(See note, Appendix A.) GOULBURN ISLANDS.

The observations were taken upon Bottle Rock, the largest of two rocky islets at the north end of South-west Bay; but the results were so doubtful and unsatisfactory, that the longitude determined by the chronometers was preferred.

The following are the observations that were taken to fix its situation, namely: Latitude by fourteen meridional altitudes of the sun l.l.on the sea-horizon, taken in various parts of the bay, and reduced by survey to Bottle Rock 11 37 24.
The difference of longitude between Bottle Rock and Cassini Island by chronometers, taken in: 1819: 7 40 47.
1820: 7 40 00.
1821: 7 38 28.
Mean difference between Cassini Island and Bottle Rock: 7 39 45.
Longitude of Cassini Island from Careening Bay, by survey: 125 38 46.
Longitude of Bottle Rock, by chronometer, from Cassini Island: 133 18 31.
The mean of the results of the lunar distances that were taken during the years 1818 and 1819, gave for the longitude of the rock 133 degrees 31 minutes 58 seconds East.

On our last voyage the mean of the Bathurst's and Dick's watches made it 133 degrees 19 minutes 40 seconds, which was finally adapted, since it accorded better with the chronometrical difference between its meridian and that of Cassini Island.

I have never been able to account for this extraordinary disagreement between the results of the lunar distances and the chronometers, since the former were taken with the sun on both sides of the moon, and seemed to be very good.
CAREENING BAY.