26 A VOYAGE TO IFrom Madeira. 
1801. much water having passed down them. By the side of one of these 
SaturdafiB. gullies, which was near the place where we lost sight of the boat, 
there was a path leading up into the interior of the island. The 
south-west and south points are low; they lie N. 14 0 W. and S. i4°E. 
and are five or six miles asunder. Between them, the land hollows 
back so as to form somewhat of a bay, which, if it afford good 
anchorage, as it is said to do, would shelter a ship from all winds 
between north and east-south-east. We did not observe any beach 
at the head of the bay, perhaps from having passed at too great a 
distance. 
No observations could be taken for fixing the situation of this 
island; but in 1795, Mr. Crosley and myself made the high land 
near the south-west point to lie in 17 0 00' north, and by uncorrected 
lunar observations, in 25° 32' west; which agrees well with the 
position of the north-west point, as given by captain Vancouver.* 
The variation from azimuth on the evening of the 14th, before 
making the land, was 13° 51' west, and 13 0 3' this evening, when 
four leagues to the west of it; the compass being placed on the 
binnacle, and the ship's head south-south- west (magnetic) in both 
cases. The true variation here, at this time, I judge to have been 
12 0 24' west. Captain Vancouver observed 12 0 32', in 1791; but it 
does not appear how the ship's head was directed. 
Some distant land opened from the south point of St. Antonio, 
at S. 75 0 E. ; which I took to be a part of the island St. Lucia. 
During the three days before making St. Antonio, the wind 
varied from the regular north-east trade, to east-north-east, and as 
far as south-east-by-east ; and about the time of seeing the land, it 
dwindled to a calm. For three days afterwards it was light, and 
variable between north and south-east ; after which it sometimes 
blew from the north-west and south-west, and sometimes from the 
eastward. These variable winds, with every kind of weather, but 
Sunday 23. most frequently with rain, continued until the 23rd, in latitude 1 1° 
* Fbyage round the World, Vol. I. page 10 a 
