Lighthouse

LIGHTHOUSE

I. LIGHTHOUSE CONSTRUCTION.

THE primary and most important consideration relating to the design and construction of a lighthouse tower which is to be built within the tide mark is the force of the waves which may be expected to assail it, and the directions and heights at which that force will act on the building. The great waves which are found in the open ocean cannot be generated in smaller seas, and, with a due regard to economy in construction, ought not therefore to be provided against. What is wanted is to ascertain in such shorter seas the height of waves in relation to the length of “fetch” in which they are generated, and next to determine their energy when on reaching the shore or a sunken rock, and so ceasing to be waves of oscillation, they enormously increase their destructive force by becoming waves of translation. Full information as to these points and as to the marine dynamometer—an instrument used for ascertaining the force of the waves on an exposed surface—will be found in the article Harbours, to which the reader is referred. It is enough here to state that the law of increase in the height of waves was found by Mr T. Stevenson to be proportional to the square root of the distance from the windward shore, and that the greatest force recorded on rocks exposed to the ocean was 31 tons per square foot. The relative forces of summer and winter gales were found to be as 1 to 3, and the vertical force, after acting on a curved sea wall, was eighty-four times greater than the horizontal force at a height of 23 feet above high water.

The history of the ancient lighthouses is of so scanty a nature that we may pass at once to more modern works, commencing with Winstanley’s Eddystone light.

Winstanley ’ s Eddystone Light.— The Eddystone Rocks, which lie about 14 miles off Plymouth, are fully exposed to the south-western seas. The lighthouse was completed by Winstanley in four seasons. In 1698 it was finished nt a height of 80 feet and the light exhibited; but in 1699, in consequence of damage by storms, the tower was increased by an outer ring of masonry 4 feet thick, and made solid from the foundation to nearly 20 feet above the rock. The height was increased to nearly 120 feet, and completed in 1700. During the well-known hurricane of 20th November 1703 the tower was destroyed. In general design as well as in details this work must be placed among the vetanda of maritime engineering. For example, in plan it was polygonal instead of circular. In his blind devotion to ornamentation Winstanley violated throughout the principles of uniformity of outer profile so as to present great obstructions to the action of the waves.

Budyerd ’ s Eddystone Tower.— This work was commenced in 1706 and completed in 1709, in the form of a frustum of a cone 92 feet high. The work consisted principally of timber, the lower part being oak carefully bolted together, and also to the rock. Above the lower structure of oak courses of stone, cramped together and fixed to the timber work and to the rock, were added in order to give weight to the structure. This lighthouse stood for forty-six years, and was destroyed by fire in 1755. In every respect the simplicity of the structure and the judicious character of the details of the design may be regarded as models of engineering. First, it rested upon a stepped level base, was circular in plan, did not wholly depend upon fixtures but upon weight, preserved a uniform external surface devoid of outside projections and ornamentation; and, above all, the engineer did not by splaying out the base needlessly throw away the small diameter which the rock afforded, but with much judgment adopted the conical form.

Smeaton ’ s Eddystone Tower.— This justly celebrated work, which consisted entirely of stone, was commenced in 1756, and the masonry was finished in 1759. Smeaton was the first engineer who adopted a structure of masonry for a sea tower and dovetailed joints for the stones, which averaged a ton in weight. This work cannot be regarded as a safe model for general imitation in exposed situations, and Rudyetd’s earlier tower was certainly as successful in resisting the forces to which it was exposed. Rudyerd unquestionably selected for so small a rock as the Eddystone a preferable form to that adopted by Smeaton. The sharply curved profile in Smeaton’s design greatly reduced the diameter of the building at a small height above the rock, and so reduced its strength.

Smeaton’s reasoning about the similarity of a tower exposed to the surf and an oak tree resisting the wind was very conclusively shown to be fallacious by the late Mr Alan Stevenson. The arching of the floors, as shown in fig. 1, was also a source of weakness which the introduction of the iron chains, shown black in the diagram, was intended to counteract. Mr Douglass in 1878 stated that “for several years the safety of the Eddystone had been a matter of anxiety and watchful care to the corporation of the Trinity House, owing to the great tremor of the building with each wave stroke.” He also stated that the projecting cornice at the top had been lifted, and that the rock itself had been considerably undermined. A new tower has now (1882) been erected in place of Smeaton’s by Mr Douglass.

Bell Rock Lighthouse Tower.— The Bell Rock, which lies 12 miles off the coast of Forfarshire, is fully exposed to the assaults of the German Ocean. The rock is of considerable extent but of a low level, the tower being covered about 16 feet at high water of spring tides. Mr R. Stevenson, of Edinburgh, when he first landed on the rock, decided to adopt a stone tower as Smeaton had done at the Eddystone, but he deviated largely from that design in the thickness of the walls, in raising the tower to 100 feet instead of 68 feet, and the level of the solid to 21 feet above high water instead of 11 feet. Instead of employing arched floors as at the Eddystone, he adopted lintel stones for the floors which formed part of the outward walls, and were feathered and grooved as in carpentry, besides having dovetailed joggles across the joints where they formed part of the walls. It will be seen on fig. 2 that the floors instead of being sources of weakness, as in Smeaton’s tower, were converted into effective bonds tying the walls together. He also used a temporary beacon or barrack on the rock for the engineer and his workmen to live in while the tower was in progress. The bill intro-dueed into parliament for this work in 1802 was not passed in consequence of financial difficulties. As the Bell Rock was scarcely dry at low water, while the Eddystone was scarcely covered at high water, the commissioners, in order to fortify Mr Stevenson’s views, consulted Mr Telford, and before going to parliament for the second time they also, on Mr Stevenson’s suggestion, obtained for the scheme the support of Mr Rennie, with whom he could afterwards advise in case of emergency during the progress of the work. The second bill was passed in 1806, and the works, which were begun in 1807, were finished in 1810, and the light was exhibited in 1811. The total weight of the tower is 2076 tons.

Skerryvore Lighthouse.— The Skerry vore Rocks, 12 miles off the island of Tyree in Argyllshire, which is the nearest land, are wholly open to the Atlantic. The works, designed and carried out by the late Mr Alan Stevenson, were commenced in 1838 and finished in 1843. The first temporary barrack was destroyed in 1838, and another erected on a more sheltered part of the rock. The tower, which is of a hyperbolic curve, is 138 feet high, 42 diameter at the base, and 16 at the top. Its weight is 4308 tons.

Bishop Rock.— The Bishop Rock, lying off the Scilly Islands, is fully exposed to the Atlantic. It was designed by the late Mr James Walker, and carried out by Afr J. N. Douglass. It is 100 feet above high water, 34 feet in diameter at the base, and 17 at top. The lowest part of the foundation of tower is covered about 19 feet at high water springs; the solid is 20 feet above high water, where the walls are 9 feet thick, and decrease to 2 feet at the top. Owing to the great tremor in this building, it has lately been found necessary to strengthen it by an internal structure of ironwork.

Wolf Rock.— This much exposed rock is about midway between Scilly and the Lizard Point, and is submerged to the depth of about 2 feet at high water. The first design for a lighthouse was in 1823, by Mr R. Stevenson, but it was not till 1862 that a lighthouse was commenced under the superintendence of Mr Douglass, from a design by the late Mr Walker. It is 116∣ feet high, 41 feet 8 inches diameter at the base, decreasing to 17 at the top, and the walls are 7 feet 9∣ inches thick, decreasing to 2 feet 3 inches. The shaft is a concave elliptic frustum, and contains 3296 tons. The lower part of the tower has projecting scarcements in order to break up the sea, but, as has been already stated in the case of the Eddystone, such projections are not in accordance with the principle of uniformity of external surface, and are, therefore, more likely to produce disturbance of the masonry than to add to its stability.

Dhu Heartach Rock Lighthouse.— The Dhu Heartach Rock, which is 35 feet above high water, is 14 miles from the island of Mull, which is the nearest shore. The maximum diameter of the tower, which is of parabolic outline, is 36 feet, decreasing to 16 feet; the shaft is solid for 32 feet above the rock; the masonry weighs 3115 tons, of which 1810 are contained in the solid part. The temporary barrack for the workmen was made of malleable iron bars with an iron drum on the top in which the workmen lived. This tower was designed by Messrs D. & T. Stevenson, and occupied six years in erection, the length of the working season being only about two and a half months in each year.

Chickens Rock Lighthouse.— The Chickens Rock lies 1 mile off the Calf of Man. The curve of the tower, which is 123 feet 4 inches high, is hyperbolic, the diameter varying from 42 feet to 16 feet. The tower is submerged 5 feet at high-water springs. The solid is 32⅜ feet in height, weighing 2050 tons, the whole weight of the tower being 3557 tons. The walls decrease from 9 feet 3 inches to 2 feet 3 inches in thickness. The work was designed by Messrs D. & T. Stevenson, and was begun in 1869 and completed in 1874.

Great Basses Lighthouse near Ceylon.— Great Basses lighthouse lies 6 miles from the nearest land, and was designed by Mr Douglass. The tower has a cylindrical base 32 feet in diameter, above which is a tower 67 feet 5 inches high and 23 feet in diameter. The walls vary in thickness from 5 feet to 2 feet. The tower, including the base, contains about 2768 tons, and the work was finished in three years.

There are several other lighthouses in Ireland, India, and America which merit more attention than our space admits of, and we shall therefore conclude with directing the attention of the engineer to the important influence of the configuration of rocks in modifying the breaking of waves. It cannot but excite surprise that some of the structures which were erected on the Eddystone should have withstood the waves so long as they did. This fact seems to lead to the conclusion that the Eddystone Rock, at one time at least, acted to some extent as a shelter to the structures which were built on it During a summer gale when Dhu Heartach lighthouse was being erected fourteen stones each of 2 tons weight, which had been fixed on the tower by joggles and Portland cement at the level of 37 feet above high water, were torn out and swept off into deep water, as shown on Plate VII. At the Bell Rock stones of 2 tons weight were several times swept away during the construction of the tower, while it is a remarkable fact that no stones were ever moved at the Eddystone. But what is more striking, the thin glass panes of Winstanley’s first tower stood successfully through a whole winter’s storms at the same level above the water as that at which the fourteen heavy blocks were swept away at Dhu Heartach, where it was found necessary from the experience acquired when constructing the lighthouse to raise the solid base of the tower to nearly the same height above the water as the glass panes in Smeaton’s tower, which have hardly ever been broken during the storms of more than a hundred years.

The conclusion then which seems fairly deducible from these facts is that the level of the plane of dangerous impact of the waves above high water depends upon the relation subsisting between their height and the configuration of the rocks above and below high water, as well as perhaps on the configuration of the bottom of the sea near the lighthouse. Thus, while the rock at Dhu Heartach, from its height above high water, forms a great protection against the smaller class of waves, it operates as a dangerous conductor to the largest waves, enabling them to exert a powerful horizontal force at a much higher level than they would had the rock been lower. The lighthouse engineer must therefore beware of taking it for granted that Smeaton’s Eddystone tower is a model for general imitation, and must carefully consider as best he can in what way the configuration of the rock may affect the stability of the tower which he has to design. Unfortunately in the present state of our information no specific directions can be laid down for his guidance iu this matter, but the following general rules of construction may be given :—

(1) The tower should have a low centre of gravity, and sufficient mass to prevent its being upset by the waves. (2) It should be throughout circular in the horizontal plane, and either straight or continuously curved in the vertical plane, so as to present no abrupt change of outline which would check the free ascent of the rising waves, or the free descent of the falling waves, or the free vent of those passing round the tower. All external scarcements in the vertical plane, or polygonal outline in the horizontal plane, are therefore objectionable. (3) Its height, caetcris paribus, should be determined by the distance at which the light requires to be seen by the sailor. The rule for determining this height will be afterwards given. (4) Where the rock is soft, or consists of ledges which are easily torn up, the tower should spring from the foundation-course at a low angle with the surface of the rock, so as to prevent its being broken up by reaction of the waves from the building; or, in other words, the tower must have a curved profile. But special care should be taken to sink the foundation-courses below the surface of the rock, as the superincumbent weight decreases with the sine of the angle of inclination of the wall. If the rock overhangs, owing to the wearing action of the waves, the tower should, if possible, be built at a distance from the place where this dangerous action is in progress. (5) Where the rock is hard and of ample area, the tower may be of such a curved form as will best suit the economic arrangement of the materials, so as to avoid an unnecessary thickness of the upper walls. (6) When the rock is hard, but of small dimensions, the diameter above the base should not be suddenly reduced by adopting a curved profile, but a conical outline should be adopted; and if the rock be hard, but of yet smaller dimensions, a cylindric form of greater height should be adopted so as to thicken the walls, and to increase the weight and therefore increase the friction, which is directly proportional to the weight of the blocks of masonry. In all cases where the rock is small the thickness of the walls should be decreased by steps or scarcements internally, but never externally. (7) The level of the top of the solid part of the tower, and the thickness of the walls above it, should, in different towers having the same exposure, be deteπnined in each case by the level and configuration of the rock and of the bottom of the sea. (8) The best position for the tower is not necessarily the highest part of the rock. It should, in each case, be selected so as to secure the greatest protection in the direction of the maximum fetch and deepest water near the reef. (9) The tower should not, if possible, be erected across any chasm which divides the rock, or in the direction of any gully which projects into it, especially if it be of converging form, which would concentrate wave action. (10) No permanent fixture of the tower to the rock is required for increasing the.stability of the structure. The foundation-course (unless where a curved profile is adopted) becomes, indeed, in the end the most stable of all, because it has the greatest weight above it, to keep it in its place. (11) The stones should, however, be sufficiently connected together, and fixed to the rock, in order to prevent their being washed away during the construction of the work, when they have no superincumbent weight to keep them in their beds. (12) The tower should rest on a truly level base, or on level steps cut in the rock. (13) The pressure of all the materials within the tower should act vertically, so as not to produce a resolved force acting laterally as an outward thrust. (14) The tower should be of such height and diameter, with walls of such thickness, as to prevent the masonry being disturbed by the impact of the waves. (15) The entrance door should be placed on that side of the tower where the length of fetch is shortest, or where from the configuration of the reef and the depth of water the force of the waves is least. This was determined at the Bell Rock by the distribution of the fucus which grew on the lower parts of the tower during the first winter, the vegetation being least where the waves were heaviest. (16) The materials should be of the highest specific gravity that can be readily obtained, and, in some special cases, lead, or dove-tailed blocks of cast iron set in cement, might perhaps be employed.

Plate VII. shows sections on the same scale of a number of the more remarkable lighthouse towers.

Fig. 3 shows an iron pile light erected at Haneda, in the Bay of Yedo, Japan.

Modes of Uniting the Stones and Courses of Masonry.— Fig. 4 (p. 618) shows the mode of combining the stones during construction at different lighthouses in the United Kingdom and in America.

II. LIGHTHOUSE ILLUMINATION.

What is required of every lighthouse apparatus is either the equal distribution of the rays constantly or periodically over the whole horizon, or else their unequal distribution over certain azimuths only.

The first of these two cases, viz., the equal distribution of the light, will be best understood by explaining the different manner in which the rays are operated on by the apparatus for a fixed light and by that for a revolving light. The characteristic of a fixed light, which is that of being seen constantly and always of the same power round the whole horizon, might no doubt be perfectly produced by a naked flame without any apparatus, but then all the rays which did not fall on the navigable track of shipping would be lost to the sailor. In order then to intercept and utilize those rays which, instead of falling on this navigable track, would either go upwards to the sky or downwards on the shore close to the lighthouse tower, and on that part of the sea which is very near the shore, we must have recourse to optical agents both for bending down the rays which naturally point too high, and for bending up those which point too low. It thus appears that the apparatus for a fixed light should bend the rays in the vertical plane only, but should not interfere with their natural horizontal divergence in azimuth.

The demands which are made on a light that has to revolve are not nearly so great as on one that is fixed, for the revolving light does not, like the fixed, require to illuminate the whole horizon simultaneously, but only each point of it at successive intervals of time. When the dark intervals occur, the rays from the flame which are then pointing in the direction of the dark spaces should therefore have their directions so altered laterally as to pass into the adjoining light spaces and thus to increase the power of the luminous flashes. A revolving light, though supplied by a flame of the same power as a fixed, will thus necessarily be far more intense, as it does not lose its power by diffusing the rays constantly over the whole horizon, but gathers them up into a number of separate bundles or beams of great intensity. The apparatus of a revolving light has consequently more optical work to do than that of a fixed, for the rays must be bent not only in the vertical plane but laterally in the horizontal and in all intermediate planes as well.

In the construction of lighthouse apparatus either metallic or glass agents may be employed, but it has been found by experiments that a great saving of light (about 25 per cent.) is effected when glass only is used. All kinds of apparatus may conveniently be ranged either under the catoptric system, where metallic reflexion only is used, the dioptric where the material employed is wholly glass producing refraction and total reflexion, or the catadioptric, in which both glass and metal are employed.

Catoptric System of Illuminating every Azimuth with Light of equal Power either Constantly or Periodically.

Passing over the early and rude expedients of such night marks as open coal fires or naked candles placed in glazed lanterns, we shall confine our attention to the gradual development of those optical designs which are now or werelately adopted for lighthouse apparatus.

Parabolic Reflectors.— In 1763, or at latest before 1777, parabolic reflectors were first used for lighthouse illumination by Mr Hutchinson, dockmaster of Liverpool. In his work on Practica» Seamanship, published in 1777, he states that the Mersey lights were fitted with reflectors (figs. 5 and 6) formed of small facets of silvered glass, and made, as he says “as nearly as they can be to the parabolic curve.” This is unquestionably the earliest published notice of the use of parabolic reflectors for lighthouse illumination. Up to 1782 the wicks of the lamps were of a flat form, but in that year Argand introduced wicks and burners of a hollow cylindric form u¾ich admitted a central current of air through the burner so as to ignite the cone of gas issuing from the wick both within and without. Rumford afterwards split up the cone of gas into several concentric shells.

“It is remarkable,” says Mr J. T. Chance in his excellent memoir {Miιι. Ins. Civil Eng., vol. xxvi.), “how many inventors have contributed their respective parts to the multiple burner: — Argand, the double current; Lange, the indispensable contraction of the glass chimney; Carcel, the mechanism for an abundant supply of oil; and Count Rumford, the multiple burner, an idea made feasible by these contrivances, and finally realized by Arago and Augustin Fresnel.”

Optical Properties of the Parabolic Reflector.— In the parabolic reflector all rays diverging strictly from the focus and falling on the paraboloid emerge in one beam of parallel rays. But as an oil light is not a mathematical point, but an object of considerable magnitude, the rays from the outside of the flame being exfocal will, after reflexion, emerge as a cone whose divergence is directly proportional to the radius of the flame and inversely to the focal distance of the reflector. Its intensity must consequently vary as the squares of the distances from the lighthouse. Optical apparatus does not then prevent that divergence which is due to the flame being of sensible magnitude.

Defects of the Paraboloid,.— It will be seen from fig. 7 that the parabolic mirror a is at best but a very imperfect instrument, for even if the radiant were strictly a mathematical point, the cone of rays (shown undotted) escaping past the lips of the mirror must be lost Mode of Employing Reflectors for Fixed and Revolving Lights.— In order to produce, on the catoptric system, a fixed light showing all round the circle, a number of reflectors (o, o, o, fig. 8) are fixed round the outside of a stationary chandelier n. As the ordinary paraboloid has about 14° of divergence, twenty-five reflectors were needed to light up continuously (though not equally) the whole horizon. If again the light was to revolve, then a revolving chandelier (figs. 9 and 10) was employed having a certain number of flat faces, on each of which was fixed a number of separate lamps and reflectors with their axes parallel to each other. When the chandelier revolved, and one of the flat sides was turned towards the sailor, he would, when at some, distance from the shore, receive a flash at once from each of the mirrors which were on that face, but when the face was turned away from him a dark period would intervene until the next face came round again.

Bordier Marcet’ s Fanal Sidéral, 1819.—In order strictly to equalize a fixed light over the whole horizon, which could not possibly be done with separate reflectors, Marcet proposed this ingenious instrument, which is generated by the revolution of the parabolic profile pp' (fig. 11) round its parameter as a vertical axis, instead of round a horizontal axis, as in all former reflectors. The vertices of the parabola are cut off, so as to permit of a common focus for thθ flame. The rays will therefore be reflected parallel to the horizontal axis in the vertical plane, while the natural divergence of the light in azimuth will not be interfered with. By this excellent contrivance the light was for the first time spread equally round the horizon in one continuous zone. But even though the radiants were reduced to a mathematical point, very many of the rays (shown in hard lines in the elevation, fig. 11) are allowed to escape past the lips of the reflector, and this loss takes place all round the circle.

Beginning in 1822, Augustin Fresnel, the eminent physicist and mathematician, entirely revolutionized the previously existing lighthouse system by means of his annular lenses, cylindric refractors, and totally reflecting prisms. Before describing these and their combinations it is necessary to state that the size of the flame produces divergence with lenses as well as with reflectors. The measure of this divergence for any point of the lens is the angle whose sine is

Radius of flame

Distance of point from centre of flame Fresnels Optical Agents. Annular Lens, 1748-1822.—Buffon in 1748 suggested a new form of lens for burning purposes in order to save the loss of heat by absorption of the sun’s rays in passing through a thick lens of large size whose outer profile is continuously spherical. He proposed to grind out of a solid piece of glass a lens in steps or concentric zones in order to reduce the thickness to a minimum (figs. 12 and 13). Condorcet, in his Éloge de Buffon, in 1773 (Paris edition, 1804, p. 35) proposed the capital improvement of building up Buffon’s stepped lens in separate rings, and pointed out that the cutting of the surface into steps had the effect of correcting to a large extent the spherical aberration, or divergence from the parallel, of the rays emitted by any continuously spherical lens. Sir D. Brewster, in

1811, also described the same plan. But both these writers designed their lenses for burning purposes only, and not for operating on light, while all the surfaces of their lenses were spherical. In 1822 Fresnel constructed a built-up lens for lighthouse purposes, in which the centres of curvature of the different rings receded from the axis according to their distance from the centre, so as practically to eliminate spherical aberration, the only spherical surface left being the small central part a (fig. 13). These lenses are used for revolving lights only.

Cylindric Refractor. —This instrument was introduced by Fresnel for effecting dioptrically by refraction in front of the flame what had been done before catop-trically by Marcet ’ s reflector by reflexion from behind the flame. It consisted of a zone or hoop of glass (figs.

14 and 15) generated by the revolution round a vertical axis of the middle section of the annular lens just described, which lens, on the other hand, being generated by the same profile round a horizontal axis, parallelized the rays in every plane, whereas the cylindric refractor does so in the vertical plane only.

Totally-Reflecting Prisms. —Fresnel next conceived the admirable improvement of employing the principle of “total” or internal reflexion by glass prisms. The ray Fi (fig. 16) falling on a prismoidal ring, ABC, is refracted and bent in the direction ι'R, and falling on the side AC, at an angle of incidence greater than the critical, is totally reflected in the direction Re, and, impinging on the side BC at e, it undergoes a second refraction, and emerges horizontally. The highest ray FA after refraction by AB and reflexion by AC must (in order to avoid superfluous glass) pass along AB, and after a second refraction at B emerge horizontally. The lowest ray FB after refraction by AB must, for like reason, pass along BC, and after reflexion by AC and a second refraction by BC also emerge horizontally.

Every other ray incident on the prisms between A and B is, after one reflexion and two refractions, emitted horizontally.

Straight Refracting Prism.— Fresnel’s straight refracting prism which refracts the rays that fall on it, but in one plane only, requires no further explanation, as it is simply a straight prism of the same horizontal cross section as one of the prisms of his cylindric refractor, so as when placed in front of his fixed apparatus to produce a beam of parallel rays like a lens.

Great Central Lamp.— We will now go on to describe the manner in which Fresnel utilized the four new optical agents which he originated, by first referring to his central burner system. In all lighthouses prior to 1822 the mode of getting up the required power was by employing a sufficient number of separate reflectors, each of which (unless we except Bordier Marcet’s mirror) required its own separate lamp. Instead of numerous independent lamps and reflectors, Fresnel used a single lamp which had four concentric wicks, and was fed with oil by a pump worked by clock work. Surrounding this burner was a stationary cylindric refractor for a fixed light, and annular lenses "evolving outside of it for a ’•evolving light.

Fresnels Combinations of his Optical Agents. Caiadioρtric Fixed Light.— This apparatus (figs. 17 and 18), in which a central burner is used, consists of a dioptric cylindric refractor with zones of silvered mirror above and below similar in profile to Bordier Marcet’s reflector. By the adoption of the refractor the whole of the wasteful divergence which occurs in Marcet’s reflector is prevented. We have here a geometrically perfect combination, but it is not so physically, because metallic reflexion is used. This physical defect Fresnel obviated in his next design.

Dioptric Fixed Light.—First Application of Total Reflexion to Fixed Lights.— In this apparatus Fresnel substituted his totally reflecting prism ρ and lens R for Marcet’s reflector, and thus distributed the whole light equally over the horizon by means of dioptric agents alone. This was the first application of total reflexion to lighthouse apparatus, and this beautiful instrument continues till now in universal use. Figs 19 and 20 represent an elevation and section of this apparatus.

Fresnel's Revolving Light. — In this form of revolving light (fig. 21) the central burner is surrounded by annular lenses L, and a compound arrangement of inclined trapezoidal lenses L' and plane silvered mirrors Μ. The inclined lenses fit closely to each other and form a pyramidal dome, and the light, intercepted by them, is sent upwards in inclined beams until, falling upon the plane mirrors M, it is sent outwards in horizontal parallelized beams. All these optical agents are made to revolve round the central lamp, and the sailor receives a full flash when the axis of one of the emerging beams passes his eye, and as soon as it passes him he is in darkness until the next beam comes round. This design, unlike that of his fixed light, is imperfect on account of the employment of metallic reflexion, and because two agents are employedfor all except the central portion of the rays.

Fixed Light varied by Flashes.— This distinction (figs. 22 and 23) Fresnel produced by placing his straight refracting prisms / on a revolving frame outside of his fixed light apparatus rr, so that when the upright prisms come in line with the observer the light is increased to the power of the revolving light, a broad flash as in the annular lens being produced in place of a narrow strip of rays as in the unassisted fixed light.

Alan Stevenson's Improvements.

Mr Stevenson was the first to introduce the dioptric system into Britain, and in doing so he made the following improvements.

1. Refractor of a Truly Cylindrical Form.— Owing to difficulties in construction, Fresnel adopted a polygonal instead of a cylindric form for his refractor, but Mr Stevenson succeeded in getting Messrs Cookson of Newcastle to construct a first order refractor of a truly cylindric form.

2. Helical Glass Joints for Fixed Lights.— Mr Stevenson further improved the apparatus by constructing the refractor in rhomboidal instead of rectangular pieces (figs. 24 and 25), thus producing helical joints and preventing serious obscuration of the light in any azimuth.

3. Helical Metallic Framings.— The internal metallic frame work for supporting the upper cupola of prisms was also, for the same reason, made by him of a helical form.

4. Diagonal Framed Lantern.— The astragals or sash bars of the lantern were likewise made diagonal and constructed of bronze instead of iron in order to reduce their sectional area. A.small harbour light with inclined astragals was made in 1836 by Mr E. Sang. Mr Stevenson also prepared a design in 1846 for Start Point, Orkney, in which he extended the helical arrangement to the astragals, but it was never carried out. Mr Douglass independently designed and afterwards carried into practice this form of lantern.

Alan Stevenson's Improved Revolving Light for Skerryvore.— In 1835 Mr Stevenson, in a report to the Northern Lighthouse Board, proposed to add fixed reflecting prisms p (fig. 26) below the lenses of Fresnel s revolving light, and he communicated this proposal to Μ. I*. Fresnel, who approved of his suggestion, and assisted in carrying out the design in 1843. This combination added, however, but little to the power of the flash, and produced both a periodically flashing and constantly fixed light; but it must be remembered that the prism for fixed lights was the only kind of reflecting prism then known. The combination of trapezoidal lenses and mirrors was also, for the same reason, still used. The prisms for Skerryvore were the first that were made of the large size (first order), and were constructed by Μ. Soleil at Paris under the superintendence of Leonor Fresnel, the successor of his brother Augustin, then deceased.

Besides the designs which have been described for improving Fresnel’s revolving light, there were many others, among which may be mentioned those of Leonor Fresnel, Sir David Brewster,W. H. Barlow, F.R.S., A. Gordon, Lepaute, Letourneau, Reyιιaud, and Tabouret; but in all these, as well as in those which have been described, the rays were either not parallelized in every plane or else unnecessary agents were resorted to for that purpose.

T. Stevenson's Holophotal System.

It was not till 1849-50 that the problem was first solved of condensing the whole sphere of rays diverging from a flame into a single beam of parallel rays without any unnecessary reflexions or refractions.

Catadioptric Holophote.— In figs. 27 and 28 part of the anterior hemisphere of rays is intercepted and at once υarallelized by the lens L, whose principal focus (i.e., for parallel rays) is in the centre of the flame, while the remainder is intercepted and made parallel by the paraboloid a, and thus the double agents in Fresnel’s design are dispensed with. The rays of the posterior hemisphere are reflected by the spherical mirror b back again through the focus, whence passing onwards one portion of them falls on the lens and the rest on the paraboloid, so as finally to emerge in union with and parallel to the front rays. This was the first instrument which intercepted and parallelized all the rays proceeding from a focal point by the minimum number of agents. It is therefore geometrically perfect, and was ca∏ed by Mr Stevenson a holophote, but it is not physically so, for it employs metallic reflexion, and with an ordinary oil flame and burner some of the rays reflected by the spherical mirror would fall upon the burner and be lost. This instrument was first employed at the North Harbour of Peterhead in 1849.

Holophotal Catadioptric Apparatus Revolving round a Central Flame.— If in place of Fresnel’s compound arrangement of trapezoidal lenses and plane mirrors there are substituted mirrors R, R (fig. 29) generated by the revolution of a parabolic profile round a horizontal axis, all the light will be at once sent out in parallel beams by them and the lenses L, and the apparatus is therefore geometrically perfect, but metallic instead of glass agents are still employed.

T. Stevenson's Holophotal Dioptric Agents.

The agents for rendering the holophote purely dioptric fall next to be described.

Holophotal Prisms, 1850.—If prisms are generated by the revolution of sections of Fresnel’s prisms round a horizontal instead of a vertical axis, as shown in fig. 30, rays incident upon them will be made parallel in every plane instead of in the vertical plane only as in Fresnel’s prisms J

In 1852 it was stated that A.

Fresnel, in 1826, tried prisms of this kind on lamp-posts for the quays of a canal in Paris, but their use was discontinued. They were not so arranged as to be applicable to lighthouses, and no account of them was published.

Double Reflecting Prisms giving Two Internal Reflexions instead of One, 1850.—The object of these prisms is to reflect the rays back through the focus. The rays from,/ (fig. 31) pass normally through the surface be and fall upon the surfaces ab, ac, by which they are totally reflected and sent back to the flame. Optical Combinations of Dioptric Holophotal Agents. Dioptric Holophote with Dioptric Spherical Mirror.— If rings of holophotal prisms p, ρ (fig. 32), combined with a central refracting lens L subtending together 180°, be placed before a flame, the whole of the front half of the diverging sphere of rays will be at once condensed by refraction and total reflexion into one beam of parallel rays. Next, if there be placed behind the flame a dome of glass a formed of zones generated by the revolution of the cross section of the double reflecting prisms round a horizontal axis, the back hemisphere of rays will, after the double reflexion, be returned by it through the flame so as to diverge along with the front rays; for this dome is a perfect mirror, not only for the faint light that is superficially reflected by the first concave surface, but also for the main portion of the rays which enter the substance of the glass and are there totally reflected. Thus the whole light is parallelized entirely by glass agents. This combination should therefore produce the light of maximum intensity, being both geometrically and physically perfect when the radius is a luminous point. Still, some of the rays reflected by the upper and lower parts of the mirror will not clear the burner of an oil light, but falling on it will be lost.

Dioptric Holophotal Revolving Bight.—Application of Total Reflexion to Revolving Lights, 1850.—If sections of the front half of the dioptric holophote be arranged as a polygonal cage round a large central flame and made to circulate round that flame, a revolving light with its periodic flashes will be produced (fig. 33), which is therefore geometrically and physically perfect, and the double agents used in Fresnel’s revolving light are thus wholly dispensed with. Mr J. T. Chance says, “on the whole the modern plan (holophotal) must give light five or six times more intense than that of the former (Fresnel’s) arrangement.” The greater simplicity and compactness of this light may be seen in comparing figs. 21 or 26 with fig. 33.

The holophotal revolving light is now the only one that is employed for all new lights. It was first introduced in 1850 on the small scale by Messrs Stevenson at Horsburgh Rock, Singapore, and on the larger scale at North Ronaldshay in Orkney, the prisms of which were most successfully made in 1851 by Letourneau of Paris.

Holophotal Fixed Light varied by Flashes.— Fresnel’s double agents (figs. 22, 23) are here also dispensed with by the single agency of panels of fixed light apparatus p', p' (fig. 34) and cylindrical refractors L', L', alternating with panels of holophotal apparatus p, p, L, L, both of which revolve together round the central burner.

Back Prisms.— It will be noticed that, while perfection was attained by single agency in Fresnel’s fixed apparatus and also in the holophotal revolving apparatus, a physical defect still attached to every form of holophote which compressed the light into a single beam where the rays require to be sent back through the flame. The defect is not a geometric one, and would be non-existent were the luminant a mathematical point. In order to remedy this objection, Mr T. Stevenson and Mr Brebner designed what they termed “back” prisms capable of deviating the rays through an angle of 130 , so that the engineer becomes virtually independent of the critical angle. Professor Swan of St Andrews also independently proposed the same kind of prism, accompanied by general formulae for its construction. The ray ab (fig. 35) is refracted at b, totally reflected at c, and again refracted at d, so as to pass out parallel to the horizontal axis. These prisms may be formed by the revolution of the generating section round either a vertical or horizontal axis, or they may be made straight as shown in elevation and section in the diagram. The “back prisms” which were first used in Islay, Argyleshire, were made by Messrs Chance in accordance with Professor Swan’s formula.

Perfect Form of Dioptric Holophote for an Oil Flame.— By combining the back prisms ga, he (fig. 36) just described with a semi-holophote abc subtending 180° and a portion of the dioptric spherical mirror ijk, no light is lost on the burner, and all the rays are parallelized, so that this apparatus, being all of glass, is both geometrically and physically perfect.

«7. T. Chance's Improvement of 1862 on Stevenson's Dioptric Spherical Mirror.— Mr Chance says, “The plan of generating zones round the vertical axis (instead of the horizontal) was introduced by the author, who adopted it in the first complete catadioptric mirror which was made and was shown in the exhibition of 1862 by the Commissioners of Northern Lighthouses, for whom it was constructed, in order to further the realizing of what Mr T. Stevenson had ingeniously suggested about twelve vears previously. During the progress of this instrument the idea occurred to the author of separating the zones (fig. 37) and also dividing them into segments like the ordinary reflecting zones of a dioptric light; by this means it became practicable to increase considerably the radius of the mirror and thereby to render it applicable to the largest sea light without overstepping the limits of the angular breadths of the zones, and yet without being compelled to resort to glass of high refractive power.”

Mr Stevenson states (JZin. Civ. Eng., vol. xxvi.) that he originally intended to generate the prisms round a vertical axis, but adopted the horizontal as more practicable where flint glass which was taken out of the melting pot on the end of a rod had to be employed for the most important and smaller central parts of the apparatus, which at that date were the only ones that could be constructed.

Professor Swan's Designs.— Among several ingenious arrangements and new forms of agents proposed by Professor Swan is the mode of sending rays from prisms through interstices left between other prisms placed in front, and also a form of agent which he termed the triesoptric prism, in which the rays would undergo two refractions and three reflexions. In fig. 38 a are the front and b the triesoptric prisms. The two upper and lower prisms a are constructed of flint glass of high refractive power. It will be observed from the drawing that this ingenious arrangement is nevertheless open to objection, for cones of light of 30° in front and of 65° at the back are lost through the interstices.

T. Stevenson's Azimuthal Condensing System for Distributing the Light Unequally in Different Directions cither Constantly or Periodically.

Previous to 1850 all apparatus was designed to emit light of equal power in every azimuth either constantly or periodically, so that when applied to illuminate seas extending farther in some directions than in others, such, for example, as long narrow sounds, the power was either needlessly great in some azimuths or not nearly strong enough in others. The only exception was where a tower was placed on a long straight coast, when a spherical mirror was fixed behind the flame to prevent the light passing landward. The light that would have been lost on the land was, in this way, sent seawards, but no attempt was ever made to condense the light over a given arc or to allocate this auxiliary light in proportion to the varying lengths of the different ranges and the amplitudes of the arcs over which it had to be seen, nor, where a light had to show all round the horizon, to weaken its intensity in one arc and with the rays so abstracted to strengthen some other arc which from its range being longer ought to be of greater power. As the optical agents which we have described were insufficient for this purpose, it became necessary to devise new agents possessing special optical properties for distributing the rays not equally, but equitably. Some of the more important of these agents and their combinations will be described, but before doing so it is necessary to make some preliminary explanations.

The form of the emergent light from condensing apparatus is neither a solid beam of parallel rays like that from an annular lens nor yet a zone of rays diverging naturally in azimuth all round like that from an ordinary fixed apparatus, but is intermediate between these, being a solid angle or wedge of light strengthened by those rays which would naturally diverge in other directions, but which are diverted and spread over the given sector as shown perspectively in plan in fig. 39, in which L represents the position of the lighthouse, LD'

the range, or distance at sea from which the light can be seen, and I)7WLD the solid horizontal angle that is to be illuminated and into which all the light DAD' which would naturally diverge over the rest of the circle must be compressed, and over which it must be uniformly spread.

Requirements for Fixed Condensing Lights.— 1. Where alight has to be seen constantly over only one arc of the horizon, the apparatus must compress all the rays within that one arc whatever its amplitude may be, and spread them uniformly over it

2. Where the whole horizon has to be constantly illuminated, but the light has to be seen at greater distances over some parts of the sea than over others, as much light must be abstracted as can be spared from the shorter ranges and diverted to the longer so as to allocate the rays in the compound ratio of the number of degrees and the squares of the distances from which the light has to be seen in each arc; and the light thus diverted from one arc to strengthen another must be spread uniformly over the one that has to be strengthened. By this mode of abstraction and addition there is produced a constant equitable distribution over the whole horizon of all the rays from any single flame. Where the whole horizon has not to be lighted, and where more than one arc has to be strengthened, the rays which would be lost on the land or can be spared in some other direction must be diverted to and spread uniformly over these arcs in proportion to their amplitudes and ranges.

Requirements for Revolving Condensing Lights.— 1. Where a light has to give its flashes periodically over only one arc of the horizon, all the rays must be collected and sent out periodically in solid beams of equal power over that arc.

2. Where the whole horizon has to be periodically illuminated, but where its flashes have to be seen at greater distances over some parts of the sea than over others, the apparatus must be made (as in number 2 of the fixed condensing light) to vary proportionally the power of the flashes whenever they begin to pass over those parts of the sea where the ranges are of different lengths so as to produce an equitable periodic distribution of all the rays.

3. "Where the whole horizon has not to be lighted, and where more than one arc has to be strengthened, the rays which would be lost on the land or can be spared in some other direction must be αiverted to and spread uniformly over these arcs so as to strengthen the flashes in proportion to their amplitudes and ranges.

If n be the number of degrees in an arc to be illuminated, and d the distance in miles to be traversed by the light, then, neglecting atmospheric absorption, the quantity of light to be allocated to that arc will be proportional to nd? ; but if we take account of atmospheric absorption, supposing q to be the quantity out of a unit of transmitted light which escapes absorption after passing through a mile of air, then the whole light needed by the arc to be illuminated will be proportional to m = ncPq d . Supposing now that Lis the whole 360’ of available light from the apparatus, the quantity to be apportioned to the given arc will be mlf½m, where Zm denotes the sum of the several numbers m computed for the respective arcs of the horizon.

The Principal Optical Condensing Agents. Condensing Straight Prisms.— These, either by reflexion or refraction or both, cause a ray fr (fig. 40) proceeding in any compass bearing from a fixed light apparatus AA to emerge in the direction,- c.g., parallel to the corresponding ray fb, which proceeds in the same compass bearing from another part of the apparatus and so of any other ray fc which is bent parallel to the ray fa. Right-Angled Expanding Prisms. — These (fig. 41) are right-angled vertically, while in piano two of them C 1 C 2 are semi-rings, and the third C 3 a semi-cone. A semi-cylindric beam of parallel rays passing vertically upwards and falling normally on the bases a of the prisms enter the glass, are reflected by the sides b, and pass out horizontally and normally to the other sides; but, as the prisms are bent through a circular segment d'ad in piano, the emergent rays will be spread over the same angle in azimuth, and this will be true of any angle in azimuth subtended by the prisms. As those in the diagram subtend 180°, the light will in this case be spread over half the horizon d'ad. Twin Prisms.— These are for carrying out Professor Swan’s mode of causing light coming from prisms behind to pass through spaces left for the purpose between others in front. The twin prisms (fig. 42) are formed by cutting out the apex (shown black on diagram) of a straight prism so as to leave a sufficient space between it and the next prism for the passage of rays coming from behind. The length of glass traversed by the rays is lessened by this arrangement, and the size of the apparatus and lantern can at the same time be very greatly reduced.

Differential Lens.— Horizontal divergence may be obtained to any required amount by varying the radius of curvature of the inner face of an annular lens. The outer face (fig. 43) is the same as that of an ordinary annular lens, while the other face (fig. 44), though straight in the vertical, is ground to the required curve in the horizontal plane. The rays ffc (fig. 45) falling upon the lens x converge to the vertical focal plane ff' and afterwards diverge through the smaller horizontal angleaf/*«/, and so for any other case.

Differential Refractor.— This is the application to the cylindric refractor of the same principle which has been described above for the ordinary annular lens.

Condensing Catoptric Spherical Mirror.— If a large arc of a spherical mirror be cut opposite to a short range of sea, and a smaller arc be cut in the mirror opposite to the longer range which has to be strengthened, then the light passing through the larger cut is received on an elliptic reflector placed behind, so as, with the help of an additional reflector and lens, to cause the larger cone of rays to be compressed into the smaller arc which has to be strengthened.

Spherical Mirror of Unequal Area.— This mirror is cut down in height in such a way that its different heights represent inversely the different distances of the neighbouring land from the lighthouse, so as to reflect less light in the shorter and more in the longer ranges.

Application to Fixed Condensing Apparatus. Fixed Condensing Light for a Single Sector, 1850.—The holophote pLp (fig. 46) throws its whole light on straight condensing prisms c, each of which distributes the rays over the required sector.

Condensing Apparatus for Steamers' SideLights.— Bymeans of this application of the condensing principle (figs. 47 and 48) all the light can be dis-< tributed with strict equality < over 112’ 30', which is the arc prescribed for steamers by the Board of Trade. Several of the Transatlantic and other steamers have adopted this kind of apparatus, which is hung on gimbals and placed in iron towers, having an entry from below the deck, which can be made use of in bad weather.

Condensing Quadrank— The fixed apparatus bbb, with spherical mirror behind, throws its rays directly through the angle of 90° afa', while the supplementary rays falling on the straight condensing prisms p, p, p, p , (fig. 49) are sent out parallel to the corresponding rays in the unobstructed central quadrant of the main apparatus. The whole light will therefore be condensed equally over 90°.

Condensing Octant.— The central fixed apparatus bb (figs. 50, 51) with spherical mirror dd throws its rays directly over the angle of 45° pgp, while the supplemental rays fall upon the straight condens ing prisms p, each of which spreads the incident rays parallel to the corresponding rays in the central angle ρgp. In this way the whole of the front hemisphere of rays is parallelized in the vertical plane and spread equally over the 45° in azimuth. The hemisphere of back rays is condensed into the same arc seawards by the dioptric spherical mirror, which returns the rays incident upon it through the flame, where mingling with the front rays they are finally distributed over the arc of 45°. The light passing above the spherical mirror is parallelized by a half holophote ih (fig. 50), and sent vertically upwards, when, falling on the expanding prisms gfe, it also is finally spread over the required arc. Thus the whole light is condensed and distributed with strict equality over the 45° by means of six different optical agents. Two apparatus of this kind form the leading lights to the River Tay.

Condensing TwinBrism Light.— Fig. 52 shows part of the Lamlash light in the Firth of Clyde. Its action will be easily understood by the numbers shown on the diagram.

Fixed Condensing Lights for more than one Sector of Unequal Range.— Fig. 53 is a chart of the Sound of Glenelg, in which the Isle Oronsay light has to be seen generally at a distance of 3 or 4 miles across the Sound, but in the direction AB down the sound an angle of 10° has to be seen about 15 miles and another up the sound between C and D an angle of 10° has to be seen about 7 miles. Fig. 54 shows in plan the apparatus, which was for certain local reasons not constructed according to the formula, but 193° of the spare light which would otherwise have fallen upon the land were allocated down Sleat Sound and 139° down the Sound of Glenelg. That for Sleat Sound was given directly by the main apparatus through the arc δ, and that to Glenelg through the arc β, but these arcs were respectively strengthened by portions of the holophotes B and C, the light from which was condensed respectively over the arc δ by each of the straight prisms a and over the arc β by each of the prisms b. The condensed light at Isle Oronsay, which is of the fourth order, appears to be equal to that of a first order, so that with a lamp consuming annually only 170 gallons of oil a light is obtained in the only direction in which great power is required equal in effect to a lamp consuming 800 gallons per annum.

Fixed Condensing Lights of Unequal Range which constantly Illuminate the whole Horizon.— For this purpose the condensing spherical mirror or spherical mirror of unequal area will be found applicable.

Revolving or Ί ntermittent Lights for Condensing the Rays into one Sector.— The holophote in fig. 55 throws its rays on straight condensing prisms P, each of which spreads the light over the prescribed sectors, while masks M turning horizontally on pivots cut off the light either slowly or suddenly so as to produce a revolving or an intermittent light, both of which condense all the rays uniformly over the one sector.

Application of Condensing Principle to Revolving Lights of Unequal Range, which do not Illuminate the whole Horizon. Repeating Light.— Plane mirrors M (fig 56) revolve on an endless chain placed outside of the apparatus and alter the direction of the flashes after they pass into the dark arc on the landward side so as to cause the lenses L, L to repeat their flashes over the seaward arc which requires strengthening. The condensing spherical mirror and mirror of unequal areas will also be found applicable in cases where the flashes do not require to sweep over the whole horizon.

Condensing Revolving Lights τohich Periodically Illuminate the whole Horizon, but which Vary the Strength of the Flashes in Passing over Certain Sectors.— The spherical mirror of unequal areas and the condensing mirror are equally well suited for those that revolve.

Condensing Intermittent Lights.— Figs. 57 and 58 show straight refracting or reflecting prisms, which revolve and intercept as they pass round certain of the rays from a central fixed light apparatus so as to produce perfect darkness over the sectors which they subtend at the time, while they spread the rays which they intercept uniformly over and thus strengthen the intermediate sectors which are illuminated directly by the central apparatus. The peculiar property of this arrangement is that the power is increased in proportion to the duration of the intervening periods of darkness. Thus, neglecting the loss by absorption, &c., the power is doubled when the periods of light and darkness are equal, trebled when the dark periods are twice as long as the light, and so on in proportion, while in every case the rays are spread uniformly over each illuminated sector.

Intermittent Condensing Light with Differential Refractor.— Figs. 59 and 60 show the new apparatus of Mull of Galloway in which ABA is the differential refractor, by the compound horizontal and vertical action of which single agent the whole condensing intermittent eflect is produced, so that condensing prisms are done away with opposite the central, which is the most important, part of the apparatus. The centre of the inner curve of the refractor is at O in fig. 60. Though there is no relative motion in this apparatus, every part of which revolves together, the parts may be arranged so that the condensers only move. This apparatus was constructed in the most satisfactory manner by Messrs Barbier & Fenes-tre, Paris.

Alteration of Fixed to Intermittent Apparatus. —Any existing fixed light can at once be made intermittent so as to show either equal or unequal periods by simply causing condensing prisms to circulate round it, while the power will be increased in proportion to the ratio of the duration of light to dark periods.

Beacons and Buoys.

Beacons in exposed situations are constructed sometimes ol stone, and cement-concrete or cement-rubble, but generally of castiron columns let into heavy base plates which are fixed to the rock by strong lewis bats. The small class iron beacons are generally of malleable iron and the larger of cast-iron, but steel or bronze might with advantage be used in very exposed places. Fig. 61 shows a first order castiron beacon as used in Scotland.

T. Stevenson's Apparent Light.— This kind of light is specially useful at places where there is a sunk rock with little sea-room round it, and at the pierheads of harbours which must be closely hugged by vessels seeking entrance. It consists of certain forms of optical apparatus for reflecting and redistributing at a beacon placed on a submerged rock parallel rays which proceed from a lamp and apparatus placed on the laud. An optical deception is thus produced, as the sailor naturally supposes that there is a lamp burning on the beacon itself. The first light of this kind, which has been in use since 1851, was placed in a beacon on a sunk reef in Stornoway Bay, and is shown pictorially in fig. 62. This light is 530 feet distant from the lighthouse where the lamp is placed. Others have since been established at Grangemouth, which is 535 feet from the light, and at lesser distances at the harbours of Ayr and Arbroath, at Odessa in the Black Sea, and at Gat Combe" Head, Queensland.

Beacon Lights.— Lamps without glass chimneys, as usedin the early experiments with paraffin, and as used with gas in Pintsch’s buoy, having iron tubes placed at certain distances above the flam.?, and supplied with very large cisterns of crystal oil, have been kept continuously burning in Scotland for about a month without trimming. These lamps are for rocks at sea which can only be reached when the weather is moderate.

Illumination of Beacons by Gas to Produce a Fixed Light.—A light near Port-Glasgow has been illuminated by gas since 1861. The tower is about 300 feet from the shore, and the supply and pressure of gas are regulated by self-acting arrangements on the shore.

Pintsch's Gas Illuminated Buoy for Producing a Fixed Light.— These buoys are gas and water tight, and are charged to a pressure of ten atmospheres. The buoy is filled with gas sufficient to bum night and day for three or four months; the burning pressure is kept constant by Pintsch’s regulator. These buoys promise to be of great service, and have now been in use for more than a year at several places both in England and Scotland.

Automatic Meter for Producing Intermittent Lights by the Flow of the Gas.— In order to distinguish one buoy or beacon from another, Mr T. Stevenson has proposed to make the flow of the gas itself produce automatic intermittent action without using clockwork. The full flame continues to burn until the action of the meter reduces the supply, but leaves a small jet still burning in the sockets of the burner till the full flame is again produced. This meter has been tried successfully on shore for twenty-eight hours consecutively.

Uniform System of Beacons and Buoys.— The first proposal for dealing systematically with the colours of beacons and buoys was in 1828, when Mr R. Stevenson proposed for the navigation of the river Forth that red buoys should be taken on the starboard and black on the port side in going up the river. This system has since been adopted generally in the United Kingdom. In 1857 Admiral E. J. Bedford suggested a uniform system for all the coasts of the country, which was carried out in Scotland by Mr A. Cuningham. A different uniform system has been adopted in England by the Trinity House. In 1859 Mr J. F. Campbell suggested what is undoubtedly the best system, that a buoy shall indicate by its shape and colour the compass direction in which a vessel should sail after seeing the buoy.

Courtenay’s Automatic Buoy.— This useful form of buoy (fig. 63) first used in America, has a cylindric tube A projecting below the level of the water, on the upper part of which cylinder there is a whistle N and also two inlet valves F, F, so that the rising and falling of the buoy draws in and expels the air alternately which sounds the whistle. It is said that an undulation of even 12 inches is sufficient to sound the whistle.

Tolling of Bells by the Hydrostatic Pressure of the Tide.— In 1810 Mr R. Stevenson designed apparatus for tolling a bell on the Carr Rock beacon, which was to have been effected by a float when ascending with the pressure of the flowing tide, when descending with the fall of the ebb tide and by means of a peculiar arrangement of weights for keeping the bell in action during the time of slack water.

Illuminants.

After Argand’s ordinary single-wick oil burner, Fresnel still adheringto Argand’s principle of the double air current introduced two, three, andfour concentric wick burners, Mr AlanStevenson the five, and Mr Douglass the six wick burner. These burners are suitable for the consumption of animal and vegetable oils, which were in use in the various lighthouse services of the world till within the last few years. These oils are, however, supersede, I by hydrocarbons, which at one half the cost give an increased luminous intensity for an equal consumption, but they require a larger supply of air for their combustion, and consequently an improved form of burner. Μ.

Maris in 1856 devised a single-wick burner for this purpose which gave good results, but many and all attempts to construct a multiple-wick mineral oil burner failed till 1868, when Captain Doty’s form was devised. By a happy choice of proportions in the various parts of his burner (fig. 64), and by the addition of an exterior cylinder surrounding the outer wick, and a central disk, both placed in such a manner as to throw a current of air into the flame at the right place, Captain Doty succeeded in producing single and multiple-wick hydrocarbon burners, which carry a flame of great luminous intensity and regularity. The following table gives the details of these burners, and also the candle power and consumption, as determined by Dr Stevenson Macadam :— [table]

Mr J. N. Douglass has also introduced a mineral oil burner, which, however, Doty claims as similar in principle to his. By igniting only half the number of wicks (the outer in a six-wick for instance) the power of the burner can be diminished by one half. This is Mr Douglass’s lamp of single and double power for use in clear and foggy -weathers respectively.

The mineral oil employed most extensively in lighthouse illumination is Scotch paraffin. The specific gravity, which is a test of the relative richness of the oil, should be from 0 - 8 to 0 , 82 at 60° Fahr., and the flashing point or temperature at ιvhich it begins to evolve inflammable vapour should not be lower than 125° nor higher than 135° Fahr. The Board of Northern Lighthouses was the first lighthouse authority in Britain that adopted paraffin and Doty burners, and by doing so an annual saving of between £4500 and £5000 in the maintenance of the lights on the Scotch coast was effected.

Single-wick burners are supplied with oil from the cistern by the capillary action of the wick alone. But in the case of multiplewick burners other methods must be employed to secure a sufficient supply. If the cistern be placed above the level of the top of the burner, the flow of the oil to the wick cases is effected by the direct action of gravity, regulated by a contrivance which maintains a constant head. If, however, the cistern be placed below this level, either a mechanical lamp is employed, in which the oil is forced into the burner by pumps worked by clockwork, or a moderator lamp, in which this is effected by the pressure exerted by a weighted piston descending in a cylinder forming the cistern.

Coal-Gas.— Coal-gas was first used as a lighthouse illuminant at Salvore, near Trieste, in 1817. For many years it has been used in the harbour lights of Great Britain when in the neighbourhood of gas-works. Mr J. R. Wigham has designed a compound or crocus burner consisting of a group of twenty-eight vertical tubes, each carrying an ordinary double fish-tail burner, and the ignited gases issuing from all these jets unite into one large flame. Additional groups of twenty jets each can readily be arranged around the first, which forms a central nucleus, and in this way, depending on the state of the atmosphere, the power of the burner can be made at will 28, 48, 68, 88, or 108 jets. Fig. 65 shows the arrangement for 28 jets, and fig. 66 one for 108 jets. In his triform or quadriform systems Mr Wigham places two, three, or four of the burners already described vertically one above the other (fig. 67), with lenses opposite to each. The following table gives the candle powers, &c., of Mr Wigham’s burners. [table]

Faraday and Holmes's Magneto-Electric Light.— In 1853 Professor Holmes made the first magneto-electric machine for lighthouses, which was tried by the Trinity House in 1857, and the electric light was first shown to the mariner in 1858. The Trinity House subsequently introduced it at Dungeness in 1862 and at Souter Point in 1871. The optical apparatus for these lights was designed by Mr J. T. Chance. In 1869 Holmes constructed for the Trinity House a dynamo-electric machine (in which no permanent magnet is used), giving a light of about 2800 candles. The magneto-electric light of Holmes as exhibited from a third order dioptric apparatus at the South Foreland light has been estimated at 152,000 candles, or twenty times that of the old first order dioptric fixed white light. The following table shows the results which have been obtained by Dr Tyndall and Mr Douglass by the magneto-electric and dynamoelectric machines, tried at the South Foreland, 1876-77: [table]

It must be kept in view that in the electric as well as in every other light the following requirements must be fulfilled:—(1) it shall be constantly in sight during those periods of time at which it is advertised to the mariner as being visible; (2) it shall be seen as far as possible in a thick and hazy atmosphere; (3) it shall constantly maintain the distinctive character of the station where it is employed so as not to be mistaken for another light; and (4), when revolving, its flashes shall remain long enough in view to let the sailor take the compass bearing of the light. The electric light practically fulfils all these conditions, and, when we consider its transcendent intensity, and the smallness of the luminous radiant which enables the engineer to adapt it to any required conditions far more strictly than oil light, we may certainly conclude that the electric is the best though the most costly of all illuminants.

The only question which has not yet been decided is whether an electric light of equal initial power will penetrate a hazy atmosphere as far as an oil light. Experiments made at Edinburgh in 1866 seemed to show that the highly distinctive flash of the electric light when acted on by optical apparatus is perhaps not so much due to a greater amount of light as to the more complete parallelism of the rays arising from the smallness of the radiant. The apparatus of a small size which was first used both in England and France necessarily produced a wasteful vertical divergence, and has therefore been now justly discarded.

Allard's Statistics of Lighthouse Apparatus.

The following useful formulae are taken from Μ. Allard’s very valuable Mémoir. sur l’intensité et la Portée des Phares, Paris, 1876.

Consumption of Oil in Relation to Diameter of Burner.— If c denote the consumption of mineral oil in grammes per hour, d the diameter of the burner in centimetres, then c=4∙9d 2 ∙ 22 .

Luminous Intensities.— A Carcel burner consuming 40 grammes of colza oil per hour being taken as unity, if I denote the intensity for mineral oil in a burner of diameter d in centimetres, then

I=0∙22d s ∙ 1 .

Luminous Intensities of Apparatus.—Loss due to Reflexion, Absorption, and Framing of Apparatus.— The loss due to surface reflexion on entering and leaving the glass may be valued at Ό50, •052, - 058, ‘075, ‘120, ’ 230, for angles of incidence respectively of 0°, 15°, 30°, 45°, 60°, 75°. In totally reflecting prisms the luminous ray suffers three deviations instead of two, therefore the above values should be multiplied by ⅜.

The loss by absorption in the glass, although properly given by an exponential formula, may, with sufficient accuracy, be taken as - 03 per centimetre of glass traversed.

The loss due to the horizontal joints of the lenses, and to the intervals between the reflecting rings, varies from Ό2 to Ό3, or from Ό1 to '04, in passing from the first to the fifth order.

Coefficients. — These are the ratios in which the intensity of the lamp is increased by the apparatus. Where m is the coefficient,/the focal distance, d the diameter, and h the height of flame in a fixed light apparatus, they can, if expressed in centimetres, be calculated from the formula—

-W“—*∙*<iΓ∙

Annular Lens.— The intensity of an annular lens is obtained by multiplying that of the corresponding fixed light by ¾ i where φ is α the angle subtended by the annular lens, and α the horizontal semi divergence. The intensity of light from an ordinary fixed light apparatus is increased 38 per cent, bv the use of the dioptric spherical mirror.

Distinctions of Lights.

The most important characteristic distinctions of lights are the following—(1) The∕aκd light. Two of these are sometimes shown at different heights from the same tower. (2) The revolving light, which at equal and comparatively long periods comes slowly and gradually into full power, and then as gradually disappears. (3) The flashing light, which at short periods (such as a few seconds) comes very quickly, though gradually, into full power, and as quickly and gradually fades away, was first introduced by the late Mr Robert Stevenson in 1825. (4) The coloured light, which is obtained by using coloured media, and was first employed by Mills of Bridlington. (5) The intermittent light, introduced by the late Mr R. Stevenson in 1830, which bursts instantaneously into full power, and after remaining as a fixed light for a certain length of time is as suddenly eclipsed and succeeded by a dark period. When the periods are very short the intermittent is now called an occulting light (6) The intermittent light of unequal periods, first proposed by Mr R. L. Stevenson in 1871, showing from the same apparatus different durations of the dark periods with equal periods of light, such as fixed for 2" eclipsed 5 , fixed 2" eclipsed for 2", and then, as at first, fixed for 2". (7) Group flashing light. This distinction consists in giving a series of groups of two or more flashes separated by short eclipses, the groups being separated from one another by a larger eclipse. Mr Wigham of Dublin first proposed this distinction, and applied it to a dioptric revolving apparatus illuminated by gas, the effect being produced by raising and lowering the gas as Mr Wilson of Troon had done in 1827. Dr Hopkinson produced the same effect optically and for any illuminant by altering the arrangement of the ordinary revolving apparatus. (8) Fixed lights illuminating the whole horizon, but shoιcing revolving or intermittent characteristics over small arcs. These can be produced by masks arranged like Venetian house blinds placed opposite the arc where the characteristic is to be different from the main light. This arrangement and No. 9 were lately proposed by Mr T. Stevenson. (9) Intermittent flashing light. This is produced by condensing prisms revolving round a fixed light apparatus. The flashing light continues giving its rapidly recurring flashes for a certain period, which is succeeded by a dark period. The light which is not required during the dark periods is condensed by the prisms so as to strengthen the light periods.

Uniform System of Characteristics.— Various modes of distinguishing lights have been at different times proposed, such as Babbage’s, which represented the stations by numbers, and Sir W. Thomson’s by the Morse alphabet, or dots and dashes; but the following results seem to show that the present system is satisfactory, while it is free from many objections to which the other modes are liable. The Board of Trade records of shipwrecks show that during the space of twenty years (1857 to 1876) 2716 vessels were wrecked by stranding on the coasts and islands of Scotland and the Isle of Man, and that of all these there were only twenty-six, and the great majority of them were small vessels, in which it was even alleged that the cause was a failure in identifying the distinctive characteristic of some one of the Northern lighthouses, and in one-third of these cases the sailors were foreigners, and made such glaring mistakes as taking a fixed for a revolving light. It is surely more probable that in many of these cases they did not confound the characteristics at all, but really did not know what they were. But, waiving all these considerations, there were after all only about 1 per cent, of all the wrecks which were alleged to have been caused by mistaking one light for another.

Power of Coloured Lights.— Experiments, and particularly those of Μ. Reynaud, have shown that red of the same initial power as white penetrates a fog to a greater distance than the white, owing to the red raysnot being so much dispersed byrefraction. Experiments made at Edinburgh in 1868 show that a light should, before passing through red glass, have four and a half times the initial power of a white light. The same results have since been obtained by Μ. Allard of Paris.

An objection to all coloured lights is the fact that some persons are unable to detect any difference between certain colours. According to the experiments of Professor Holmgren on from sixty to seventy thousand persons in Europe and America, about 4‘2 per cent, are congenitally colour-blind, and the commonest form of this defect is inability to distinguish between red and green, which are the only tints employed for lighthouse distinctions. This objection, however, must gradually lose its weight, as the Board of Trade has now adapted the “colours test” as part of their examinations for certificates in navigation.

Distribution of Lights on a Coast. —Mr Alan Stevenson gives the following as a few of the chief considerations which should guide us in the selection of sites and characteristic appearance of the lighthouses to be placed on a line of coast.

“1. The most prominent points of a line of coast, or those first made on oversea voyages, should be first lighted; and the most powerful lights should be adapted to them, so that they may be discovered by the mariner as long as possible before his reaching land.

“2. So far as is consistent with a due attention to distinction, revolving lights of some description, which are necessarily more powerful than fixed lights, should be employed at the outposts on a line of coast.

“3. Lights of precisely identical character and appearance should not, if possible, occur within a less distance than 100 miles of each other on the same line of coast which is made by over-sea vessels.

“4. In all cases the distinction of colour should never be adopted except from absolute necessity.

“5. Fixed lights and others of less power may be more readily adopted in narrow seas, because the range of the lights in such situations is generally less than that of open sea-lights.

“6. In narrow seas also the distance between lights of the same appearance may often be safely reduced within much lower limits than is desirable for the greater sea-lights. Thus there are many instances in which the distance separating lights of the same character need not exceed 50 miles; and peculiar cases occur in which even a much less separation between similar lights may be sufficient.

“7. Lights intended to guard vessels from reefs, shoals, or other dangers should, in every case where it is practicable, be placed seaward of the danger itself, as it is desirable that seamen be enabled to make the lights with confidence.

“8. Views of economy in the first cost of a lighthouse should never be permitted to interfere with placing it in the best possible position; and, when funds are deficient, it will generally be found that the wise course is to delay the work until a sum shall have been obtained sufficient for the erection of the lighthouse on the best site.

“9. The elevation of the lantern above the sea should not, if possible, for sea-lights, exceed 200 feet; and about 150 feet is sufficient, under almost any circumstances, to give the range which is required. Lights placed on high headlands are subject to be frequently wrapped in fog, and are often thereby rendered useless, at times when lights on a lower level might be perfectly efficient. But this rule must not, and indeed cannot, be strictly followed, especially on the British coast, where there are so many projecting cliffs, which, while they subject the lights placed on them to occasional obscuration by fog, would also entirely and permanently hide from view lights placed on the lower land adjoining them.

‘ ‘ 10. The best position for a sea-light ought rarely to be neglected for the sake of the more immediate benefit of some neighbouring port, however important or influential; and the interests of navigation, as well as the true welfare of the port itself, will generally be much better served by placing the sea-light where it ought to be, and adding, on a smaller scale, such subsidiary lights as the channel leading to the entrance of the port may require.

“11. It may be held as a general maxim that the fewer lights that can be employed in the illumination of a coast the better, not only on the score of economy, but also of real efficiency. Every light needlessly erected may, in certain circumstances, become a source of confusion to the mariner; and in the event of another light being required in the neighbourhood, it becomes a deduction from the means of distinguishing it from the lights which existed previous to its establishment.

“12. Distinctions of lights, founded upon the minute estimation of intervals of time between flashes, and especially on the measurement of the duration of light and dark periods, are less satisfactory to the great majority of coasting seamen, and are more liable to derangement by atmospheric changes, than those distinctions which are founded on what may more properly be called the characteristic appearance of the lights, in which the times for the recurrence of certain appearances differ so widely from each other as not to require for their detection any very minute observation in a stormy night. Thus, for example, flashing lights of five seconds’ interval, and revolving lights of half a minute, one minute, and two minutes, are much more characteristic than those which are distinguished from each other by intervals varying according to a slower series of 5", 10", 20", 40", &c.

“13. Harbour and local lights, which have a circumscribed range, should generally be fixed instead of revolving; and may often, for the same reason, be safely distinguished by coloured media. In many cases also, where they are to serve as guides into a narrow channel, the leading lights which are used should, at the same time, be so arranged as to serve for a distinction from any neighbouring lights.”

Table of Distances at which Objects can be seen at Sea, according to their respective elevations, and the elevation of the eye of the observer. [table]

Example.— A tower 200 feet high will be visible 20'66 nautical miles to an observer whose eye is elevated 15 feet above the water; thus, from the table :—

15 feet elevation, distance visible 4 ’ 44 nautical miles.

200 ,, ,, ,, 16∙22 „ „

20’66 Floating Lights.

Prior to 1807 the only kind of floating light was a shin with small lanterns suspended from the yard-arms or frames. The late Mr Robert Stevenson introduced in that year a lantern which surrounded the mast of the vessel, and was capable of being lowered down to the deck to be trimmed. By his plan a lantern of much greater size could be used, and with this increase of size a larger and more perfect apparatus could be employed which admitted of gearing for working a revolving light, as adopted by the Trinity House. In fig. 68, aa is the mast, b tackle hook, c, c brass flanges for fixing parts of lantern together, e and g weather guards, h plate glass front of lantern, i shutter by which lamps are trimmed, k lamps, l silver reflector. Revolving catoptric apparatus was applied to floating lights in England, and Μ. Letourneau, in 1851, proposed to employ a number of sets of dioptric apparatus in one lantern; figs- 69 and 70 show elevation and plan of one of the Hooghly floating lights on the dioptric principle, designed for the Indian Government by Messrs Stevenson. It will be observed that not fewer than four of the separate lights are always in view from every part of the compass at once. The “Seven Stones’’lightship, which has all the recent Trinity House improvements, is moored off the Lands End in 42 fathoms. Her moorings consist of a 40 cwt. mushroom anchor, and 315 fathoms of lj inch studded chain cable. The vessel is timber-built, copper-fastened throughout, sheathed with Muntz metal. Her length is 103 feet between perpendiculars; extreme breadth, 21 feet 3 inches; depth of hold from the strake next the timbers to the upper side of the upper deck beams, 10 feet 3 inches. In the event of the vessel breaking adrift, she is provided with sails, the mizzen being frequently used for steadying the vessel at her moorings. The cost of the vessel, fully equipped for sea, with illuminating and fog signal apparatus complete, was about £9500.

Lanterns are generally constructed of diagonally framed astragals (fig. 71) to avoid the obscuration of light in any one azimuth as would be the case with dioptric fixed light apparatus were vertical astragals used. The astragals are made of gun-metal, having a tensile strain of 33,000 lb to the square inch; the dome is made of copper plates rivetted together; and the glass is the best plate ∣ inch thick. Storm panes are kept at hand which can be applied in a few minutes in the event of a pane being broken, which sometimes occurs from large birds driving against the lantern and by stones thrown up from the face of cliffs by sea and wind during storms. Mr Douglass’s cylindric or helical lanterns (fig. 72) have steel astragals, and the panes, which are also cylindrical, are ⅜ inch thick.

Lightning Conductors.— Each lighthouse tower is furnished with a copper lightning conductor consisting of a I inch copper rod or wire rope. These conductors terminate about 18 inches above the top of dome of lantern in a forked shape with two platinum points; the lower end bifurcated and furnished with large copper earth-plates is sunk into moist earth or into the sea. The conductor is put into metallic connexion with all large masses of metal outside or inside the lightroom. See Lightning Conductors. Machinery.— Revolving lights are propelled by clock-work machinery actuated by a falling weight. The machines are provided with a maintaining power which keeps the apparatus moving while the machine is being wound up, and a bell is attached which rings to warn the keepers when the machine requires winding up. A clock face is also attached which keeps time with the lightroom clock when the apparatus is revolving at the proper speed.

Masking Screens.— For the purpose of cutting off sharply the light proceeding from any fixed light apparatus, the late Mr J. Μ. Balfour devised what he called radial masking screens. These screens, which are in use at several lighthouses where the fairway for ships is narrow, consist of a senes of thin plates placed with their edges towards the apparatus, and with such a distance between them, and of such a length as will secure that only rays having the desired divergence can pass through between them.

With revolving lights it is not possible to effect a “cut off” as in the case of fixed lights, on any particular bearing; for the direction of the axis of the luminous beam projected by the lens or reflector is being continually changed in the horizontal plane by the revolution of the frame on which it is fixed. Mr Stevenson’s mode of overcoming this difficulty is by fixing a light metallic screen or coloured glass shade outside of the revolving apparatus and on the safety side of the danger arc. These screens are fitted with rollers for nιnning on a slightly inclined plane or circular path. Small projecting rods or snugs fixed to the sides of each lens are, in revolving, brought against the edge of the screen, and gradually press the screen before it up the inclined plane. By the time the lens reaches the edge of the danger arc the screen has been pushed to the top of the inclined plane, and the full beam of light, coming from the now entirely uncovered lens, points in the required line of cut off, which is the border of the danger arc. But whenever the further revolution of the apparatus causes the snugs to pass clear of the edges of the screen so as to free it from their pressure it immediately runs back again to its original position in front of the lens, so as to prevent any of the light being now sent seawards. By this continued reciprocative movement of the screen, as lens after lens comes round, the same effect will be successively produced and the light will always be cut off on the lines of obscuration, so that the flashes can never be seen within the danger arc.

Mathematical investigations and formulae for constructing the optical instruments for lighthouses will be found in the books undemoted by Alan Stevenson, Swan, Chance, Nehls, Beynaud, Allard, &c., to which the reader is referred.

Literature. —Smeaton, Narrative of the Eddystone Lighthouse, London, 1793; Stevenson, Account of the Bea Rock Lighthouse, Edinburgh, 1824; Fresnel, Mémoire sur un nouveau système d^ éclairage des phares, Paris, 1822; Brewster, Treatise on Burning Instruments. 1812; Fanale di Salvore, nell' Istrta, Illuminato a Gas, Vienna, 1821; Account of a Nein System of Illumination for Lighthouses, 1827; Alan Stevenson, Account of Skerryvore Lighthouse, with Notes on Lighthouse Illumination, 1847; Stevenson, Treatise on the History, Construction, and Illumination of Lighthouses, London, 1850; Report of the Lighthouse Board of America, washington, 1852; J. T. Chance, “On Optical Apparatus used in Lighthouses," in Proc. Inst. C. E., vol. xxvi.; “Dioptric Apparatus in Lighthouses for the Electric Light,” Proc. Inst. C. E., vol. lii; Douglass, “The Wolf-Rock Lighthouse," Proc. Inst. C. E., vol. xxx; "Electric Light applied to Lighthouse Illumination.” Proc Inst. C. E ’.,vol. ]vii.; Thos. Stevenson, Lighthouse Illumination, 2d ed., Edinburgh, 1871 (German translation with appendix, Uel>er die Berechnung von Leuchtthurmhohen und Leuchtthurm·apparatus, by Chr. Nehls, Hanover, 1878); L. Reynaud, Mémoire sur Γeclairage et le balisage des côtes de France, Paris. 1864; Thos. Stevenson, Lighthouse Construction and Illumination, London. 1881; W. Swan, “New Forms of Lighthouse Apparatus," Trans. Roy. Scot. Soc. of Arts, vol. viii.; Μ. E. Allard. Mémoire sur l’intensité et la portée des phares, Paris, 1876; Id., Memoirs sur les phares électriques, Paris, 1881; Proc. Inst. C. E., passim.