163 



detach itself from the horizon. This limb was 

 not visible till 4* 56' 56". The disk of the sun, 

 much flattened, was well defined ; during the 

 ascent, there was neither double image nor 

 lengthening of the lower part. The duration * 

 of the sun's rising being triple that which we 

 might have expected in this latitude, we must 

 suppose, that a fog bank, very uniformly ex- 

 tended, concealed the true horizon, and followed 

 the sun in it's ascent. Notwithstanding the li- 

 bration of the stars f which we had observed 

 toward the east, we could not attribute the 

 slowness of the rising to an extraordinary re- 

 fraction of the rays occasioned by the horizon of 

 the sea ; for it is precisely at the rising of the 

 sun, as Le Gentil daily observed at Pondicherry, 

 and as I have several times remarked at Cu- 

 mana, that the horizon sinks, on account of 



* The apparent duration was 8' V instead of 2' 41". 

 Though my journals contain near eighty observations of the 

 rising and setting of the sun, made either during the voy- 

 age, or on the coasts, I have never perceived any sensible 

 retardation. 



+ A celebrated astronomer, Baron Zach, (Mon. Corres- 

 1800, p. 306) has compared this phenomenon of an apparent 

 libration of the stars to that described in the Georgics (lib. i, 

 v. 365) . But this passage relates only to the falling stars, 

 which the ancients, as well as our mariners, considered as a 

 prognostic of wind. The Latin poet appears to have imitated 

 the verses of Aratus. (Diosem. v. 926, edit. Buhle, i, p. 206. 

 Lucret. W, v. 143.) 



M 2 



