Kangaroo Island.-] TERRA AUSTRALIS. 185 
sav, that it was the sole place found by us where the hope of pro- 1802. 
1 • j April. 
curing fresh water could be entertained. Monday 5. 
Having received on board a good stock of wood, the launch 
was hoisted in, and every thing prepared for going to sea. Next 
morning, so soon as the sun was sufficiently elevated to be observed Tuesday a. 
in the artificial horizon, I landed to take the last set of observations 
for the time keepers ; after which the anchor was weighed, and we 
steered out of Nepean Bay with a light breeze from the south-west. 
Towards noon it fell calm, and finding by the land that the ship was 
set westward, an anchor was dropped nearly in our first place off 
Kantjuroo Head ; and Mr. Westall took the sketch given in the (Atlas, 
0 Plate XVII. 
Atlas. View 11.) 
The rates of the time keepers were obtained, for the sake of 
expedition, from single altitudes of the sun's upper and lower limbs, 
taken from a quicksilver horizon with a sextant fixed on a stand; 
the time being noted from Arnold's watch, compared with Earn- 
shaw's time keepers before going on shore and immediately after 
returning. From the altitudes of the 3rd, 4th, and 6th, in the morn- 
ing, the rates of the two time keepers and their errors from mean 
Greenwich time, reduced to noon there on the last day, were as 
under : 
Earnshaw's No. 543, fast o h o' i8",03 and losing S",^6 per day. 
No. 520, slow o 45 29,56 - - - 18,07 
In deducing these errors, the longitude given by the time 
keepers on our first arrival from Spencer's Gulph, which I consider 
to be equally good with that of Port Lincoln, was used, with a cor- 
rection of 1' 20" for the change of place. The medium of the Port- 
Lincoln rates was something greater than that now found ; which 
corresponded with the time keepers having given the longitude of 
Kanguroo Head less on the second than on the first arrival. This 
was some proof that the letting down had not affected the rates, and 
tended to give me confidence in their accuracy. 
