48 QUANTFTY OF ttORIZONTAL REFRACtlOW. 



Appirentdlftor- The firft inftance that occurred to me, of obfervable refrac- 

 of vefr-^ S" *'^" ^^'^'^ ^'^^ furlace of the Thames, was wholly accidentaL 

 over tie fufice I was fitting in a boat near Chelfea, in fuch a pofition that 

 •t the Thames, ^^y eye was elevated about half a yard from the furface of the 

 water, and had a view over its furface, that probably fortie- 

 what exceeded a mile in length, when I remarked that the 

 oars of feveral barges at a diftance, that were then coming 

 up with the tide, appeared bent in various degrees, according 

 to their diftance from me. The mofl diftant appeared nearly 

 in the form reprefented, Plate I. Fig. I. dd being my vifible 

 horizon by apparent curvature of the water; ab the oar itfelf 

 in its inclined pofiuon ; and b c an inverted image of the por- 

 tion be. By a little attention to other boats, and to buildings 

 on (hore, I could difcern that the appearance of all diflant 

 objefts feen near the furface of the water was affeded in a 

 fimilar n.anner, but that fcarcely any of them afforded images 

 fo perfedly diftin6t as the oblique line of an oar dipped in the 

 water. 

 Thefe effeas A perfon prefent at the time (as well as fome others to 



ared!fte^nt yvhom 1 have fince related the circumffancel was inclined to 

 iro! what u i 



roighf b" caufed attribute the appearance to refledion from the furface of the 

 bj jreficaion. water ; but, by a moderate (bare of attention, a very evident 

 difference may be difcovered between the inveriion occafioned 

 by refiedlion, and that which is caufed by atmofpherical refrac- 

 tion. In cafes of refledion, the angles between the object and 

 image are fliarp, the line ofcontaft between them ilraightand 

 well defined, but the lower part of the image indefinite and 

 confufed, by means of any flight undulation of the water. But, 

 when the images are caufed by refradion, the confines of the 

 objed and its inverted image are rounded and indiftinft, and 

 the low-er edge of the image is terminated by a ftraight line at 

 the furface of the water. In addition to thefe marks of diffe- 

 rence^ there is another circumftance which, if attended to, 

 muft at once remove all doubt ; for, by bringing the line of 

 fight near to the furface of the water, boats and other fraall 

 objects are found to be completely hidden by an apparent hori- 

 zon, which, in fo fbort a diflance, cannot be owing to any 

 real curvature of the water, and can aiife folely from the bend- 

 ing of the rays by refradion. 

 Tbfiy appear When I rcfleded upon the caufes which were probably in- 



referribie to an ftrumcntal in the produdion of thefe phenomena, they appeared 



higher tempera- * * ' j t r 



tute in the referriblc 



