162 
SURVEY OF THE INTERTROPICAL 
1822 . Peron and Bouvard, and about five miles from 
Jan. 12. the shore, which from the former extended in a 
N.E.b.N. direction, still low and sandy. 
At noon the latitude was observed to be 32® 
30' 42", but by the land it was only 32® 23' 30", 
a difference of 7' 12". This error was occasioned 
by the haze which concealed the true horizon, 
and caused an appearance of land all round us, 
on which rocks, sandy beaches, and trees were 
so plainly formed, that the officer of the watch 
actually reported two islands on the western hori- 
zon. This was the most remarkable instance of 
“ mirage"' that we ever witnessed ; the haze had 
only commenced a few minutes before noon, whilst 
the observation for the latitude was in the act of 
being taken ; and immediately after I was em- 
ployed upon the chart for half an hour, puzzling 
myself in attempting to reconcile the observed lati- 
tude with the bearings of the land. This curious 
phenomenon was also witnessed by the French 
in Geographe Bay. During the time this magical 
appearance continued, we had very light airs 
from the southward: the barometrical column 
fell to 29.76 inches, but the hygrometer indicated 
an extraordinary dryness of the air. At sunset 
the haze cleared away, when Rottnest Island 
was seen, bearing between N. 10° and 32® E. 
(mag.); a breeze then freshened from W.S.W., 
