LIGHTHOUSE. 258 LIGHTHOUSE. 3 represents in vcrtica] si-ction this, which is the most perfect of Frcsnel's inventions in light- bouse illiuuination, especially when made in pieces of the rhonihoidal form, and used in con- nection with a diagonal framing introduced by Allan Stevenson. In Fig. 3 p shows the re- fracting and totally reflecting prisms, r the cy- lindric refractor, and f the radiant. From what ha-s been stated, it will l>e readily seen that, in so far as regards fi.ed lights, which are required to illuminate constantly the whole of the horizon with equal intensity, the dioi)tric light of Fresnel with Allan Stevenson's improvements is a perfect instrument. But the case is different as regards revolving light-s, or those where the whole rays have to be concentrated into one or more l)cams of parallel rays. To revert to the paral)olic re- flector, it must be obvious that all rays which escape past the lips of the reflector never reach the eye of the mariner, while, if we return to the dioptric revolving light of Fresnel, we find that those rays which escape past the lens are acted on by itro agents, l)oth of which cause loss of light by absorption. The loss occasioned by the inclined mirrors, and in p.issing through the pyramidal inclined lenses, was estimated by Fresnel himself at onc-hntf of the whole incident rays. In order to avoid this loss of light, Thomas Stevenson proposed in 1S4!) to introduce an ar- rangement by which the use of one of these agent-s is avoided, and the employment of the total reflection which had l>een successfully em- ployed by Fresnel for fi.ed lights was intro- duced with great advantage for revolving lights. He combined an annular lens. L (Fig. 4): a jiambolic conoid, n, triuicated. at its parameter, or between that and its vertex; and a portion PLAN OF THOMAS STEVESSOX's SYSTEM. of a spherical mirror, b. The lens and mirror were placed at the proper focal distance from the flame, so that all the rays of light com- ing from the front of the flame were intercepted either by the paraboloid or the lens. The rays emanating from the flame may be regarded as divided into two hemisplieres. 'llie lens intercc|)ts part of the anterior hemisphere of rays, and makes them parallel while the remainder fall on the paraboloidal surface, and are reflected parallel. The rays forming the posterior hemisphere fall- ing on the spherical mirror are reflected forward again through the' focus in the same lines, but in opposite directions to those in which they came, and then are in part refracted by the lens, and the rest are made parallel by the paraboloid. The result is that rays finally coming from the posterior hemisphere emerge horizontally com- bined with those from the anterior hemisphere. Thus the entire s|ihere of diverging rays is col- lected into one beam of parallel rays without employing any unnecessary agents. What has iH'cn just described was termed by Stevenson a laloptric holupholc. What follows is a description of his dioptric hnfopliotr, in which total reflection is the underlying principle. The front half of the rays is operated upon by totally reflecting glass prisms, similar in section to those applied by Fresnel for fixed lights; but, instead of being curvilinear in the horizontal plane onlj-, they are also curvilinear in the verti- cal plane, and thus |)rodnce. in union with an annular lens, a beam of panillel rays similar to what is ])roduce<l by the paraliolic mirrur. The rays proceeding backward fall upon glass prisms, which produce two reflections u|ion each ray, and cause it to pass back through the llamc. so as to fall ultimately in the proper direction upon the dioptric holophote in front, so that the whole of the light proceeding from the flame is thus ultimately parallelized by means of the smallest number and the best kind of optical agents. It is a remarkable property of the spherical mirror that no ray ])asses through it, so that an ob- server, standing behind the instrument, perceives no light, though there is nothing between him and the flame but a screen of transparent glass. Where the light is produced by great central sta- tionary burners, the apparatus assumes the form of a polygonal frame, consisting of sectors of lenses and holophotal prisms, which revolves round the flame, and each face of which produces a beam of parallel rays. Hence, when the frame revnivi's round the central llame the mariner is alternately illuminated and left in clarUness. ac- coriling as the axis of each successive face is p<iinted toward him or from him. In the re- volving holophotal light one agent is enabled to do the work of two agents in the revolving light of Fresnel, as total reflection, or that by which least light is lost, is substitute*! for me- tallic reflection. The dioptric holophotal system, or that by which total reflection is used as a portion of the revolving apparatus, was first em- ployed on a small scale in 1S50 at the Horsburg lighthouse, and on a large scale in IS.ll at North I?onaldshay in Orkney. Since that date this system has been all but universally intro- duced into Europe and America. AznriTHAi, CoNDEXSiXG LioHT. The above is a description of the general principles on which lighthou.ses are illuminated. In placing a light in some situation, regard, however, must be had to the physical peculiarities of the localities. The following plans of Thomas Stevenson may W cited as examples. Where it is only necessary to illuminate a narrow sound, it is obvious that the requirements are very different than where the light is distributed equally all around the hori- zon. On the side next the shore, no light is re- quired at all: across the sound, a feeble light is all that is necessary, because the distance at which it has to be seen is small, owing to the narrowness of the channel ; while up the sound and down the sound the sea to be illuminated is to be of greater or lesser extent, and requires a corresponding intensity. If the liglit were made sufficiently powerful to answer for the greater distance, it would be much too powerful for the shorter distance across the sound. Such an ar- rangement would occasion an unnecessary waste of oil, while the light that was cast on the land-
