ATMOSPHEROLOGY. — llEFR ACTION. 391 



stern, and a peculiar distortion of the masts ; a di- 

 vision of a ship in the middle, and a lengthening of 

 the hull, &c. ; but the preceding having been more 

 attentively studied in connection with the state of 

 the atmosphere, they are more satisfactory, and con- 

 sequently more worthy of particular detail. 



From the whole of these facts, the following re- 

 marks may be deduced. 



1^^, That the curious refractions of the atmos- 

 phere in the polar regions, as far as they have been 

 observed, have usually occurred in the evening or 

 night, after a clear day. 



2d, That they are most frequent on the com- 

 mencement or approach of easterly winds. And, 



3<:/, That they are, probably, occasioned by the 

 commixture, near the surface of the land or sea, of 

 two streams of air of different temperatures, so as to 

 occasion an irregular deposition of imperfectly con- 

 densed vapour, which, when passing the verge of the 

 horizon, may produce the phenomenon observed*. 



Those phenomena, considered as the effects of re- 

 fraction, &c. which remain to be mentioned, are not 



• Perhaps the refraction of the dense vapour incumbent on 

 the surface of the Thames, which at high water brings into 

 the view of a spectator on the opposite bank, objects that are 

 invisible at low water, may, in some measure, illustrate, or 

 serve to accomit for this phenomenon. 



