at the Mast-head in the River Plate. 215 



his hand to it, on which it burnt with a hissing noise resembling 

 the burning of a portfire, at the same time emitting a dense 

 smoke, without any sensible smell. On taking away his hand 

 it resumed its former appearance ; but, when he applied the 

 sleeve of his wet jacket, it run up it and immediately went out, 

 and never again appeared. The one on the vane-stafF retained 

 its position for upwards of an hour, but, on account of the 

 heavy rain, and probably also from having been struck by the 

 vane attached to the staff, it went out ; but resumed its position 

 after the rain had ceased, although with a less deo-ree of brio-ht- 



ness. 



The second time I observed the same phenomenon was in the 

 month of December of the same year, while off the coast of Pa- 

 tagonia, in Lat. 36° 40' S., and Long. 54° 40' W., a little to the 

 southward of the River Plate. The night on which it was ob- 

 served was exceedingly dark, but rendered as light as day by 

 the continued flashes of lightning, which spread over the sky 

 and around the whole horizon, so much so indeed that the firma- 

 ment presented a complete mass of fire. At this time I ob- 

 served the same luminous appearance, and attached to the same 

 staff as before mentioned. It retained its situation for some 

 hours, but when the lightning subsided it became gradually 

 fainter, and at last entirely disappeared. 



I may here remark, that the above lights were not attached 

 to any kind of metal in particular ; for the one on the yard-arm, 

 and the other on the vane-staff, were both attracted by metal, 

 the former to an iron-bolt, the latter to a spindle of copper ; 

 while the one seen in the month of December was attached to a 

 spindle of hard- wood, the copper having been removed to guard 

 against lightning. 



After any of the above phenomena having been seen, we al- 

 ways had very bad weather, commencing with heavy and sud- 

 den squalls, generaUy from the sSouth-west, but varying a few 

 pomts each way, and settling in a iew hours to a steady gale. ^ 



In addition to these remarks, I may mention, that, on the 

 coast of South America, the barometer is very sensible of any 

 change In the weather, particularly with regard to the pampe- 

 ros. Its general height in fine weather is 30.10, varying very 

 little from that point ; but, on the approach of the pampeso, the 



