at the bow of the ship, where the current firft begins, and acts fairly upon the palate, in preference to the ftern, where the tumultuous cloting of the waters caufes a wake, vifible to a great diftance. The palate D is funk nearly as low as the keel, that it may not be influenced by the heaping up of the water and the dashing of the waves at and near the water line. The arch K is to afcertain how many knots or miles the would run in one hour at her then rate of failing. But the graduations on this arch must be unequal; because the refiftance of the fpring I will increase as it becomes more bent, fo that the index will travel over a greater fpace from one to five miles than from five to 12. Laftly, the palate, rod, fpring, and all the metallic parts of the inftrument, fhould be covered with a ftrong varnish, to prevent ruft from the corrofive quality of the falt water and fea air. This machine may be confiderably improved as follows: Let the rod or fpear A B (fig. 5.) be a round rod of iron or fteel, and inftead of moving on the fulcrum or joint, as at C, let it pafs through and turn freely in a focket, to which focket the moveable joint must be an nexed, as represented in fig. 6. The rod muft have a fhoulder to bear on the upper edge of the focket, to prevent its flipping quite down. The rod muft alfo pafs through a like focket at F, fig. 5. The joint of the lower focket must be fixed to the bow of the ship, and the upper joint or focket must be connected with the horizontal arm E. On the top of the uppermost focket let there be a fmall circular plate, bearing the 32 points of the mariner's compafs; and let the top of the rod AB come through the centre of this plate, fo as to carry a small index upon it, as is reprefented in fig. 4. This small index muft be fixed to the top of the rod on a square, so that by turning the index round the plate, the rod may alfo turn in the fockets, and of courfe carry the palate D round with it; the little index always pointing in a direction with the face of the palate. The fmall compass plate should not be faftened to the top of the focket, but only fitted tightly on, that it may be moveable at pleasure. Suppofe then the intended port to bear SW. from the place of departure, the palate must be turned on the focket till the SW. point thereon looks directly to the fhip's bow; fo that the SW. and NE. line on the compass plate may be precifely parallel with the fhip's keel, and in this position the plate muft remain during the whole voyage. Suppofe, then, the ship to be failing in the direct courfe of her intended voyage, with her bowfprit pointing SW. Let the little index be brought to the SW. point on the compafs plate, and the palate D will neceffarily prefent its broad face toward the port of deftination; and this it must always be made to do, be the fhip's courfe what it may. If, on account of unfavourable winds, the ship is obliged to deviate from her intended courfe, the little index must be moved so many points from the SW. line of the compass plate as the compass in the bin. nacle fhall fhow that the deviates from her true courfe; fo that in whatever direction the fhip fhall fail, the palate D will always look full to the SW. point of the horizon, or towards the port of def. tination, and confequently will prefent only an

oblique furface to the refifting medium, more or lefs oblique as the fhip deviates more or less from the true courfe of her voyage. As, therefore, the refiftance of the water will operate lefs upon the palate in an oblique than in a direct pofition, in exact proportion to its obliquity, the index II will not how how many knots the veffel runs in her then courfe, but will indicate how many the gains in the direct line of her intended voyage.-Thus, in fig 9. if the fhip's course lies in the disection of the line AB, but he can fail by the wind to nearer than AC; fuppofe, then, her progreffive motion fuch as to perform AC equal to five knots or miles in an hour, yet the index II will only point to four knots on the graduated arch, because she gains no more than at that rate on the true line of her voyage, viz. from A to B. Thus will the difference between her real motion and that pointed out by the index be always in proportion to her deviation from her intended port, until he fails in a line at right angles therewith, as AD; in which case the palate would prefent only a thin fharp edge to the refifting medium, the preffure of which fhould not be fuffi-` cient to overcome the friction of the machine and the bearing of the spring L. So that at whatever rate the thip may fail on that line, yet the index will not be affected, fhowing that the gains no thing on her true course. In this cafe, and alfo when the veffel is not under way, the action of the fpring L fhould cause the index to point at O, as reprefented by the dotted lines in fig. 5. and 8. As the truth of this inftrument muft depend on the equal pressure of the refifting medium upon the palate D, according to the hip's velocity, and the proportionable action of the spring L, there should be a pain or screw at the joints Cnd F, fo that the rod may be readily unshipped and taken in, in order to clean the palate from any foulness it may contract, which would greatly increafe its operation on the index H, and thereby render the graduated arch falfe and uncertain. Further, the spring L may be exposed too much to injury from the falt water, if fixed on the outfide of the fhip's bow. To remedy this, it may be brought under cover, by conftructing the machine, as reprefented by fig. 8. where AB is the rod, C the fulcrum or centre of its motion, D the palate, E the horizontal arm leading through a small hole into the forecastle; M is a strong chain ́ fastened at one end to the arm E, and at the other to a rim or barrel of the wheel G, which by means of its teeth gives motion to the femicircle I and index H. The spring L is fpiral and enclosed in a box or barrel, like the main fpring of a watch. A fmall chain is fixed to, and paffing round the barrel, is faftened by the other end to the fuzee W. This fuzee is connected by its teeth with the wheel G, and counteracts the motion of the palate D. N, N, are the two fockets through which the rod AB paffes, and in which it is turned round by means of the little index R. S is the mall compass plate, moveable on the top of the upper focket N. The plate S hath an upright rim round its edge, cut into teeth or notches, fo that when the index R is a little raifed up, in order to bring it round to any intended point, it may fall into one of thefe notches, and be detained there;

other wife

otherwise the preffure of the water will force the palate D from its oblique pofition, and turn the rod and index round to the direction in which the hip fhall be then failing. Should it be apprehended that the palate D, being placed fo far forward, may affect the fhip's fleerage, or obftruct her rate of failing, it should be confidered that a very small plate will be fufficient to work the machine, as one of three or four inches in diameter would probably be fufficient, and yet not large enough to have any fenfible effect on the helm or thip's way. The greateft difficulty, perhaps, will be in graduating the arch K, (if the machine is conftructed as in fig. 5. the unequal divifions of which can only be atcertained by actual experiment on board of each ship respectively, inasmuch as the accuracy of these graduations will depend on three circumftances, viz. the pofition of the fulcrum C with respect to the length of the rod, the fize of the palate D, and the ftrength or bearing of the spring L. When thefe graduations, however, are once afcertained for the machine on board of any one veffel, they will not want any future alterations, provided the palate D be kept clean, and the spring L retains its elafticity. But the unequal divifions of the graduated arch will be unneceffary, if the machine is conftructed as in fig. 8; for as the chain goes round the barrel L, and then winds through the spiral channel of the fuzee W, the force of the main fpring muft operate equally, or nearly fo, in all pofitions of the index, and confequently the divifions of the arch K may in fuch cafe be equal. After all, it is not expected that a fhip's longitude can be determined to a mathematical certainty by this inftrument. The irregular motious and impulfes to which a fhip is continually expofed, make fuch an accuracy unattainable perhaps by any machinery: But if it should be found, as we flatter our felves it will on fair experiment, that it anfwers the purpose much better than the common log, it may be confidered as an acquifition to the art of navigation. It fhould be observed, that in afcertaining a fhip's longitude by a time-piece, this great inconvenience occurs, that a small and trifling mistake in the time makes a very great and dangerous error in the distance run; whereas the errors of this machine will operate no farther than their real amount, which can never be great or dangerous, if corrected by the usual obfervations made by mariners for correcting the common log. A like machine, made in its fimple form (as in fig. 5.) fo conftructed as to ship and unship, might occafionally be applied alongside about midships, in order to afcertain the leeway; which, if rightly shown, will give the fhip's precife longitude. As to fea currents, this and all other machines hitherto invented must be subject to their influence, and proper allowances must be made according to the skill and knowledge of the navigator. Laftly, fome difcretion will be neceffary in taking obfervations from the machine to be entered on the log-book; that is, the most favourable and equitable moment fhould be chofen for the obfervation; not whilft the fhip is rapidly defcending the declivity of a wave, or is fuddenly checked by a ftroke of the fea, or is in the very act of plunging, in all cafes, periods may be found in which a fhip

proceeds with a true average velocity; to dif cover which, a little experience and attention will lead the skilful mariner. An ingenious mechanic would probably conftruct this machine to better advantage in many refpects. The author only meant to fuggeft the principle; experiment alone can point out the best method of applying it. He is fenfible of at least one deficiency, viz. that the little index R, fig. 4. will not be ftrong enough to retain the palate D in an oblique pofition when the fhip is failing by the wind; more especially as the compafs plate S, in whole notched rim the index R is to fall, is not fixed to, but only fitted tight on the focket N. Many means, however, might be contrived to remedy this inconveni-' ence.

(21.) SHIPS, WHICH CANNOT FOUNDER, PRINCIPLES FOR CONSTRUCTING. Patrick Miller, Efq. of Dalfwinton, propofes to conftruct ships and vessels which cannot founder, on the following principles. The vessel is to be kept afloat, with out the aid of it fides, folely by the buoyancy of its bottom, which is flat; the bottom never being fo deeply immersed as to bring the upper surface thereof on a level with the water; fuch veffels not being conftructed for the purpose of carrying cargoes, but for that of carrying paffengers, with the neceflary ftores and provifions; and as thefe veffels are not kept afloat by the aid of their fides, but by the buoyancy of their bottom, as above described, they cannot fink, and therefore pumps are not required, nor are they in any respect neceffary for the prefervation of fuch veffels. The faid veffel is put in motion during calins, and against light winds, by means of wheels. Thefe wheels project beyond the fides of the vessel, and are wrought by means of capftans; the number and the dimenfions of the wheels depend upon the length of the veffel. Thefe wheels are built with eight arms, which confift entirely of plank. Sliders are used to work and to keep the veflel to windward when under fail. These sliders are placed in the centre of the veffel, from flem to ftern; they are made of plank, and the number and dimenfions muft depend on the length of the veffel; and they are raifed and let down either by the hand or by means of a purchase, according to the fize of the veffel. Veffels of this conftruction draw water in proportion to their dimenfions, as follows: a veffel of 40 feet in length, and from 13 to 19 feet in breadth, will draw from 13 to 16 inches of water. One of 50 feet in length, and from 17 to 24 feet in breadth, will draw from 15 to 18 inches of water. One 60 feet long, and from 20 to 28 feet broad, will draw from 18 to 21 inches of water. One 70 feet long, and from 23 to 32 feet broad, will draw from 21 to 24 inches of water. One 80 feet long, and from 27 to 37 feet broad, will draw from 24 to 27 inches of water. One go feet long, and from 30 to 42 feet broad, will draw from 27 to 30 inches of water. One of ico feet in length, and from 33 to 47 feet in breadth, will draw from 30 to 33 inches of water. As, from the principle upon which this veffel is constructed, the cannot fink, the invention must prove a means of faving many lives; and as it will give more room and height between the decks than any veffel of the fame dimenfions of another con

ftruction,