548 



was 95° 119) is quite white. I have observed this oscil- 

 lation for several days. The variations of colour are the 

 prognostics of a change in the depression. Ought we not to 

 admit, that the luminous rays sent to us by the most distant 

 part of the sea, that which grows paler, are inflected in such 

 a manner, during their passage through the inferior strata of 

 the atmosphere, that in their convex curvature toward the 

 surface of the Earth, they are confounded with the rays of 

 the nearer stratum of the sea ? We can judge only from the 

 direction of the rays ; and those rays, which really arrive 

 from the most distant strata of the water, being inflected, 

 appear to us to belong to the nearest. It is this circumstance, 

 which occasions the appearance of streaks, and which aug- 

 ments that intensity of colour or brightness, remarked in the 

 new horizon. 



(( All these phenomena are equally observable in the bar- 

 ren steppes of Caraccas, and on the borders of the Oroonoko, 

 in those places where sandy plains surround the river. We have 

 frequently seen the mirage this year, 1800, between Calabozo 

 and San Fernando de Apure ; and at the Oroonoko, to the north 

 of the Mission of Encaramada. The little hills of San Juan and 

 Ortiz, the chain called Galera, appear suspended, when seen 

 from the steppes, at three or four leagues distance. The 

 sand, at noon, was heated (in the sun) as high as 42° Reau- 

 mur. At eighteen inches above the surface, the thermometer 

 kept at 32° Reaum. At six feet (in the shade) it marked 

 29 5° R. Palm-trees, standing single in the Llanos, ap- 

 peared to be cut off at bottom : it seemed as if a stratum of 

 air separated them from the ground. The plains, destitute 

 of vegetation, appeared to be meers, or lakes. It is the illu- 

 sion so common in the African deserts. At the Mesa de 

 Pavones, in the middle of the steppes of Caraccas, Mr. Bon- 

 pland and I saw cows suspended in the air. Distance one 

 thousand toises. Measuring with the sextant, the breadth of 

 the aerial interval, we found the animal's feet elevated above 

 the soil 3' 20 v . Simple suspension, no double image. I was 



