The writer thanks reviewers, one of whom was Professor Y. Battiau-Queney, for constructive suggestions, and Dr Jennie Bourne for her usual critical reviews. I also express my gratitude to Professor Battiau-Queney for kindly translating the abstract, abridged version, keywords and captions.
‘Embarrassing questions tend to remain unasked,
and if asked, to be answered rudely.’ (P. Medawar)
« Les questions embarrassantes ont tendance à ne pas être posées et, si on les pose, la réponse est brutale » (P. Medawar)
‘If at first an idea is not absurd, then there is no hope for it.’ (A. Einstein)
« Si à première vue une idée n’est pas absurde,
alors elle n’aura pas d’avenir » (A. Einstein)
1Acceptance of Huttonian continuous change (Hutton, 1795) and Davisian ubiquitous degradation (Davis, 1899) have together ensured that, with the exception of some exhumed forms, landscapes are regarded as youthful. Most are considered to be of later Cenozoic age. Thus, all the forms and assemblages discussed in a volume of papers dedicated to Long-termLandscape Development in Southeastern Spain (Mather and Stokes, 2003) date from the Neogene and most from the Late Tertiary or Quaternary. A similar temporal perspective characterises the volume on long-term landscape development edited by B.J. Smith et al. (1999), and though they note the possibility of older forms persisting in lowland Scotland, the other inherited features found in Britain and considered by P. Migon and A.S. Goudie (2001) were assigned to the Cenozoic.
2Such youthful landscapes dominate the continents. In Australia (fig. 1A), for instance, the vast interior plains of the Carpentaria and Lake Eyre basins were shaped by riverine erosion and deposition, though later, during the Quaternary, some were modified by the wind. The virtually featureless Nullarbor Plain is an etch surface developed on Miocene limestone, and a similar age has been attributed to the laterite that caps many of the tablelands of northern Australia. And so on – youthful land surfaces are widely distributed.
3But for almost a century, in Australia and elsewhere, there have been reports and claims of substantial landscape elements that are of much greater antiquity, and that, indeed, date from the Mesozoic (Twidale, 2007). They are not exhumed though such forms, which are of various ages, are recorded in the landscape and in the literature. Many are of subCretaceous date, but a few are as old as Archaean. The very old surfaces considered here are of etch type. Even duricrust-capped surfaces lack an A-horizon. Their character was concisely expressed by E.S. Hills when discussing the high plains of eastern Victoria: ‘While these high surfaces of low relief clearly represent preserved relics of old surfaces which have escaped deep dissection, they have naturally suffered some reduction and modification in detail during the long periods of time to which they have been exposed to weathering and erosion, but this is relatively minor.’ (Hills, 1975, p. 300).
4They are an order of magnitude older than is implied in conventional theory, and consequently, but not surprisingly they are viewed with suspicion, amounting to disbelief by many, perhaps most, geomorphologists. What, then, is the evidence on which the ancient landscapes concept is based, and what are its implications? Examples of very old landscape remnants and the evidence adduced for their ages are described and discussed. First, however, a summary is given of the geological and geomorphological characteristics of the Australian continent that are germane to the dating of events and landscapes, and to the conservation of landforms.
Fig. 1 – Location map, Australia (A). Location map, Gulfs region of South Australia (B).
Fig. 1 – Carte de localisation, Australie (A). Carte de localisation, région des Golfes de l’Australie méridionale (B).
5Structurally, Australia can be regarded as comprising the western shields and cratons, of various Precambrian ages but embracing Palaeozoic fold belts at the eastern margin; and together occupying the western half of the continent. The east coast is backed by the complex of Palaeozoic orogens that underlie the Eastern Highlands. The lowland between shields and highlands were marine embayments during the later Mesozoic and Miocene, and are now underlain by various sedimentary basins.
6Several factors have favoured the persistence and dating of surfaces in the ‘Island Continent’. Australia is located far from plate boundaries and though not atectonic is relatively stable. It has not experienced orogenesis since the Palaeozoic, though earth movements caused either by crustal convection (cf. Burke, 1996; Partridge, 1998) or isostatic adjustments have caused regional and local uplift, and minor but widespread, crustal joggling. Such relative tectonic quiescence has allowed planation to proceed unchecked for long periods but disturbed by relative uplift leading to stream rejuvenation and landscape revival. The resultant stepped landscapes are part of a low continent, with a mean altitude of only some 330 m, for about 40 % of the continent is less than 200 m above sea level, with a maximum elevation of 2227 m above sea level in Mt Kosciuszko and a low point of 15-16 m in Lake Eyre South.
7On one hand, prolonged weathering of the planation surfaces has produced several widely distributed duricrusts that have been dated stratigraphically and thus form regional morphostratigraphic markers (fig. 2). Various marine transgressions, but particularly those of the Cretaceous (Frakes, 1987), penetrated deep into the continent and provide further evidence of the age of associated planation surfaces. Valuable temporal benchmarks are provided also by dated volcanic rocks; first, at the southern edge of the continent and associated with the separation of Australia and Antarctica, and later, in the Eastern Highlands where hotspot volcanicity has occurred throughout the Cenozoic.
8On the other hand, the compact shape of the continent has limited the impact of Late Cenozoic glacio-eustatic sea-level changes. The mid- to low-latitude position of the continent reduced the direct impacts of Quaternary glaciation and periglacial activity, which were confined to the uplands of southeastern Australia. The Late Palaeozoic glaciation, however, proved significant for Australian landscape studies. All of southern and central Australia was covered by ice sheets during the Early Permian (Sakmarian) and with the exception of certain older exhumed forms, contemporary landscape evolution can be taken as beginning with the recession and melting of the ice and the exposure of the glaciated land surface (BMR Palaeogeographic Group, 1992).
Fig. 2 – Dissected silcrete-capped plateau near Innamincka, northeastern South Australia.
Fig. 2 – Dissection du plateau couvert de silcrète près d’Innamincka, nord-est de l’Australie Méridionale.
9Structurally, the Gulfs region of South Australia (fig. 1B) comprises the Gawler Craton (Flint, 1993) including the Stuart Shelf and the Adelaide Geosyncline (Preiss, 1987), on which are developed the Flinders and Mt Lofty ranges and Kangaroo Island.
10The massif consists of rows of bornhardts developed mainly on silicic volcanic rock (1592 Ma) but also, in the west, on intrusive granite (1585 Ma). The origin of the volcanic rocks remains a matter of debate (Blisset et al., 1993; Allen et al., 2003, 2008) but both volcanics and granites are subdivided by orthogonal fracture patterns. In addition, concentric arcuate fractures associated with the 600 Ma Acraman meteorite impact are prominent in the western part of the upland (Williams, 1994).
11Presumably, the columnar jointing that is widely developed and exposed in the volcanics developed shortly after their effusive emplacement. The fracture systems that define the bornhardts are intruded by the Gairdner Dyke Swarm, which is mainly doleritic and dates from 1050 Ma, showing that the fractures, which proved critical to later landscape development, were already in place at that time. Sheet fractures (‘offloading joints’) that are confined within the fracture-controlled major orthogonal and rhomboidal blocks transect the banks of columnar joints. They probably formed as a result of shearing some time after the orthogonal fracture system was in place (Twidale et al., 1996), but prior to the period of subsurface weathering that developed and defined the bornhardts beneath a weathered planation surface, the notional Beck Surface.
12Toward the end of the Mesozoic, during the separation of Australia and Antarctica, uplift of a complex en échelon fault zone at the southern margin of the massif caused it to be tilted down to the north. Rivers flowing north stripped most of the regolith, exposing the bedrock as a bornhardt massif of etch type (fig. 3 A and B). Only small patches of corestones set in a matrix of weathered country rock have so far been located, perched high in the local relief. The regolithic detritus, including cobbles of the distinctive volcanic rock, was deposited as the Mt Anna Sandstone in the Eromanga Basin where it is interbedded with fossiliferous marine strata of Neocomian-Aptian age (fig. 3C; Wopfner, 1969, p. 152-156).
13This, then, is also the age of the etched bornhardt landscape (Campbell and Twidale, 1991). This Nott Surface is derived from the degradation of the Beck Surface which predates the Neocomian-Aptian, but postdates the withdrawal of the ice sheet that covered most of southern and central Australia during the Sakmarian.
14Silcrete developed in valley floors and in areas bordering the massif, probably during the Eocene (Wopfner et al., 1974; Firman, 1983), has been dissected at the margins of the upland. Apart from Quaternary faulting that has caused blockage and reversal of the drainage in the catchment now occupied by the Lake Gairdner salina and recent gullying, the Early Cretaceous landscape remains virtually untouched.
15Thus, stratigraphic, geomorphological and topographic considerations together permit some old surfaces, like that preserved on the Gawler Ranges massif, to be dated with assurance. Moreover, not only the ancient surface but many of the events that imposed their characteristics on the host rocks and on the landscape can be identified and dated. A long sequential development or lineage can be reconstructed (Twidale and Vidal Romani, 1994).
16The Early Cretaceous etched summit bevel of the Gawler Ranges may be correlated with the dimpled crests of the higher inselbergs of northwestern Eyre Peninsula (Mt Wudinna, Ucontitchie Hill) and the high shoulder preserved on Carappee Hill, located in the same general area. Several of these residuals are stepped and imply episodic exposure and the development of increased relief amplitude, and not degradation or lowering, through time.
Fig. 3 – A: Bevelled bornhardts in dacite in the southern Gawler Ranges, northern Eyre Peninsula, South Australia, and seen from Spring Hill.
Fig. 3 – A : Dômes résiduels (bornhardts) de dacite tronqués par un aplanissement dans le sud des Gawler Ranges, au nord de la Péninsule d’Eyre, Australie Méridionale, vus de Spring Hill.
Fig. 3 – B: The summit bevel of the Gawler Ranges seen from the air (E.M. Campbell).
Fig. 3 – B : Le sommet aplani des Gawler Ranges vu du ciel (E.M. Campbell).
Fig. 3 – C: Mt Anna Sandstone with cobbles of Gawler Range Volcanics, exposed in the Eromanga Basin (H. Wopfner).
Fig. 3 – C : Grès du Mt Anna avec blocs et galets issus des formations volcaniques de la chaîne affleurant dans le Bassin d’Eromanga (H. Wopfner).
17The Early Cretaceous marine transgressions and associated sediments (Frakes, 1987) proved useful in the dating of the Nott Surface also provided crucial evidence concerning the age of the high plain remnants preserved on the Arcoona Plateau (fig. 4). The Plateau is developed on a sequence of Neoproterozoic strata that have their equivalents in the Geosyncline to the east. But they lapped on to the Stuart Shelf, a covered sector of the Gawler Craton, and remain essentially undisturbed. The sequence, which dips gently northward, includes several quartzites and the plateau is developed on the successively exposed quartzite caprocks. The plateau was dissected prior to the Early Cretaceous when the sea invaded valleys and basins, leaving behind fossiliferous deposits. Thus the plateau surface is older than the Early Cretaceous but postdates the Permian glaciation.
18Rounded cobbles and gravels testify to the erosion of the plateau by rivers graded to the Early Cretaceous seas. Some of the regolith may have been stripped from the plateau in areas adjacent to the marine embayments but beds of mudstone and shale survive above the caprock in many places: the residual mesas are what L.C. King (1968) called domed plateaux. That the quartzitic caprocks protected the old surface by retarding scarp recession is indicated by the intense scarp-foot weathering evidenced throughout the region, resulting in depressions, false cuestas, alveolar forms and the presence of silcrete of probable Eocene age (Wopfner et al., 1974) in valley floors and piedmonts. The present landscape dates from the Eocene but the plateau surface is older. It is of at least Early Cretaceous age, having been shaped by subsurface weathering prior to the Cretaceous.
19The Kakadu or Arnhem Land massif is similarly dated as pre Early Cretaceous for the massif stood as an island in the then seas (e.g., Needham, 1982; Frakes and Bolton, 1984).
Fig. 4 – Domed plateau with, in foreground, silcrete-capped dissected valley floor, southern Arcoona Plateau, South Australia.
Fig. 4 – Plateau en dôme, avec au premier plan le fond disséqué d’une vallée couronnée de silcrète, sud du Plateau d’Arcoona, Australie méridionale.
20Though the Flinders Ranges is a region of ridge and valley topography a summit surface preserved mainly on quartzite ridges and basins is prominent throughout the upland. It is physically contiguous with a surface exhumed from beneath a cover of Early Cretaceous (Neocomian) littoral marine strata of which the only upland remnant is preserved in Mt Babbage, in the extreme north of the upland (Woodard, 1955; Alley and Lemon, 1988). The regolith preserved there at the unconformity between the Cretaceous and the Precambrian is not encountered elsewhere, indicating that the summit surface is of etch type. It may have been shaped by rivers graded to the Early Cretaceous shoreline. This suggestion finds support in the demonstrated age of the present ridge–and–valley assemblage of the Flinders Ranges which resulted from the later Cretaceous uplift of the Ranges, rejuvenation of rivers, and differential erosion of the summit high plain.
21Possibly as a result of tectonic disruptions associated with the separation of Australia and Antarctica, a lake was impounded during the Middle Eocene in the northern part of the Willochra Plain, an intermontane basin located in the southern Flinders Ranges. Silicified lacustrine strata with contained plant fossils tongue up valleys between ridges adjacent to the present intermontane basin, showing that the ridge and valley topography predates the Middle Eocene. Thus the bevelled ridge crests are older and of putative Cretaceous age (fig. 5).
22Hints of an even older planation surface are provided by sedimentological evidence from terrestrial Triassic basins preserved within the fold mountain belt. The strata in the basins fine upwards. It has been suggested that this may indicate a progressive decrease of relief amplitude, culminating in a planation surface of low relief in the areas that source the basins (Johnson, 1961), but no relics of such a surface have been identified in the field.
Fig. 5 – The Battery, showing a bevel cut across a shallow basin in quartzite, southern Flinders Ranges, South Australia (Department of Lands, South Australia).
Fig. 5 – La « Battery », replat tronquant les quartzites d’un bassin peu profond, dans le sud des Flinder Ranges, Australie méridionale (Department of Lands, Australie méridionale).
23The earliest suggestion of the existence of very old landscapes in Australia is due to P.S. Hossfeld (1926), who adduced evidence pointing to a Cretaceous age for the lateritised summit surface of the Mt Lofty Ranges (fig. 6). The upland is a horst developed on folded and faulted strata of Proterozoic and Cambrian age which in the east and south are intruded by granite of Early Palaeozoic age. It is bordered to the east by the Miocene marine embayment that is now the Murray Basin. Cenozoic sequences that include basal beds of Middle Eocene age were deposited in fault angle valleys or half-grabens developed on the western flank of the horst (Miles, 1952; Reynolds, 1953; Campana, 1958). Lateritic detritus occurs in the basal beds (Glaessner and Wade, 1958). Also, by the Eocene the River Torrens had incised its gorge and deposited coarse debris of that age where it debouched from the upland before flowing to the sea via the present site of Adelaide. Thus various lines of evidence point to the horst and the lateritic high plain preserved on it predating the (Middle) Eocene.
24However, on Kangaroo Island, a structural extension of the Mt Lofty Ranges horst, a longer perspective is provided by basalt of Middle Jurassic age extruded during the parting of Australia and Antarctica. It overlies the laterite that caps the plateau that forms the backbone of the Island and which is developed on strata of Proterozoic, Cambrian and Permian ages. The laterite therefore predates the volcanic event but is younger than Permian (Daily et al., 1974), as are the dissected lateritised high plains preserved on the adjacent Fleurieu Peninsula (the southern part of the Mt Lofty Ranges) and the Lincoln Uplands of southern Eyre Peninsula. Stratigraphic considerations point to its being Triassic in age.
25But the discomfort with extreme antiquity is exemplified by the contrasted analyses of the Kangaroo Island landscape. The first systematic geological mapping established that the (Wisanger) basalt overlies the laterite, that it postdates the Permian (Sprigg et al., 1954). This was accepted without demur and confirmed by later workers (Daily et al., 1974). Admittedly the laterite profile concerned lacks the massive pisolitic ferruginous zone of classical occurrences and found at many sites in the Gulfs region, including parts of Kangaroo Island. Yet it is typical of many exposures in the Mt Lofty Ranges with the mottled kaolinised horizon overlain by a ferruginous concentration rather than a distinct lithological horizon. Meantime, however, the basalt had been numerically dated as Middle Jurassic (McDougall and Wellman, 1976), by contrast with the Quaternary age, based on long distance extrapolation, suggested by Sprigg’s group. Consequently the laterite had to be much older that the initial Pliocene age assigned to it. The stratigraphic relationship between basalt and laterite remains the same, the physically-determined age of the basalt has been confirmed, but the age implications flowing from the physical dating of the volcanic rock are strenuously denied. Even scientists shy away from things they do not understand, whereas it is such unknowns that ought to attract attention.
26Notwithstanding this debate, the lateritic regolith is well preserved on the southern sector of the Mt Lofty Ranges, but to the north most of it has been stripped to expose the erstwhile weathering front as an etch surface from which rise lateritised remnants. Regolithic detritus was deposited in marginal basins of Eocene age located to the north and west of the upland.
27Volcanic flows were used also to date the high plains of eastern Victoria. They are located on divides between valleys floored by stratigraphically dated lavas, some of Eocene age. Thus, the palaeosurface remnants are at least of Cretaceous age (Hills, 1934, 1938), though later work suggests that some of the higher remnants may predate the Cretaceous (Jenkin, 1988). Hills’ use of valley lava flows proved critical also in demonstrating the great age of the previously identified high plains and plateaux preserved in the Shoalhaven and adjacent areas of southeastern New South Wales (e.g., Craft, 1932).
Fig. 6 – View across the Torrens Gorge of the summit plain of the Mt Lofty Ranges, South Australia (R.L. Oliver).
Fig. 6 – Vue sur la Gorge de Torrens dans la plaine sommitale des Mt Lofty Ranges, Australie méridionale (R.L. Oliver)
28The lateritised Old Plateau of the southwest of Western Australia (fig. 7 ; Jutson, 1914, 1934) is developed on the Yilgarn Craton, a complex of granite, gneiss and ‘greenstone’ (basic volcanics) with ages in the range 2900-2500 my, except in the south where the range is 1300-1100 my. The Plateau is bordered on the west by the Darling Fault and Scarp. The latter is dissected and Early Cretaceous strata are preserved in some of the valleys but not on the Plateau (Playford et al., 1976), suggesting that the Old Plateau predates the Early Cretaceous (Twidale and Bourne, 1998). The dissection of the Old Plateau to produce the New commenced in the Eocene and the incised streams continue to extend their valleys at the expense of the lateritised surface (Clarke, 1994; Salama, 1997).
Fig. 7 – Remnant of lateritised Old Plateau, preserved near Mt Magnet, central Yilgarn Craton, Western Australia. Where the laterite has been stripped, the granitic bedrock is exposed in the so-called New Plateau.
Fig. 7 – Reste du Vieux Plateau latéritique conservé près du Mt Magnet, au centre du Craton de Yilgarn, Australie occidentale. Là où la latérite a été érodée, le substrat granitique affleure sur le dénommé Nouveau Plateau.
29Regional stratigraphy suggests that the prominent weathered high plains surface of the Hamersley Ranges developed on flat-lying, though in places faulted and warped Precambrian Banded Iron Formations, is of later Cretaceous age (MacLeod, 1966; Hocking et al., 1987). Shallow pockets of a ferruginous pisolite are preserved on the summit surface. They suggest first that there was once a cover of such material, but, and second, that most of it was stripped by rivers radiating from the surface and deposited (as the economically significant Robe River Pisolite) in their valleys. There it hardened on desiccation (cf. Alexander and Cady, 1962) and thus protected the valley floors. At several sites the Pisolite overlies fossiliferous alluvia of Middle Eocene age. Erosion of the adjacent slopes caused topographic inversion, with the old valley floors left standing in local relief as spectacular sinuous mesas (fig. 8). Thus, the Hamersley Surface is mainly of etch type and of Middle Eocene age, but the remnants capped with the Pisolite imply the earlier, Cretaceous, existence of a planation surface on which the regolith developed.
Fig. 8 – The Hamersley Ranges in the northwest of Western Australia, with Mt Wall on the left skyline. The sinuous mesas in the foreground are former river valleys now capped by pisolitic iron oxide derived from the plateau beyond.
Fig. 8 – Les Hamersley Ranges dans le nord-ouest de l’Australie occidentale, avec le Mt Wall à gauche sur la ligne d’horizon. Au premier plan, les mesas ondulées sont d’anciennes vallées fluviales maintenant couronnées d’oxyde de fer pisolitique, provenant du plateau situé en arrière.
30Stepped landscapes with plains at multiple levels are preserved in several parts of Australia. They are most obvious in regions like the Yilgarn Craton where the Old and New plateaux are prominent and distinct, but are discernible elsewhere also. Such very old elements pose problems for the conventional models of landscape development based in Huttonian and Davisian tenets, for they ought to have been degraded and worn away by their long exposure to weathering and erosion. So, first, are the data susceptible of alternative explanation?
31Marine planation, favoured by some early workers such as A.C. Ramsay (1846) and S.W. Cushing (1913), can be ruled out both on general grounds (e.g., Baulig, 1952; King, 1963) and for want of evidence: there is nothing to suggest that the sea ever covered or seriously encroached upon the Yilgarn and Gawler cratons. Early Cretaceous seas invaded the already dissected Arcoona Plateau but as has been demonstrated, associated erosion by marine agencies and streams graded to the shorelines was limited. C. Fenner (1930) supposed that the then Mt Lofty Ranges were inundated by the sea during the Miocene, but though invaded basins and embayments were invaded, associated shorelines stand relatively low in the relief: as stated by M.F. Glaessner and M. Wade (1958, p. 124) : ‘The Mt Lofty Ranges were not covered by the sea’.
32Early workers in arid lands (Passarge, 1904; Jutson, 1914) emphasised the role of wind (deflation, corrasion) in regional and local planation. Jutson, for instance, attributed the widening of stream-dissected valleys to planation by the wind (though by 1934 he had accepted the primacy of river work). Working in tropical Africa, Passarge concluded that in his Banda landscapes the wind had lowered the land surface down to the level of the regional water table in a mechanism that was described as ‘leveling without baseleveling’ (e.g., Davis, 1905), for the supposed aeolian surface was developed independently of sea level. However, further reflection and later work suggests that though major deflation hollows have developed in especially propitious conditions (Breed et al., 1989) and deflation must contribute to the lowering of arid plains, erosional forms of aeolian derivation are minor and limited (e.g., Russell, 1932 ; Said, 1962). Even in the deserts water-related weathering is significant and most erosional landforms are caused by running water generated by episodic but effective rains of the present regime or inherited from past humid phases (e.g., Peel, 1941). Regional planation is the result of weathering and the work of rivers.
33Another possible explanation for summit surfaces, such as that preserved on Fleurieu Peninsula where the age of the surface is derived in part from the age of deposits in marginal basins, involves the possibility of high plains being formed above the ultimate baselevel but in accordance with local or regional datums. It has been assumed that the landscapes discussed in earlier pages have been shaped in relation to the then sea level or ultimate baselevel. They were then relatively uplifted, dissected and in most instances stripped of regolithic cover. But could there have been planation high in the relief? Taking an overall view, most of the high plains discussed here transect various structures and lithologies. They are erosional forms developed in materials unlikely to act as regional baselevels.
34But perched baselevels could develop in areas of flat-lying or only gently dipping strata. In the Arcoona Plateau, for example, could the various flat-lying resistant strata have acted as regional baselevels, and a composite high plain have been formed, say, during the Cenozoic, following the withdrawal of the Early Cretaceous seas that inundated the adjacent lowlands in the later Cretaceous ? In that particular region the plateaux are underlain by one of several gently dipping Neoproterozoic quartzites, which certainly protect and maintain the escarpments. They also restricted the extension of the valleys and basins - witness the Eocene silcrete of many piedmonts - but they did not serve as baselevels of erosion, for argillites are preserved above them on the domed plateaux, which is incomprehensible if the various caprocks had formed local baselevels.
35Assuming for the sake of discussion that the forms and surfaces cited are of the ages suggested, the question arises as to how have these alleged old surfaces survived the elements?
36Several factors can be cited but the crucial argument concerns water, which is of prime significance in weathering and erosion. Its effects cannot be overestimated. Because of its molecular structure, water is the supreme solvent. Rock rich in minerals such as quartz that are chemically of low reactivity in the conditions that commonly obtain at and near the land surface obviously are stable. In addition, groundwaters are rich not only in chemicals and organisms but also in the products of organic decay so that regolithic waters are aggressive. Accordingly, any factor that prevents a landscape element being or remaining in contact with water is protective. Thus a quartzite ridge is likely to persist, as are etch surfaces devoid of regolith. As J.R. Logan (1851, p. 326) pointed out : ‘The soil is always kept moist…’ and as a result, whereas on exposed bare rock surfaces decomposition is arrested or retarded, ‘Under-ground decomposition tends to spread unchecked on all sides.’ Protection may accrue from negative factors, namely lack of exposure to moisture (also Twidale and Bourne, 2000).
37Large rivers have exploited structural weaknesses (e.g., the Zambezi below the Victoria Falls), have achieved notable erosion in brief catastrophic flows (e.g., Kendall, 1902; Baker, 1973) and over long periods (e.g., Lucchitta, 1990). Apart from such big rivers, however, it can be argued that running water effects little erosion where in contact with solid rock such as is exposed in etch surfaces.
38Gullies are widely developed but everywhere are eroded mainly in regolithic material. Incision has virtually ceased where the stream encounters lithified rock, however weak (e.g., mudstone). Similarly, describing the landforms of J.T. Jutson’s (1914) New Plateau exposed in the Pilbara region of the northwest of Western Australia, A.J. Noldhart and J.D. Wyatt (1962, p. 46) record that the plain surface is essentially coincident with the junction between the lateritised regolith and the fresh granite basement. The granitic etch surface has undergone little vertical incision since its exposure sometime during the later Cenozoic. J.T. Jutson (1914, 1934) made the same point, though in a general or regional - rather than site - context. As summarised by E.S. Hills (1962, p. 10), it was concluded that: ‘… if a marked increase in erosion is induced on a soil-covered terrain developed on rock, there is first of all a period of rapid skimming off of the soil and rotten rock, followed by a sudden decrease in removal of detritus when fresh rock is exposed to the eroding agents.’
39Clearly, stream behaviour is a significant factor so that location vis-à-vis zones of crustal disturbance, of whatever origin, also assumes relevance, for uplift induces stream rejuvenation. Similarly, lowering of sea level (baselevel) initiates a phase of stream rejuvenation and landscape revival that first impacts near the coast and works inland, albeit slowly (Taylor et al., 1985 ; Young and McDougall, 1993). Thus, continents such as Australia, with a compact shape and low coastline-length/area ratio, are less vulnerable than, say, Europe with its long indented coastline relative to its total area (1 km of coast for every 390 km2 in Australia; 1 km for every 1.75 km2 in Europe).
40The backwearing or scarp retreat mechanism, was recognised by such earlier workers as O. Fisher (1866), J.T. Jutson (1914), and A. Holmes (1918), but was championed by Lester King as the basis of a model of landscape evolution (e.g., King, 1942, 1953 ; see also Fair, 1947, 1948). Recession is localised, with escarpments deeply breached only by major rivers. Nevertheless, backwearing involving pronounced basal slope weathering and erosion, is probably the dominant mechanism in landscape evolution viewed globally. However, because of the contrast in the distribution of soil and vegetation and hence near-surface moisture as between humid and arid lands, such basal attack is most pronounced in aridity (Penck, 1924; Jessen, 1938; Twidale and Milnes, 1983). It implies the longer survival of the ‘original’ or pre-incision surface than does a regime based in lowering or steady-state development.
41Obviously some rocks, like quartzite, are resistant and many materials and especially the well-named duricrusts become resistant when dry. But structural advantage also promotes survival by setting in train the concatenation sequence in which the factors mooted above are combined.
42Structures favourable to differential weathering and erosion result in unequal activity, that in turn introduces reinforcement mechanisms (Behrmann, 1919; Bliss Knopf, 1924; Crickmay, 1932; King, 1970; Crickmay, 1976). The operation of this sequence has ensured that some areas have been lowered, but by the same token, and as a corollary, that adjacent areas have escaped such activity and persist, albeit stripped of regolith, standing high, dry and essentially intact as etched palaeosurfaces.
43Reinforcement can be illustrated at a number of scales but can be illustrated by the development of Mt Wudinna, a bornhardt inselberg located on northwestern Eyre Peninsula. It became upstanding because it is a massive compartment of rock. It sheds rainfall and runoff. It is relatively dry, whereas the adjacent plains receive not only direct precipitation but also runoff from the upland. The rock around the hill is weathered (and may decrease in volume, resulting in compaction and surface subsidence.Alternations of scarp-foot weathering and erosion resulted in the episodic exposure of the residual, with an increase in relative relief over time (Twidale and Bourne, 1975; Bourne and Twidale, 2000).
44Having in mind the cover of soil and vegetation developed over much of the continents, it has been claimed that never before has the Earth been so well armoured against the processes of weathering and erosion (Russell, 1958). The significance of woodland can be gauged by the accelerated soil erosion induced by its removal or depletion but in addition ground cover, and particularly the grasses (Graminaceae), which developed and extended during the later Cenozoic, and from the Oligocene onwards. They protected the bedrock against raindrop impact and run-off. Dense root systems also afford protection against erosion for they bind the soil (for an extreme example: see Egan, 2006).
45Organic evolution can be traced back to micro-organisms that existed some three billion years ago, since which time, though interrupted by extinction phases like those of the Permian and end-Cretaceous, there has been a proliferation of plants and animals. Is it not possible that just as the spread of humans has instigated an epicycle of accelerated soil erosion, so the spread of bacteria, nannobacteria and other micro-organisms (e.g., Trudinger and Swaine, 1979; McFarlane and Heydeman, 1985; Folk, 1994) and of chemicals produced by particular plants (e.g., Bloomfield, 1957; Hingston, 1962) has increased the rate at which rocks are weathered, and hence more rapidly rendered them susceptible to erosion?
46If so, it can be suggested that older landforms and landscapes like The Humps, a massive granitic inselberg located just north of Hyden, in the southern Yilgarn Craton of Western Australia (Twidale and Bourne, 2004) that is a refuge for Gondwanan plant species (Hopper and Burgman, 1983; Hopper et al., 1996; Fay et al., 2001) were not subjected to this hazard, because they already stood high and dry in the landscape and thus were protected against the brunt of attack by water charged with chemicals and biota. Old high landscape remnants may have avoided the biochemical assault. Thus, it is possible that reinforcement at grand spatial and temporal scales has contributed to the preservation of very old forms in a way that has not benefited later features to the same extent.
47Evidence pointing to the great age of some landscape elements has been recorded over the past 80 years, and particularly during the last half century. Here, evidence has been cited pointing to etch landscapes of Early Cretaceous ages. There are hints of even older Triassic landscapes in several of the regions analysed. Alternative modes of development have been cited and conservative factors suggested but it must be conceded that even if all the suggested conservative mechanisms have operated in concert, they reduce rather than resolve the problems posed by the very old palaeosurfaces like those discussed here and reported from various parts of the world. However, even if the great age of some of the landscapes cited remains in question or proves to be in error, so long as others withstand critical examination and are accepted as very old landscape elements, questions of ‘how’ and ’why’ remain to challenge basic geomorphological tenets and models.
48Even so, having entered this rider, it must be recognised that, and understandably, there were and are many who cannot accept that landscape elements such as those cited above can have withstood the forces of weathering and erosion for the periods implied. Some have been influenced by a conscious or subconscious adherence to conventional theory embracing virtually ubiquitous continuous change wrought by the agents of denudation. The announcement of ‘the likely existence of pre Pleistocene landscapes in Australia’ (Bierman and Turner, 1995) reflects the same mind-set, the same instinctive and conventional feeling that the Earth’s surface is, with the exception of some exhumed forms, youthful.
49Others who harbour doubts about the ancient landscape concept appear to do so on the basis of intuition, and though Thomas Aquinas remarked that intuition was given to angels, while the common man has to be content with reason, this is not to be ignored for, based in experience, it could signal that something is awry. R.P. Bourman (1989, p. 58, 1995, p. 5), for example, considers that the long-term survival of surfaces and forms is ‘not reasonable’. It has been suggested that if a thing exists it must be possible but, to the contrary, if it exists it calls for explanation.
50Some investigators have presented specific arguments. E.H. Brown (1980, p. 12), for instance, after asserting that: ‘any land surfaces of Cretaceous age and older are extremely unlikely to form part of the present surface’, suggested that rates of chemical weathering are such that any old landscape elements must have been eliminated. This is plausible but can be countered, for whereas some areas have been weathered and eroded as Brown suggested, all parts of the landscape have not been subject to equal activity. Some have been weathered and eroded, but others have remained relatively dry and have survived. Stepped landscapes in general and stepped inselbergs in particular provide plausible evidence and argument suggestive of contrasted rates of degradation in immediately adjacent areas.
51Unfortunately, so strong is the feeling against the ancient landscapes concept that some in the Australian geoscience community, evidently have characterised it as a collective ‘con job’, perpetrated by its proponents who, presumably collectively, and over a period of almost a century, have succeeded in deceiving their ‘unquestioning colleagues’ (M. Sandiford, personal communication, 2006). Such allegations, lacking substantiation and informed argument, but nonetheless condoned, are not conducive to the professional debate, infused with an informed scepticism, that the idea merits. The concept is unconventional but its implications for general theory are considerable and most advances arise from non-conformity.
52The Victorian author, A.W. Kinglake, stated he would like to see posted over the doorway of every church in the land the words ‘Important if true’; and the same sentiment applies to the ancient landscapes concept.