the merchant service, but in the navies of this and other countries. The compass and the binnacles before you are designed to thoroughly carry out in practical navigation, the Astronomer Royal's principles. The general drawback to the complete and accurate realisation of plans for carrying out these principles heretofore, has been the great size of the needles in the ordinary compass which renders one important part of the correction, the correction of the quadrantal error for all latitudes by masses of soft iron placed on the two sides of the binnacle, practically unattainable; and which limits, and sometimes partially vitiates the other chief part of the correction, or that which is performed by means of magnets placed in the neighbourhood of the compass. Five years ago my attention was forced to this subject through my having been called upon by the Royal Society to write a biographical sketch of the late Archibald Smith, with an account of his scientific work on the mariner's compass and ships' magnetism, and I therefore commenced to make trial compasses with much smaller needles than any previously in use; but it was only after three years of very varied

trials, in the laboratory and workshop, and at sea, that I succeeded in producing a mariner's compass with the qualities necessary for thoroughly satisfactory working in all weathers and all seas, and in every class of ship, and yet with small enough needles for the perfect application of the Astronomer Royal's method of correction for iron ships. One result at which I arrived, partly by lengthened trials at sea in my own yacht, and partly by dynamical theory analogous to that of Froude with reference to the rolling of ships, was that steadiness of the compass at sea was to be obtained not by heaviness of needles or of compass card, or of added weights, but by longness of vibrational period' of the compass, however this longness is obtained. Thus, if the addition of weight to the compass card improves it in respect to steadiness at sea, it is not because of the additional friction on the bearing point that this improvement is obtained; on the contrary, dulness of the bearing point, or too much weight upon it,

The vibrational period, or the period (as it may be called for brevity) of a compass, is the time it takes to perform a complete vibration, to and fro, when deflected horizontally through any angle not exceeding 30° or 40°, and left to itself to vibrate freely.

renders the compass less steady at sea, and, at the same time, less decided in showing changes of the ship's head, than it would be were the point perfectly fine and frictionless, supposing for the moment this to be possible. It is by increasing the vibrational period that the addition of weight gives steadiness to the compass; while, on the other hand, the increase of friction on the bearing point is both injurious in respect to steadiness, and detrimental in blunting it or breaking it down, and boring into the cap, and so producing sluggishness, after a short time of use, at sea. If weight were to be added to produce steadiness, the place to add it would be at the very circumference of the card. My conclusion was that no weight is in any case to be added, beyond that which is necessary for supporting the card; and that, with small enough needles to admit of the complete application of the Astronomer Royal's principles of correction, the length of period required for steadiness at sea is to be obtained, without sacrificing freedom from frictional error, by giving a large diameter to the compass card, and by throwing to its outer edge as nearly as possible the whole mass of rigid material which it must have to support it.

In the compass before you, these qualities are given by supporting the outer edge of the card on a thin rim of aluminium, and its inner parts on thirty-two silk threads or fine copper wires stretched from the rim to a small central boss of aluminium, thirty-two spokes, as it were, of the wheel. The card itself is of thin strong paper, and all the central parts of it are cut away, leaving only enough of it to show conveniently the points and degree divisions of the compass. The central boss consists of a thin disc of aluminium, with a hole in its centre, which rests on the projecting lip of a small aluminium inverted cup mounted with a sapphire cap, which rests on a fixed iridium point.

Eight small needles from 34 inches to 2 inches long, made of thin steel wire, and weighing in all 54 grains, are fixed like the steps of a rope ladder on two parallel silk threads, and slung from the alumi

nium rim, by four fine copper wires through eyes in the four ends of the outer pair of needles.

The weight of the central boss, aluminium cup, and sapphire cap, amounts in all to about five grains. It need not be more for a 24-inch than for a 10-inch compass. For the 10-inch compass the whole weight on the iridium point, including rim, card, silk threads, central boss, and needles, is about 180 grains. The limit to the diameter of the card depends upon the quantity of soft iron that can be introduced without inconvenient cumbrousness on the two sides of the binnacle to correct the quadrantal error. If, as sometimes may be advisable in the case of a steering compass, or of a pole or masthead compass, it be determined to leave the quadrantal error uncorrected, the diameter of the compass card may be anything from 12 to 24 inches, according to circumstances. A 24-inch card on the new plan will undoubtedly have less frictional error or "sluggishness" for the same degree of steadiness than any smaller size; but a 12-inch card works well even in very unfavourable circumstances, and it will rarely, if ever, be necessary to choose a larger size unless for convenience to the steersman for seeing the divisions, whether points or degrees. Specimens of 12-inch, 15-inch, and 24-inch pole compasses have been made. The last mentioned may be looked at with some curiosity as being probably the largest compass in the world. It will no doubt be properly condemned as too cumbrous for use at sea even in the largest ship, but there can be no doubt it would work well in a position in which a smaller compass would be caused to oscillate very wildly by the motion of the ship.

The period of the new 10-inch compass is in this part of the world about 40 seconds, which is more than double the period of the A card of the Admiralty standard compass, and is considerably longer than that of the ordinary 10-inch compass, so much in use in merchant