550 



On examining the angles of altitude, under which the sum- 

 mits of the islands of Boracha and Picuita are seen, we find, 

 that the extent of the variations decreases with the magnitude 

 of the angles. The oscillations of the horizon were 7 ' 57" j 

 those of the summit of Picuita, 2 ' 25" ; those of the summit 

 of Boracha, 0' 27". The real depression of the horizon, in- 

 dependent of any refraction, ought to be 5 f 29" • I found it 

 between 6' 10" and 14' 17". In all these cases the refrac- 

 tion was negative ; that is to say, the trajectories described 

 by the luminous rays were convex toward the surface of 

 the water, at least in their inferior part. We find too> 

 that the apparent base of the island of Picuita was not al- 

 ways placed above the apparent horizon of the sea. It some- 

 times came down to the same height, as on the 5th of Sep- 

 tember at sunset. The island, of course, then appeared rest- 

 ing on the horizon. Sometimes the bases of the island ap- 

 peared even lower than the apparent horizon of the sea, as on 

 the 4th of September j and the surface of the sea was seen 

 then a little beyond the island. " During these variations the 

 trajectories of the luminous rays were convex toward the sea, 

 at least in their inferior part, as the depression of the horizon 

 proves ; but the point of contact of the limiting trajectories 

 with the surface of the sea was more or less distant from the 

 eye ; and it was this circumstance that produced the varia- 

 tions observed in the suspension of the islands, which were 

 seen sometimes more distant, sometimes nearer than the 

 limit." Biot, Recherch., p. 216, 217, 219. 



The influence of the rising and setting of the sun, which is 

 so apparent in my observations on Picuita, confirms what 

 Legentil perceived during his abode at Pondicherry. This 

 distinguished traveller constantly saw* during the winter, the 

 horizon of the sea descend thirty-six seconds, when the first 

 ray of the sun began to appear. The sun rose above the ap- 

 parent horizon, as if it rose out of Chaos. Biot, p. 225. See 

 also my Recueil oVObs. Astronoimques, torn, i, p. 153. 



I frequently remarked, that the two capes of the island of 



