oven to calcareous sands. In the railway cutting, on the north side 
of the Swan, at Upper Swan, they show signs of contemporaneous 
erosion. Lithologically they have been thoroughly described by 
Hardman and Nioolay (1, 2). Simpson refers to them as being 
of Mesozoic age (7). They afford no palaeontological evidence. 
Shells are found in the bore sections of the Perth District, but are 
absent t'rom the Guildford bores (~>). They are best considered as a 
piedmont deposit, post-dating tbe formation of the Darling escarp- 
ment. 
Drift Sands . — The Western bank of tbe Swan, between Guild- 
ford and Upper Swam is bordered at a varying distance by dunes 
which have been fixed by vegetation. The dunes belong to the coastal 
dune belt, but are here siliceous rather than calcareous. They re- 
present fhe earlier accumulations during the period of coastal pro- 
gradation, which preceded tbe cycle of drowning referred to pre- 
viously (10), All the rivers of the Swan Goastal Plain appear to 
be consequent upon the Darling escarpment, but the encroachment 
of the dunes has had a. powerful effect in dellecting the streams 
on the plain (10), and the Swan has been forced, in the Guildford 
District, to take a subsequent course over the Guildford beds, until, 
by gathering strength from the Helena and the Canning, it has 
attained supremacy over the dunes. The surface of the Guildford 
beds slopes gently to the south, but the western gradient, on tbe 
western side of tbe Swan, is sufficient to cause stagnation of tbe 
drainage from the high dunes, with the result that a number of 
swamps have formed between tbe dunes and the river. 
Thu Tkrkacks axt> Lkvki.s. In order to investigate the terraces 
four lines of levels were determined by theodolite, crossing the valleys 
of the Swan and Helena (the latter in two cases only), approximately 
al right angles. From the levels obtained, profiles were plotted on 
a uniform scale, and these, combined with an examination of the 
ground, have led us to our conclusions. Four physiographic levels 
have been clearly revealed, and a fifth is believed to exist, but is not 
so well marked nor so well preserved as tbe remainder. The upper- 
most level is that of the surface of the Guildford beds, the lower 
levels being indicated by river terraces. As the terraces are de- 
veloped symmetrically in both rivers, and correspond very closely, 
they are of the type usually ascribed to uplift (11). 
The uppermost, or Caversham level, is tbe original erosion sur- 
face of the district. It has an elevation of 44 feet above survey 
datum at Guildford and Midland Junction, and slopes very evenly 
upwards to the north, being 70 feet above survey datum at Upper 
Swam In this level the Swan and Helena eroded their original 
paths over the Guildford beds, and developed broad and preco- 
ciously mature valleys until the close of the cycle of drowning. The 
thalweg so formed constituted the Guildford level, which is now 
