1839] 



in the Madras Light H^use, 



277 



would be visible, in the same sitiialioiis the rays from the other points 

 would also be seen, constituting images of the flame or luminous body, 

 of whatever kind. 



In this manner an image of a candle will be seen by reflection from a 

 circular looking glass or mirror, in all those situations embraced by a 

 cone of rays proceeding from it in the directions pointed out by the 

 above law— namely, if A (Fig. 7 Pi- 4) be the light, and B C a circular 

 plane mirror, a reflected image of A will be visible from any point with- 

 in the frustum of the cone B C gh, formed by the reflection of the rays 

 impinging on the circular area Y^dC e. 



If the mirror be square then a reflected image of the light will be 

 visible in any part of the pyramid of rays, formed in the same way, as 

 shown in Fig. 8. 



From what has been said in the former part of this paper, taken in 

 connection with these circumstances, it will be easily seen, that in order 

 to produce the utmost beneficial effect, no part of a planemirror, if placed 

 vertically, ought to be above the uppermost part of the light to which it 

 is adapted, as in that case, the rays which would be reflected from that 

 part of it, would be those belonging to the upper hemisphere, which I 

 have already shown to be useless (vide Fig. 9). I am not aware whether 

 this circumstance has been attended to in the disposition of the reflectors 

 in the Madras Light-hous?, having never examined the apparatus, but I 

 shall suppose such to be the case, and that they are disposed to the 

 greatest advantage they are capable of, as even were it not so, it is a de- 

 fect which could be very easily remedied, and it is my object to show, 

 not that a bad use has been made of eflficient means, but that the con- 

 trivance itself, under the best arrangement, is of much less value and 

 assistance than has generally been supposed. 



I shall now proceed to a description of the lantern of this Light-house, 

 and to dissect the operation of, and calculate the value, of the assistance 

 rendered by the reflectors contained in it. In doing this, I shall as I be- 

 fore stated, suppose every mirror to be placed in the most advantageous 

 position (with regard to height), viz., so that the image reflected from 

 the upper portion of it is visible on the edge of the horizon, and that the 

 lower portions in tuiii, successively, afi^ord light to the points nearer to 

 the building itself. 



The shape of the iuntern is an octagon, three sides of which, to- 

 wards the land, are opaque, and the remaining five sides glazed. On the 

 three first opaque sides are placed the reflectors, in three rows, one 

 above the other ; they are squares of one foot each side, of looking- 

 glass, and occupy the centre of each of the sides of the polygon as shown 

 in the plan Fig. 13. 



