28th February.

At 6.0 a.m. we were again in the saddle, following a creek which had an
average west-south-west course, but the channel was soon lost in a wide
grassy flat, with polygonum and atriplex, in this flat were some large
detached pools of water, 50 to 100 yards wide and a quarter to half a
mile long, although the dry season had reduced them to much narrower
limits than usual, as they were now eight to ten feet below the level of
the plain; at 11.45 camped at a large sheet of water, just above a
remarkable ridge of sandstone rocks on the right bank of the creek.
Ducks, pelicans, spoonbills, etc., were very numerous, but so wild that
they could scarcely be approached within range of our guns; until the
present time it has been doubtful whether the creek turned towards
Cambridge Gulf, the interior, or to the coast westward of the Fitzroy,
but the first point being now 220 nautic miles to the north, and the
general course of Sturt's Creek south-west, such a course is not
probable, and it therefore only remains to determine whether it is lost
in the level plains of the interior, or finds an outlet on the north-west
coast. The careful and minute surveys of the coast from the Victoria
River to Roebuck Bay show that no rivers exist of such magnitude as the
Sturt would attain in passing through the ranges to the coast, nor does
the general abrupt character of the coast-line favour the supposition
that any interior waters would find an outlet in this space. That the
elevation of this part of the creek is sufficient to enable it to form a
channel to the north-west coast is shown by the barometric measurement:
the dividing ridge between the head of the Victoria and Hooker's Creek is
about 1200 feet, at the head of Sturt's Creek 1,370 feet, and our present
camp 1100 feet; thus the average fall of Sturt's Creek has been 270 feet
in 180 miles, or one and a half feet per mile. Now the distance to
Desault Bay (which appears the most probable outlet) is 370 miles, and
allowing an increase of 500 for deviations, there would be more than two
feet descent per mile, which would be sufficient for the maintenance of a
channel. Should the creek turn to the south and enter the sandy desert
country, the water would soon be absorbed, especially as the wet season
at the upper part of the creek occurs when the dry season is prevailing
in the lower part of its course. That it does lose itself in a barren
sandy country is, I fear, the most