Journal of the Royal Society of Western Australia, 87(3), September 2004 
been eroded. This can be construed as an early stage in 
the development of mogotes from a cracked surface (Fig 
11; 8 in Fig 1). The encrustation is, however, not iron-rich 
but mainly organic. Lichens, for instance, protect the 
surface but also they introduce the possibility that they 
may have caused expansion, for instance, sufficient to 
buckle pre-existing thin plates as their hyphae (or 
rootlets) penetrated between crystals exposed in the 
surface layer of rock. 
Subsurface initiation of tafoni and alveoles 
Though many familiar major and minor granite 
landforms are demonstrably of two-stage or etch type 
and are of subsurface origin (see e.g. Twidale 1962; 
Twidale & Bourne 1975b, 1976) the provenance of 
alveoles and tafoni (Fig 12A) has remained enigmatic. 
On the one hand, the enlargement of the hollows is 
clearly due to salt crystallisation or haloclasty (Evans 
1969; Winkler 1975; Bradley et al. 1978). Growth takes 
place from the base upwards in exposed blocks, 
boulders and sheet structures. Evidence of subsurface 
initiation is lacking, though the lateral merging of tafoni 
with flared slopes, as for example at Ayers Rock (Uluru) 
and at Kokerbin Hill (Twidale 1978; 1982a, p 257), 
suggests that some have been initiated within the 
regolith. 
This conclusion finds support on the northwestern 
spur of The Humps just below but associated with Level 
II. There, several large residual boulders display tafoni 
developed in rock that is rotten due to chemical 
weathering of the basal zones (Fig 12B). It is suggested 
that the bases of the boulders, resting on an impermeable 
rock-base - an intact sheet of rock - were altered as a 
result of contact with water. The feldspars have been 
altered to kaolin. Limonite is concentrated in the 
marginal zones or visors. This chemical alteration 
contrasts with the haloclasty which is a physical or 
mechanical process and which, exploiting the already 
weakened host rock, is, with gravity, responsible for the 
formation and enlargement of the hollows. 
Preservation of sheet structure and boulders on crests 
The smooth outlines of the bornhardt are at several 
sites interrupted by isolated blocks or boulders, or by 
groups of such residuals preserved on crests or upper 
slopes (Figs 2 & 7A). The reasons for the survival of these 
remnants of sheet structures (10 in Fig 1) are twofold. 
First, that the volume of runoff increases downslope, so 
that crestal zones are less weathered and eroded than 
downslope areas. Downslope wash removes the grus 
which otherwise would retain moisture and thus ensure 
the continuation of weathering and the eventual 
disintegration of the blocks and boulders. Second, the 
lower zones have been subjected to successive phases of 
subsurface scarp-foot weathering. Any remnant blocks 
and boulders on lower slopes either have survived 
because they are especially large and therefore persistent, 
or they have fallen to their present position. 
Large gnammas on crests 
In the sense of areal extent, and as on many other 
granite hills in the district, most of the largest gnammas, 
in the form of shallow, flat-floored pans (Twidale & 
Corbin 1963), are found on the crests of hills. Some on 
the crest of King Rocks, for instance, are more than 15 
diameter. As runoff increases downslope (see above), i{- 
might be thought that the largest gnammas ought to b^ 
developed on lower slopes. But there, heavy runoff 
ensures that the basins are filled to overflowing. Egress 
channels are eroded, so that lateral extension by solution, 
hydration and hydrolysis of the rock-forming minerals 
by standing water is limited. On hill crests, however, 
only direct rainfall which stands in basins is received. 
Moreover, the crests are the oldest, and longest exposed, 
parts of the outcrops; there has been time for pans to 
extend laterally by moisture-related weathering. 
Alcoves: armchair-shaped hollows and embayments 
Several armchair-shaped hollows are developed high 
on the midslope of the valley which drains the northern 
sector of The Humps (Figs 4, 7A & 8). One is as deep as it 
is long and wide (Fig 8; 3 in Fig 1), but they are typically 
much wider than they are in plan dimensions. 
How have they formed? Grus is washed into the floors 
of clefts (or slots) and valleys. There, the patches of 
detritus become vegetated and a soil develops. Thus a 
depression - an incipient alcove? - filled with enough 
soil to nurture small trees has formed in the floor of the 
cleft with Eucalyptus caesia. Vegetated soil patches retain 
moisture which aggressively react with the granite with 
which it is in contact by virtue of chemicals released by 
weathering and organic acids derived from rotting 
organisms. In addition, lichens favour wetter surfaces 
such as are found in valleys and even otherwise bare 
rock surfaces are attacked by the hyphae of lichens and 
by the roots of trees and shrubs. In these ways 
weathering causes vertical rotting and thickening of soil 
patches. They also effect lateral extension. Headward 
extension takes place causing steepening of the valley 
headwalls but lateral growth - what Mabbutt (1966) 
called mantle-controlled planation - takes place at both 
margins so that the depression widens twice as rapidly 
as it extends headwords. This produces alcoves or 
armchair-shaped hollows, some of them many tens of 
metres across. The side- and head-walls are commonly 
flared (as on Hyden Rock: Twidale & Bourne 2001) 
though the largest such alcove located on the western 
side of The Humps is bounded by slightly overhanging 
bluffs in which sheet fractures and structures are exposed 
(10 in Fig 1). These fractures attract weathering, so 
preventing any tendency for the development of a 
smooth convexity. 
Conclusion 
The Humps displays many morphological features in 
common with other Wheatbelt inselbergs: it is of etch 
origin and has survived by virtue of its massive 
structure, its crestal area is of great antiquity, it has been 
exposed in stages, and it displays many familiar minor 
forms including some due to contemporary seismicity. 
Yet it offers enough different forms and lines of 
evidence to make it interesting. A subsurface initiation 
for tafoni is suggested, the survival of the inselberg can 
be construed as tectonic, and the origin of several minor 
structural features can be traced back more than 2.5 
billion years to the origin of the rock of which the 
inselberg is composed. 
132 
