VOL. LXXXVII.] PHILOSOPHICAL TRANSACTIONS. Q3 



inclined surface of the land ab, fig. 6, touching it at c, fig. 7, the point e will 

 show the height of the maximum of density, where the pencil of the rays of light, 

 from thence to the eye, approach nearest the sea; for pencils of rays from this land, 

 taken at small distances from c, will form parallel curves, nearly, through the 

 refracting mediums, and c will be the point of greatest refraction; for above c, as 

 at b, the refraction somewhat decreasing, will appear below the line ab, or the pa- 

 rallel to the surface of the land, and the refractions decrease below the point c ; for 

 had they increased uniformly down to the surface of the sea, it would render the 

 apparent angle of the point of land z more acute than the angle cao, contrary to 

 all observations. 



Thus I have endeavoured to explain the phenomena of the distorted appearance 

 of the land near the horizon of the sea, when the evaporation is great; and when 

 at the least, I never found the land quite free from it when I used a telescope; and 

 thence infer, that we cannot have any expectation to find a true correction for the 

 effect of terrestrial refraction, by taking any certain part of the contained arc; for 

 the points zcb, fig. 7, will have various refractions, though they are at nearly the 

 same distance from the observer. And if the observations are made wholly over 

 land, if the ground rises to within a small distance of the rays of light in their pas- 

 sage from the object to the eye, as well as at the situation of the object and observer, 

 the refractions will be subject to be influenced by the evaporation of rains, dews, 

 &c. which is sufficiently proved by the observations of Colonel Williams, Captain 

 Mudge, and Mr. Dalby, Phil. Trans. 1795. The appearances mentioned by them 

 cannot be demonstrated on general principles, as they arise from evaporation pro- 

 ducing partial refractions. In those general principles, it is supposed that the same 

 lamina of density is every where at an equal distance from the surface of the sea, at 

 least as far as the eye can reach a terrestrial object; but in the partial refractions, 

 the lamina of the expanded or rarefied medium may be of various figures according 

 to circumstances, which will refract according to the incidence of the rays, and 

 affect the appearance of the land accordingly, which I have often seen to a sur- 

 prizing degree. But my principal view is to show the uncertainty of the dip of the 

 sea, and that the effect of evaporation tends to depress the apparent horizon at x, 

 when the eye is not above the maximum of density; and hence the difficulty of 

 laying down any correct formula for these refractions, while the law of evaporation 

 is so little understood, which indeed seems a task not easy to surmount. The 

 effect indicated by the barometer and thermometer is insufficient: and should the 

 hygrometer be improved to fix a standard for moisture in the atmosphere, and show 

 the variations near the surface of the ocean, which certainly must be taken into the 

 account, evaporation going on quicker in a dry than a moist atmosphere, the theory 

 might still be incomplete for correcting the tables of the dip. I shall therefore 

 conclude this paper, by showing a method I used in practice, in order to obviate 

 this error, in low latitudes. 



When desirous to attain more accurately the latitude of any head land, &c. in 



