ON LIGHTHOUSE APPARATUS. 



319 



Figure 3. 



a paraboloid, L lens, b spherical mirror. 



lens must subtend the same angle at the flame as the cone of rays which 

 escapes past the lips of the reflector. If this lens be put in the position 

 I have spoken of, it is quite obvious that all the rays are rendered 

 parallel, and there is no loss of light. But I have only spoken of one 

 half of the light, namely, the front half, I have said nothing about the 

 back half. Now the rays falling on the back portion of a parabolic 

 reflector are parallelized by it. The object of the lens is also to 

 parallelize the rays, and therefore, if the rays being already parallelized 

 by the reflector were to fall on the lens they would be caused to con- 

 verge, and afterwards to diverge, and thus be lost as before. In order 

 therefore to utilise the light from behind, it is necessary to cut off the 

 back of the parabolic reflector, and to beat it into a spherical form, so 

 that the rays falling normally on this surface will be returned back 

 again through the flame, and proceeding in a diverging direction will 

 be intercepted either by the paraboloid or by the lens in front. In 

 this way, the whole of the light will be utilised, and this therefore ought 

 to be the light of maximum intensity. So far, it is an obvious im- 

 provement upon what Fresnel uszd for utilising the light which passed 

 above the lens, for you observe he used two agents in his revolving light, 

 whereas, if his plane mirrors had been made portions of parabolas, 

 he could have dispensed with the inclined lenses which are inside the 

 apparatus, and one agent would have done instead of two. But after 

 all, although optically considered, this form cf holophote is geo- 



