35 
on Horizontal Refractions. 
pencils of rays a a and d d, exceed the angle made by a a , and b b , 
the apparent height of the sail of the vessel ; for were those 
angles equal, the inverted sail would only be contracted into 
the parallel of altitude of the boom b, and render the appear- 
ance confused, as in the hull of the vessel. 
Respecting the existence of two pencils of rays entering the 
eye from every point of an object not more elevated than a, or 
less than i, fig. 3, in this state of the atmosphere, I cannot bring 
a stronger proof than that of the strength of a light when the 
rays pass near the horizon of the sea, proved by the following 
observations. 
Going down Channel about five years ago in the Trinity 
yacht, with several of the elder brethren, to inspect the light- 
houses, &c. I was told by some of the gentlemen, who had been 
on a former survey, that the lower light of Portland was not so 
strong as the upper light, at near distances, but that at greater 
distances it was much stronger. I suspected that this differ- 
ence arose from the lower light being at or near the horizon 
of the sea, and mentioned it at the time ; but afterwards had a 
good opportunity of making the observation. We passed the 
Bill of Portland in the evening, steering towards the Start, a 
fresh breeze from the northward and clear night ; when we 
had run about five leagues from the lights, during which time 
the upper light was universally allowed to be the stronger, 
several gentlemen keeping watch to make observations 
thereon, the lower light, drawing near the horizon, suddenly 
shone with double lustre. Mr. Strachan, whose sight is 
weak, had for some time before lost sight of both lights, but 
could then clearly perceive the lower light. I then went aloft, 
(as well as others,) but before I got half mast up, the lower 
F 2 
