in order to prevent the too rapid condensation of heated air on the large internal surface of the lantern roof, which consists in having two domes with an air-space between them, as shewn in the enlarged diagrams in Plate XXVI. An important improvement in the ventilation of Lighthouses was some years ago introduced by Dr FARADAY into several of the Lighthouses belonging to the Trinity House, and has since been adopted in all the dioptric lights belonging to the Commissioners of Northern Lighthouses. After mentioning several proofs of extremely bad ventilation in Lighthouses, Dr FARADAY thus describes his apparatus:[83] [83] Minutes of Institution of Civil Engineers, vol. i., p. 207. “The ventilating pipe or chimney is a copper tube, 4 inches in diameter, not, however, in one length, but divided into three or four pieces; the lower end of each of these pieces for about 1¹⁄₂ inch is opened out into a conical form, about 5¹⁄₂ inches in diameter at the lowest part. When the chimney is put together, the upper end of the bottom piece is inserted about ¹⁄₂ inch into the cone of the next piece above, and fixed there by three ties or pins, so that the two pieces are firmly held together; but there is still plenty of air-way or entrance into the chimney between them. The same arrangement holds good with each succeeding piece. When the ventilating chimney is fixed in its place, it is adjusted so that the lamp-chimney enters about ¹⁄₂ inch into the lower cone, and the top of the ventilating chimney enters into the cowl or head of the lantern. “With this arrangement, it is found that the action of the ventilating flue is to carry up every portion of the products of combustion into the cowl; none passes by the cone apertures out of the flue into the air of the lantern, but a portion of the air passes from the lantern by these apertures into the flue, and so the lantern itself is in some degree ventilated. “The important use of these cone apertures is that when a sudden gust or eddy of wind strikes into the cowl of the lantern, it should not have any effect in disturbing or altering the flame. It is found that the wind may blow suddenly in at the cowl, and the effect never reaches the lamp. The upper, or the second, or the third, or even the fourth portion of the ventilating flue might be entirely closed, yet without altering the flame. The cone junctions in no way interfere with the tube in carrying up all the products of combustion; but if any downward current occurs, they dispose of the whole of it into the room without ever affecting the lamp. The ventilating flue is in fact a tube, which, as regards the lamp, can carry everything _up_ but conveys nothing _down_.”