546 



of the horizon, at the moment of the sun's setting, 95° 

 115*7 ; and twelve minutes after, during the twilight, 95° 

 114-7. C, or Picuita, 95° 112. Still later, C, 95° 111-3. 

 F, 95° 116* 3. There can be no error in these observations, 

 for the signal A remains at the same height, while the hori- 

 zon of the sea undergoes such rapid changes. On other days,, 

 I have seen the small islands rest on the horizon a little before 

 the setting of the sun. The setting does not always occasion 

 the same change of temperature, and of the terrestrial refrac- 

 tions. 



" The 24th of September. Since the 18th, sky constantly 

 cloudy. The weather changes in the night of the 23d. Great 

 transparency, the stars very brilliant, but no twinkling, even 

 at the horizon. The 24th great dryness. Hyg. at 21 h morn- 

 ing, 32° Deluc (67 Sauss.). Therm. 21-5° Reaum. Depres- 

 sion of the horizon, the greatest I ever observed. Water of 

 the sea, 22°. The arid soil of the shore, 32*7°. Boracha is 

 entirely in the air. The sky very blue. Cyan. 21. Small 

 fishing- boats suspended, swimming in the air, three or four 

 minutes above the horizon of the sea, which is extremely well 

 defined. One of the boats, viewed through the telescope, 

 presents an inverted image. 



A 94° 62 



B 95° 106 



C 95° 116-2 



E 95° 932 



F 96° 12, consequently near 8' greater than 

 on the 4th of September. 

 te Picuita appears often double, and inverted during the 

 remainder of the day. The inverted image is as large and 

 as high as the direct image. This last is entirely suspended 5 

 but the inverted image, the intensity of the light of which is 

 very weak, trenches on the horizon of the sea : it covers a 

 portion of the last strata of the ocean. At 22 h in the morn- 

 ing, therm. 23 5 y 5 hyg. 31-5°. 



