for the purposes of Illumination in lighthouses. 47 



Sir ISAAC NEWTON has himself remarked, " That every irregula- 

 rity in a reflecting superficies, makes the rays stray five or six times 

 more out of their due course, than the like irregularities in a refract- 

 ing one /' and we may therefore easily conceive what a scattering 

 and dispersion of the rays must take place from a surface ham- 

 mered into a parabolic curve. This dispersion may not appear so 

 conspicuous, when we examine the reflected beam near the reflec- 

 tor itself; but at moderate distances even, it must exercise an 

 enormous influence, in weakening the intensity, disturbing the pa- 

 rallelism, and consequently destroying the uniform density of the 

 reflected column of light. The second source of imperfection of 

 surface, namely, the scratches and striae, will be easily under- 

 stood by those who have examined the beautiful Iris ornaments 

 of Mr BARTON. All the light which falls upon the scratches 

 on a metallic surface, is reflected in coloured pencils to a distance 

 from the direction of the rest of the light ; and this distance in- 

 creases with the number and closeness of the scratches. Not 

 a single ray of this coloured light can ever enter the main beam 

 of a lighthouse reflector, so that it is entirely lost. By standing 

 in front of one of these reflectors, it will be seen, that these 

 scratches are so numerous, that the surface has the appearance 

 of being covered with the finest hair. If the surface had been 

 regularly ground and polished upon pitch, like the specula of te- 

 lescopes, no such effect would be produced ; but this cannot be 

 done with parabolic reflectors. 



3. On the Imperfection of the Parabolic figure. The practical 

 execution of a parabolic surface for optical purposes, has long 

 been regarded as a very difficult operation, even when effected 

 by the nicest machinery. Hence, the operation of forming a 

 parabolic surface by a gauge and a hammer, directed solely by 

 the eye of a workman, is not likely to be successful. Had 

 such a surface been intended for that of a solid for ornamen- 



