553 



pareant, austri mutationem indicant. When a distant island 

 is of a height very unequal, it is the variations in the de- 

 pression of the horizon, and not the inverted image of the 

 sky formed on the lowest parts of the island, that make it 

 look as broken, or divided into several small islands. If 

 Theophrastus meant a multiplicity of images placed one be- 

 neath the other, he would not have neglected to mention in- 

 verted images. Aristotle in his Meteorologica, Lib. iii, cap. 

 ix. p. 577 C. Duval's edition, speaks also of the suspension 

 of capes, and considers it as the effect of a refraction in con- 

 densed air. 



In the course of my observations on the mirage, I carefully 

 made a distinction between the very frequent cases, when 

 there is suspension without inversion. Mr. Biot has related 

 the circumstances, in which this phenomenon takes place. 

 He shows (Rech. p. 261) that the inverted image may be 

 reduced to such small dimensions, as to become impercepti- 

 ble. Respecting the variations of colour, to which the hori- 

 zon of the sea is subject, and the prognosticks of a change of 

 refraction derived from the bands or black streaks*, this phe- 

 nomenon has not always the appearance of several horizons 

 separated by aerial intervals (Biot, p. 10, 151, 183 and 265) . 

 I have never observed these intervals formed by the reflected 

 air j I have merely seen that a great change in the depres- 

 sion is preceded by the formation of streaks in that place, 

 where the new horizon was about to be formed. I have 

 proved above, (chap, iii, vol. ii, p. 67) that near the equator 

 the surface of the sea is almost habitually 1° or 1*5° warmer 

 than the surrounding air. This difference of temperature is 

 considerable enough, to be looked upon as one cause of the 

 mirage. Mr. Woltmann observed, that there was a double 

 image, or simply suspension, on the banks of the Elbe, when 

 the water was only two degrees of Fahrenheit (0 8° cent.) 

 higher than the air. We cannot then be astonished, that the 



* See p. 547. 



