The perception is correct, but its meaning is not interpreted properly
The Searchlight Effect was reported, unwittingly as usual, on the front page of the Sunday Times for 2 January 2000: describing a powerful green laser beam, Bryan Appleyard wrote "Its beam followed the meridian north out of the park, passing well to the left of the dome and then across the Thames where, to our eyes, it vanished although apparently it went on for another seven miles." Sometimes called the Searchlight Illusion, this is not an optical illusion but simply a misinterpretation of what is quite correctly seen.
When gazing at the 'end' of the beam, Mr Appleyard was looking in the exact direction of the beam, seeing it at a very small angle so that the intensity of the beam appeared great, though it was actually being attenuated by atmospheric absorption as it went. The weaker the beam became, the longer the length of his line of sight along it. This effect keeps the beam bright right up to its apparent end, which makes the effect quite striking. When a number of strong searchlights are directed upwards, each shaft of light appears to end in a point whose location depends on the direction of the beam. Stars can be pointed out by a lecturer to her audience using a strong laser, and the beam will seem to end near the same star for any observer.
The figure illustrates the origin of the effect. A is seen behind the beam, the beam seems to pass behind B, and ends just before reaching C, with regard to the observer. The effect persists for an observer a considerable distance from the source of the beam, but when this seems to happen with a distant searchlight, the reason for the termination of the beam is more likely to be the limit of dust or haze in the atmosphere. The difference can easily be identified if the beam is moved about, however. If what is observed is attenuation of the beam, it will also be obvious at a distance from the weakening of the beam, in contrast to its rough constancy in the true Searchlight Effect.
Composed by J. B. Calvert
Created 2 January 2000
Last revised 17 April 2000