88 THE LABRADOR PENINSULA. cnAr. xxvi. 



Gulf; and a telescope generally enables the observer to 

 detect in the confused and highly distorted image of 

 a ship or boat, high above the horizon, as many as three, 

 and sometimes five, images of the object, blended together 

 and overlapping one another. Local changes in the 

 temperature of the surface water, caused by puffs of wind 

 bringing the cold water to the surface, and mingling it 

 with the warm superficial stratum arising from rivers or 

 proximity to land, are the chief causes of mirage. Dr. 

 Kelly states that, during Admiral Bayfield's survey of 

 the Gulf, mirages were most frequently observed at Bic, 

 Point de Monts, Mingan, and the Straits of Belle Isle. 

 Some of these displays are so striking, that a description 

 of one will convey a good idea of this very beautiful 

 phenomenon, which is often as grandly displayed in the 

 Gulf as in the magnificent source of the St. Lawrence, 

 Lake Superior. 



We were off Metis on the afternoon of September 14, 1835. 

 There was a light easterly wind and cloudy sky ; the tem- 

 perature of the air 48°, the dew point 40*5°; the surface water 

 39-5°. The barometer 29*90° falling. Some light rain fell 

 two or three times during the afternoon, and we had very heavy 

 continuous rain after night falL 



Several vessels were in sight between 3 p.m. and 4 p.m., and 

 all presenting a variety of appearances from refraction. The 

 most remarkable was that in which a vessel with all sail set 

 at one moment looked like an immense chest, no sails or masts 

 being visible. On observing her for a time, the black body 

 seemed to separate horizontally into two parts ; and two sets of 

 mingled sails occuj)ied the intervening spaces, with one set of 

 very small sails above. The figures afterwards became more 

 distinct, and three images were clearly discerned. 



Captain Bayfield and Mr. Bowen observed five distinct images 

 of another vessel, after I left the deck. 



