On the 13th, in latitude 4° 44' south and longitude 23° 17' west, a swallow, a gannet, and two sheerwaters were seen; and from six to eight in the evening, the officer of the watch and myself thought the water to be much smoother than before, or than it was afterward. Had it been in an unknown sea, I should have been persuaded that some island, or shoal, lay at no great distance to the south-eastward of our situation at that time.
SUNDAY 20 SEPTEMBER 1801
The trade wind continued, with some little variety in its direction, to blow fresh until the 20th, when it became light, and sometimes calm. We were then approaching the small island Trinidad. Many gannets were seen at twenty-four leagues off, but none at a greater distance. On the 23rd [WEDNESDAY], the island was in sight; and at noon, when our latitude was 20° 1' south, and longitude 29° 13' west, a peaked hummock near the eastern extremity bore S. 25° W., nine or ten leagues. The western extremity bore S. 29° W., and at first appeared to be a bluff head; but it afterwards assumed the form of a conical rock, and was, in all probability, the Nine Pin of captain D'Auvergne's chart. One of the rocks called Martin Vas, was visible from the main top, and angled 49° 43' to the left of the peaked hummock; its bearing was consequently very near S. 25° E.
Mons. de la Pérouse, who sent a boat on shore to Trinidad, lays down the latitude of the south-east point at 20° 31' south, and longitude from lunar observations, 28° 37' west of Greenwich. The latitude appeared to agree with our observations; but in the longitude there is some difference. According to Earnshaw's two time keepers, No. 465 and 543, which kept better rates than the remaining four, the longitude of the Nine Pin is 29° 25½' west; which being reduced to the south-east point, will place it in 29° 23', or 46' west of the French navigator.* The longitude in captain D'Auvergne's plan of Trinidad, constructed 1782, is 29° 55', or 32' still further west. From two sets of distances of the star Altair to the west, and two of Aldebaran east of the moon, I made the longitude of the south-east point to be 29° 19' west; the difference from the time keepers, which I consider to have given the best longitude, being no more than 4'.
[* The error of No. 465 was found, at the Cape of Good Hope, to be 10' 57", 2 to the east, and of No. 543, to be 39' 21", 5 east, contracted in 96 days upon their English rates. To obtain the above longitude, a proportional part of these errors according to the number of days, has been applied to the keepers; and the difference between them is then no more than 2".]
THURSDAY 24 SEPTEMBER 1801
Azimuths taken upon the binnacle in the morning, with three compasses, and the ship's head at S. W. by S., gave variation 3° 54'; and in the evening, at S. W., 3° 50'; but next morning, when Trinidad was just disappearing from the deck in the N. 60° E., other azimuths then showed the variation to be 1° 35' west, the ship's head being S. S. W.; it therefore appears, that there is a difference off the north, and off the south-west sides of the island. From the first observations I deduce the true variation to be 4° 14' west, and from the last 1° 50' west. Captain D'Auvergne marks the variation 0° 45' west, in 1782; but under what circumstances it was ascertained, does not appear.
TUESDAY 29 SEPTEMBER 1801
The trade wind having again arisen from east-south-east, we were enabled to make between eighty and ninety miles a day. It afterwards veered gradually round, by the north-east and north, to the westward, and blew fresh; so that on the 29th, our latitude was 31° 2' and longitude 26° 0' west. This was 17' to the south, and about 6° west of the situation usually assigned to Saxemberg; an island which has been frequently sought by the East-India, and other ships, in the place which it still occupies in the charts; and not finding it there, they have run a few degrees to the eastward, in the same parallel, but always without success. The opportunity which presented itself of now adding 6° of longitude to the examined space, and on the opposite side, I should have thought myself culpable in neglecting; and therefore, having placed the ship in the supposed parallel of the island, we steered due east for it; adopting the same regulations for the look-out at night, as when searching for St. Paul's.
We had seen an unusual number of pintado and sooty petrels on the preceding afternoon, as also of a brown bird, apparently one of the sea-swallow tribe, having a white belly and the form and size of a woodcock; and this evening it was reported to me from the mast head, and confirmed by others on deck, that a turtle was seen lying upon the water. These indications of land gave me some hope that the long lost Saxemberg might be brought to light. On the following noon [WEDNESDAY 30 SEPTEMBER], the observed latitude was 30° 41' and longitude 22° 46'; and nothing further had transpired to betoken the vicinity of land. Next day [THURSDAY 1 OCTOBER], our observations gave 30° 34' south, and 20° 28' west; and I then steered east-south-east, a course which should have taken us almost directly over the supposed situation of Saxemberg, if the same current of 11' north had prevailed, as on the preceding day. But this not proving to be the case, our track lay a few miles to the south; though sufficiently near for us to be satisfied of the non-existence of the island in the place assigned to it, if that could any longer admit of a doubt.*