1 5 3 Effea of Refrait'ion an the Tijible Horizon. 



tion ; for above C, as to B, the refraflion fomewli.it decreafing, will appear below ihe line 

 a .*, or the parallel to the furface of the land, and the refracVions dccrealc below the point C; 

 for, had they increafed uniformly down to the furface of the fea, it would render the apparent 

 angle of the point of land z more acute than the angle C a O, contrary to all obfervations. 



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

 land near the horizon of the fea when the evaporation is great, and when at the leaft I never 

 found the land quite free from it wlien I ufed a telefcope, and from thence infer that wc 

 cannot have any expectation to find a trye corre£lion for the effetl of terreRrial refraftion 

 by taking any certain part of the contained arc; for the points z CB, Fig. 7, will have various, 

 refrailions, though they are at nearly the fame diftance from the obferver. And if the ob- 

 fervations are made wholly over iand, if the ground riles to within a fmall diftance of the 

 rays of light, in their paflage from the object to the eye, as well as at the fituation of the 

 objedt and obferver, the refractions will be fubjccl to be influenced by the evaporation 

 of rains, dews, <5cc. ; which is fulHciently proved by the obfervations of Colonel Williams, 

 Captain Mudge, and Mr. Dalby, Phil. Tranf. 179J, p- 583. 



The appearances mentioned by Colonel Williams, Captain Mudge, and Mr. Dalby (Phil. 

 Tranf. 179;, p. 586, 587) cannot be deraonftrated upon general principles, as they arife 

 from evaporation producing partial refraclions. In thofe general principles it is fuppofcd 

 that the fame lamina of denfity is everywhere at an equal diltance from the furface of the 

 fea, at leaft as far as the eye can reach a terreflrialobjeft; but in the partial refra£lions, the 

 lamina of the expanded or rarefied medium may be of various figures, according to clr- 

 cumftances, which will refraft according to the incidence of the rays, and affeiSl the ap- 

 pearance of the land accordingly, which I have often feen to a furprifing degree. But my 

 principal view is to (hew the uncertainty of the dip of the fea, and that the efFeft of eva- 

 poration tends to deprefs the apparent horizon at a- when the eye is not above the maximum 

 of denfity ; and from hence the difficulty of laying down any correft formula for thefe re- 

 fra£lions whilft the law of evaporation is fo little underftood ; which indeed feems a taflc 

 not eafy to furmount. The efl^eft indicated by the barometer and thermometer is infuf- 

 ficient ; and (hould the hydrometer be improved to fix a ftandard for moifture in the at- 

 mofphere, and (hew the variations near the furface of the ocean, which certainly muft be 

 taken into the account (evaporation going on quicker in a dry than a moift atmofphere), 

 the theory might ftill be incomplete for corrcfting the tables of the dip. I fliall therefore 

 conclude this paper, by (hewing a method I ufed in practice, in order to obviate this error 

 in low latitudes. 



When I was defirous to attain more accurately the latitude of any headland, &c. ia 

 fight, I frequently obfcrved the angular diftances of the fun's neareft limb from the horizons 

 upon the meridian both north and fouth, beginning a few minutes before noon, and taking 

 alternately the obfervations each way, from the poop, or fome convenient part of the (hip, 

 where the fun and the horizon both north and fouth were not intercepted ; and having 

 found the greateil and lead difiances from the refpc£live horizons, which was at the fun's 

 palfing the meridian, and correfted both for refra£lion, by fubtrafting from the leaft, and 

 adding to the greatcft altitude, the quantity given by the table ; and alfo having corrected 

 for the error of the inftrument, and the fun's femi-diameter, the fum of thefe two angular 

 ^ftances redoce<l as above — 180° 16 equal to double the dip, as by the foUowing 



Example. 



