Page:Northmost Australia volume 2.djvu/297

632 Our course was about due west, and after climbing some low granite ridges we found ourselves on the western fall. [SEE MAP F.] Before we reached the COLEMAN RIVER, the granite country changed to schistose. This river had a fairly wide sandy bed, with very low banks, and water-holes far between and small. After running it down for about 6 miles, about south, we followed it down on a westerly course to a point about 30 miles from the Gulf (approx. 142 8' E.). The river was well denned until we reached this point, but here it began to get very small, the flood waters evidently flowing in all directions over the flat country. (The same thing happens with the other Gulf rivers, the Lukin, Edward, Alice, Mitchell, Staten and Red.) The country bordering the river all the way down was poor, the undulating schistose formation changing to Desert Sandstone and very flat sandy country, chiefly timbered with teatree.

"Nothing particular happened while we ran the river down, except an occasional brush with the blacks. The BLACKS on the Gulf Coast make very poor provision against the wet season. They either employ bent sheets of messmate bark or arch some sheets of teatree bark over a frame formed by a few bent sticks. Their practice in this respect contrasts strongly with that of the Princess Charlotte Bay natives, who sleep in well-constructed bee-hive gunyahs, designed to protect them from mosquitoes. For this latter purpose the Gulf natives construct a fragile platform, or often two platforms, of sticks forming a sort of two-story sleeping-place without walls. The lower platform is occupied by the gin, whose duty it is to keep up the smoke-generating smouldering fire. The man reposes on the upper platform, reaping the benefit of the smoke but taking no part in the work.

" Leaving the COLEMAN, we turned NORTH, passing over the same kind of flat sandy teatree country until within a mile and a half of the Edward River, when the country changed to grassy open box flats, with very hard grey soil.

"The EDWARD RIVER, which I named after my brother Dr. Edward Embley, of Melbourne, was, where we cut it, about 40 yards wide, with low banks, and contained some fairly good rocky water-holes, the rock being Desert Sandstone. It runs from east to west and enters the Gulf, after splitting up, in the low flat sandy teatree country, into several channels, in about 14 45' latitude. We followed it down to the sea and returned to where we first struck it. The box flats border the river on both sides for a considerable distance, but before reaching the head of the river the country becomes poor. The head (approx. 142 30' E.) is on flat country, in a pebbly ironstone-conglomerate formation with very little grass and timbered with quinine bush and teatree.

" On our way up the river from its mouth, we came across an ironwood tree, marked J (an old mark), on the bank of a small channel. This may have been one of the Jardines' midday camps.

" It is not uncommon in this region to see a group of small tree-stumps with the roots uppermost, stuck in the middle of a swamp and forming a sort of platform. In the neighbourhood of the Edward River we found the explanation of this singular arrangement. The roots are used for the support of corpses. On one occasion we saw three, each done up in a well-tied covering of folds of teatree bark. Presumably the idea is to protect the relics from dingoes, ants, etc.

" Continuing north from where we had first struck the Edward, we passed first across box flats and then over a stretch of sandy country until we came within a mile of the LUKIN RIVER, when grassy box flats were again met with. In the Lukin we