COASTS OF AUSTRALIA. 
57 
on account of the deceptions appearance of the i82i. 
land; indeed the strait which we discovered Aug. i. 
leading to sea, was not seen until we were 
within two hundred yards of it, and would also 
have escaped our observation, had not the 
channel been so direct, that the sea horizon was 
exposed to our view. At the bottom of this 
arm are two deep bays, which were partially 
but sufficiently examined. In most parts of 
Munster Water there is good anchorage, amongst 
several small rocky islands, on one of which we 
landed, and climbed its summit, but saw nothing 
to repay us for the trouble or the danger of the 
ascent: the surface was composed entirely of 
loose blocks of sand-stone, which, when trod 
upon, would crumble away, or roll down the 
nearly perpendicular face of the rock ; and it 
was only by grasping the branches of the acacias 
and other trees that were firmly rooted in the 
interstices of the less-decomposed rocks, that we 
were saved from being precipitated with them. 
On our return we passed through the channel 
on the west side of the Midway Isles, which 
we found to be very deep, and the stream very 
strong. 
The next day we pulled through the strait that 
insulates Greville Island, and found that it com- 
municated with Munster Water, at a part where 
