X. LIGHTHOUSES, ETC. 541 



phenomenon known in optics under the name of " total reflection," and 

 proposed to substitute for the common reflectors glass rings, within which the 

 luminous rnys should be reflected without appreciable loss. 



Fresnel's first conception for these circular rings was to place the fascets 

 through which the luminous rays pass perpendicularly to these rays, so as 

 not to alter their direction ; the reflecting surface would then have preserved 

 the shape of the mirrors to be replaced, but hence resulted inconvenience, 

 and a too great weight of glass. Fresnel found out that a direction inclined 

 to the rays could be given to the fascets, and thus these inclinations be com- 

 bined, as well as the shape of the reflecting surface, so as to cause the rays 

 to emerge horizontally. The transverse section of the rings then became 

 triangular, instead of showing four sides, and the dimensions were reduced. 



The apparatus exhibited is that to which Fresnel first applied this invention. 

 Its diameter is reduced to 0'20 m ; the cylinder is generated by an echeloned 

 section composed of three elements and fills up a half circumference. The 

 rays passing above this cylinder are gathered by four total reflection rings, 

 and this is effected by turning around the vertical point of the focus the 

 section of the catadioptric triangles just spoken of. Thus is obtained a fixed 

 light apparatus, lighting up half the horizon. The lamps of the St. Martin's 

 Canal having to be erected at 70 metres distance, it became necessary to give 

 them a greater lateral than frontal intenseness. Fresnel succeeded in this 

 by placing on each side a half annular dioptric lens, generated by the rotation 

 of the section of the cylinder around an horizontal axis, parallel with the 

 longitudinal direction of the quay, but he had moreover the happy idea of 

 making the section of catadioptric triangles to revolve around this axis so as 

 to form an annular lens, collecting around the focus an angle of great 

 amplitude, and comprising at the same time dioptric and catadioptric rings. 



The manufacture of these different circular rings offered serious difficulty, 

 and Fresnel was obliged to set up a factory. 



A first apparatus was completed in 1826, and submitted to the Lighthouse 

 Committee towards the end of December. Four of these new lamps were 

 finished at the beginning of 1827, but they could not be tried until after the 

 inventor's death. 



This study shows how Fresnel came to invent not only the section of cata- 

 dioptric rings, and the use of these rings in fixed light apparatus, but also 

 to apply them to annular lenses for flashing lights or for fixed lights. By 

 uniting the pieces of dioptric elements and of catadioptric rings manufactured 

 in Fresnel's time for the apparatus of the St. Martin's Canal, the annular lens 

 exhibited under No. 6 was formed, and it may be considered as the type 

 of all the annular lenses used in the lighthouses of different order. 



The model in wood, No. 7, represents an apparatus similar to the preceding, 

 but having a diameter of 0'25 m instead of 0'20 m . It is a study of Fresnel's 

 which he did not carry out. 



Lamp Burners. 



No. 8. One of the first burners, with four concentric wicks, constructed 

 after experiments made by Arago and Fresnel in 1819-20. 



No. 9. Burner, with two wicks and outer casing for directing the draught, 

 constructed by Henry-Lepaute in 1845. 



No. 10. Burner, with five wicks, of graduated shape, for mineral oil, with 

 the last improvements adopted in the lighthouses of France, 1876. 



When Fresnel undertook the improvement of lighthouses, he had to solve 

 not only the problem of construction of the lenses, but also that of lamps with 

 several wicks. The chemist Guyton-Morvau had already studied the ques- 

 tion. In a paper read by him at the Institute in 1797, he stated that he had 

 constructed, 10 years before, a lamp on the Argand principle, with three 



