Twidale Caves Granite

Embed Size (px)

Citation preview

  • 8/11/2019 Twidale Caves Granite

    1/23

    Cadernos Lab. Xeolxico de LaxeCorua. 2008. Vol. 33, pp. 35 - 57

    ISSN: 0213 - 4497

    Caves in granitic rocks:

    types, terminology and origins

    TWIDALE, C. R.

    1

    and BOURNE, J. A.

    1

    (1) School of Earth and Environmental Sciences, Geology and Geophysics, University of Adelaide,Adelaide 5005, South Australia Tel.: +618 8303 5392; fax: +618 83034347; E-mail address:[email protected]

    Recibido: 1/11/2007 Revisado: 4/3/2008 Aceptado: 20/6/2008

    Abstract

    Caves or openings of various shapes and sizes are well and widely developed in granitic rocksas well as in other lithological environments. Some are caused by preferential water-relatedweathering, e.g. hydration, others to sapping, but haloclasty plays a crucial role in the deve-lopment of tafoni. These are especially well represented in granitic exposures. This can beexplained partly because the inherent strength of the crystalline rock permits hollowed blocks,boulders and sheet structures to remain standing. The hollows themselves owe their origin

    partly to the susceptibility of feldspar and mica to hydration and other forms of water-relatedalteration, and also to the capacity of haloclasty to rupture and break down the rock. On theother hand, dry granite is relatively stable and, particularly if it is cemented by salts concen-trated at and near the surface by lichens and mosses, it forms the crusts or enclosing visors thatare an essential part of tafoni morphology.

    Key words: cave, niche, shelter, alveole, tafone, hydration, haloclasty

  • 8/11/2019 Twidale Caves Granite

    2/23

    INTRODUCTION

    To the layman a cave is an undergroundhollow with access from the ground surface orfrom the sea (HANKS, 1986, p. 253), a state-ment that is almost identical to the formal ortechnical definition of a cave as a naturalunderground open space, usually with anopening to the surface (BATES and JACK-SON, 1987, p. 105; unless otherwise stated,all definitions cited in this paper are takenfrom this source). A cavern is a large cave or acomplex of caves. In general parlance, howev-

    er, any natural shelter or overhang also isreferred to as a cave.

    Caves are of various shapes, sizes, and ori-gins.

    HOLLOWS ASSOCIATED WITH STEEP

    SLOPES: NICHE, SHELTER, SLOT,

    ALCOVE, NOTCH

    A niche is a shallow cave, recess or re-

    entrant produced by weathering and erosionnear the base of a rock face or bluff. A shel-ter is a long and deep niche or coalescenceof niches. The term alcove used both in itsgeologic and general senses denotes a deepniche formed in a precipitous bluff or wall.Notch used in a coastal context is compara-ble to shelter but in its broader sense theword can denote a small alcove or, and morecommonly, a narrow passageway or slot.

    Niches and shelters are characteristic offaceted slopes, and are typically locatedwhere a permeable caprock or regolith is incontact with an impermeable lower forma-tion. Rarely, where the bluff has regressedand simultaneously migrated upslope andbeen reduced in height as the debris slopeextended upwards, the remains of formershelters with indurated walls and ceilings arepreserved.

    In granitic terrains niches and shelters area fairly common occurrence. Some are attrib-utable to weathering along sheet fractures to

    produce sheet tafoni (see below) but othersare spaces of roughly triangular cross-sectionleft vacant by the weathering and dislocationof wedges of rock at the exposed ends ofsheet structures and generated by shearingalong sheet fractures (TWIDALE et al., 1996;figure 1a). Others are caused by the weather-ing, near the present or former ground level,of obliquely intersecting cross joints or ofweaker rocks (figure 1b). In addition they areespecially common where the fresh graniteunderlies either an indurated veneer (figure1c) or a regolithic duricrust capping such as

    laterite or silcrete (figure 1d). Thus onplateaux near Cue and at The Granites, nearMt Magnet, both in the central Yilgarn Cratonof Western Australia, shelters are frequentlyformed at the base of the bluff (figures 1e and1f). They are developed in kaolinitic mottledand pallid zones of the laterite. The locationof such shelters varies according to the geom-etry of the faceted slope for they occur atbluff and debris slope, the extent of which

    varies spatially and in the long term temporal-ly. Shelters at plain level are referred to ascliff-foot caves.

    Niches and shelters are most commonlydeveloped in sedimentary terrains as a resultof seepage at the base of a bluff, at the inter-face of the permeable rock exposed in the rockface and the impermeable detritus accumulat-ed in the so-called debris slope (so-called,because by contrast with talus or scree slopes,

    the debris from which it takes its name is mostcommonly discontinuous and/or only a fewcentimetres thick).

    Though developed in Miocene MannumLimestone, rather than granite, bluffs bound-ing the Murray Gorge in South Australia dis-play every gradation between slopes consist-ing wholly of a bluff, to a graded slope com-prising upper convexity and lower concavedebris slope, distributed according to positionwith respect to the meandering river (TATE,1884; TWIDALE, 2000). They provide evi-dence of how shelters are developed.

    CAD. LAB. XEOL. LAXE 33 (2008)36 Twidale and Bourne

  • 8/11/2019 Twidale Caves Granite

    3/23

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 37

    Figure 1. (a) Void left by thedislocation of a triangularwedge of rock, eastern flankof Ucontitchie Hill, EyrePeninsula, South Australia.Scale provided by the lateGeorge Sved. (b) Shelter dueto ground level weathering,Middle Tor, Dartmoor, south-western England. (c) Shelterhigh on slope developed

    beneath indurated crust, but

    in massive granite, KokerbinHill, southwestern YilgarnCraton, Western Australia.

    (a)

    (b)

    (c)

  • 8/11/2019 Twidale Caves Granite

    4/23

    CAD. LAB. XEOL. LAXE 33 (2008)38 Twidale and Bourne

    (d)

    (e)

    (f)

    Figure 1. (d) Shelter inkaolinised zone developed

    beneath local silcrete (possi-bly a stream deposit in pre-dominantly lateritic carapace),The Breakaway, betweenHyden and Norseman, YilgarnCraton. (e) Shelters developedhigh on the slope at the base of

    bluffs in weathered (lateri-tised) granite near Cue, north-ern Yilgarn Craton, and (f)

    near plain level at TheGranites, Mt Magnet, centralYilgarn Craton.

  • 8/11/2019 Twidale Caves Granite

    5/23

    The Limestone is a calcarenite that con-sists of some 65% CaCO3 and is more suscep-tible to water attack than is granite. The outershell of the limestone is cavernous and is rid-dled with tubes and hollows. Where debrisslopes are developed the junction betweenbedrock and detritus is marked by numerousniches which in the past have merged to formshelters. From about 3000 years ago and up toperhaps the end of the Nineteenth Century theshelters were occasionally occupied byIndigenous/Aboriginal Australians (e.g.HALE and TINDALE, 1930; MULVANEY et

    al., 1964). Archaeological excavations haverevealed steepened bedrock slopes the risersof which display niches comparable to thosedeveloped at the back of modern shelters. Theshelters were occupied periodically and thestepped morphology of the buried bedrockslope is related to the rates of aggradation ofthe shelter floor. This varies according towhether the particular shelter was deserted,when natural relatively slow accumulation of

    debris derived from particles and fragmentsfalling from the shelter walls and ceilingallowed time for backwall niches to developand steps to form. When the site was occu-pied, however, natural accumulation was aug-mented by the ash from fires, shells and otherdebris (the midden), the rate of accretionincreased and bedrock risers developed(TWIDALE, 1964). Given that these eventstook place in a time span of hundreds of years

    it is clear that the backwall niches, presentlydeveloping and evidenced on risers in theexcavation, develop quite rapidly consideredin geological terms.

    Shelters are optimally developed in lime-stone terrains in the humid tropics. Here someswamp slots or cliff-foot caves extend manymetres beneath the walls of karst towers andcupolas and extend also to depths of severalmetres. They are critical to the development oftowers from domical hills (NEWSON, cited inSCRIVENOR, 1928, p. 189; PATON, 1964;JENNINGS, 1976; TWIDALE, 2006).

    Thus niches are caused by weathering andflushing of groundwaters at the permeable-impermeable interface. They increase in size

    and coalesce to form small shelters. Thereafterthe elongate hollows are enlarged by the gran-ular disintegration of the walls and ceilings as aresult of water seeping through the permeablecountry rock and dissolving the cement thatbinds the calcite and silica fragments. Similarprocesses are responsible at suitable sites ingranitic terrains, only at a slower rate, reflect-ing the greater stability of mica, feldspar andquartz as compared with calcium carbonate.

    ALVEOLES (OR HONEYCOMB WEA-

    THERING)

    Alveole is preferred to the term honey-comb because whereas the latter implies geo-metric regularity and depth, the former have arandom plan distribution and involves shallowpenetration. Alveoles are small hollows devel-oped on exposed bare and essentially fresh rock

    surfaces. They are typically a few centimetresdiameter and a couple of centimetres deep.They are well and widely developed in arid andsemi-arid coastal zones as well as inland sites(see MUSTOE, 1982, p. 108). They are bestdeveloped on sandstone, shale, and basic crys-talline rocks such as dolerite, as well as onlimestone and basalt. They are not as welldeveloped in fresh granite. Some with well-defined septa have been noted in humid tropi-cal north Queensland, in the monsoonal north

    of Western Australia and on The Humps nearHyden in the southwest of Western Australia(figure 2a). Others are developed in weatheredgranitic rocks where fractures are indurated andthe enclosed corestones have fallen away. Butalveoles are rare in fresh granite. Irregular andill-defined small hollows in which fresh rock isexposed attest the activity of haloclasty but themargins are diffuse and the intervening septathat delimit and define alveoles are not devel-oped, by contrast with adjacent sandstone,dolerite and, though more rarely, phytokarsticoutcrops of calcarenite.

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 39

  • 8/11/2019 Twidale Caves Granite

    6/23

    be ascertained, it is difficult to conceive of apetrologic or lithologic structure or texturethat is developed in the varied types of rock inwhich tafoni occur and that would plausiblyexplain their distribution. Some septa are cov-ered by green algae (MUSTOE, 1982) and it istempting to ascribe these separating walls tothe protective action of the biota. But is thealgal coating cause or effect?

    Two scenarios can be envisaged. First, thealgae covered the entire rock surface but theveneer was breached where salt settled andstarted to crystallise, causing disintegration ofthe rock and destruction of the algal coat inthose areas. The algae persist on the dry non-saline areas between haloclastic hollows.Second, the rock surface was randomly bro-ken where salt began to crystallise and thealgae colonised the relatively stable spaces

    between. The septa were reduced in area asthe hollows were enlarged, so that some rock

    CAD. LAB. XEOL. LAXE 33 (2008)40 Twidale and Bourne

    Alveoles have been attributed to wind,frost, and salt weathering but there is nowgeneral agreement that they are due in part atleast, and probably in large measure, to saltprecipitation and associated processes (e.g.WELLMAN and WILSON, 1965; WIN-KLER, 1975, 1981; BRADLEY et al., 1978;YOUNG, 1987). Many coastal occurrencesretain efflorescences of halite in the hollows,and the forces exerted by crystallising salts aresufficient to rupture rocks (WINKLER andSINGER, 1972; see also TWIDALE andBOURNE, 2008).

    At most sites the alveoles are randomlydistributed with no patterns suggestive, forinstance, of the control of weathering by frac-tures or fissures. The rock of the interveningribs or septa apparently is of the same compo-sition as that in which the hollows are shaped

    (MUSTOE, 1982). Though the nature of therock weathered and eroded cannot of course

    Figure 2. (a) Alveoles in granite, Emu Rock, near Mareeba, north Queensland.

  • 8/11/2019 Twidale Caves Granite

    7/23

    faces eventually became devoid both of alve-oles and algae

    Evidence from the development of alve-

    oles on natural rocks incorporated in humanconstructions of known date demonstrates thatthey form in a matter of decades (e.g.BARTRUM, 1936; GRISEZ, 1960; GILL etal., 1981; MUSTOE, 1982). Some workersregard alveoles as small or incipient tafoni(see e.g. the title of the GILL et al., 1981 ref-erence, and the entry in BATES and JACK-SON, 1987, p. 671), and in some instancesthis appears to be the case; though not in gran-

    ite (figure 2b). But at Cape Cassini on thenorth coast of Kangaroo Island, for instance,alveoles are well-developed on benches(exposed bedding planes) that comprise frac-ture-controlled blocks of flat-lying, fine-

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 41

    grained Proterozoic sandstone. All stages ofsurface lowering are developed in a sequencebeginning with isolated alveoles set into the

    original surface, to surfaces with a few alve-oles remaining, to fresh rock surfaces bound-ed by iron-indurated rims but devoid of alve-oles (figure 2c). In sandstone, at least, theeffect of the development of numerous alve-oles has been to eliminate the layer of rock inwhich they are formed. Such alveoles havenot coalesced to form tafoni. Evidently theprotection afforded by algae colonising thesepta is not enduring. The neat stripping of

    sandstone layers by alveolar weathering maybe facilitated by bedding. Certainly no suchsystematic lowering has been noted ongranitic or doleritic surfaces affected by themechanism.

    Figure 2. (b) Alveoles in argillite, Beda Valley, southern Arcoona Plateau, South Australia. Note the hollows thatmay be due to the coalescence of alveoles.

  • 8/11/2019 Twidale Caves Granite

    8/23

    That some rock basins are initiated on podsof minerals susceptible to weathering at thesite is suggested by field evidence (e.g.BOURNE and TWIDALE, 2002). Similarexploitation of weak minerals has beenadduced in explanation of pitting(TWIDALE and BOURNE, 1976) and could

    also account for the initiation of alveoles (andalso for the initiation of tafoni, for hollows ofirregular shape and varied dimensions andreferred to as pecking TWIDALE andBOURNE, 2001, see belowoccur on flaredslopes). Even so the distribution of some ispuzzling both at the local and regional scales.For instance, at Hallett Cove, on the coastsouth of Adelaide, South Australia, thin bedsof fine-grained sandstone interbedded andfolded with phyllitic mudstone display alve-oles, though the intervening argillitic rocks donot. Regionally, the absence of well-defined

    alveoles in granite on arid and semiarid coastsimplies that granite may be too tough to allowrapid disintegration under attack by eitherhaloclasty or hydration, or that algae cannotreadily gain a foothold or that colonisation isoutpaced by granite disintegration. Also, theoccurrence of well-defined alveoles in granite

    in seasonally humid northern Australia arguesagainst a haloclastic origin, for salts areflushed through the system in such environ-ments. This suggests either that the alveolesare inherited from a former drier climaticphase, or that hydration has been the dominantprocess in their formation. But if inherited, andgiven the known rapidity of chemical weather-ing in the tropics, how have (admittedly few)granitic alveoles survived? Also if hydration isso effective, why are such forms not morecommon in the monsoonal north of Australiaand in similar environments elsewhere?

    CAD. LAB. XEOL. LAXE 33 (2008)42 Twidale and Bourne

    Figure 2. (c) Alveoles in sandstone, Cape Cassini, north coast of Kangaroo Island, South Australia. Note the twolevels, the higher riddled with alveoles, the slightly lower almost clean presumably as a result of the coalescence

    of hollows but with a second generation beginning to develop.

  • 8/11/2019 Twidale Caves Granite

    9/23

    TAFONI

    The most common cavernous formsdeveloped in granitic rocks are tafoni (singu-lar tafone). Tafone is an Italian word mean-ing variously an aperture or cavity or (inCorsica) window. Tafoni are well developedin dry regions and on coasts in aridity or semi-aridity. They have been compared to alveolesand up to 10 cm deep but this is misleading.Of course, tafoni are initially small and atsome early stage of development (pecking?)must be less than the stipulated limit, but

    recognisable tafoni are characteristicallymuch larger. Many are large enough to pro-vide shelter for a family or a class of students.They are best defined as hollows developed

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 43

    on the undersides of blocks, boulders or sheetstructures, most frequently in granitic hostrocks (figures 3a and 3b). They extendupwards into the rock mass and in mostinstances are more-or-less enclosed by a toughcrust or visor of the country rock (granite).Some blocks and boulders are hollowed tosuch an extent that only a relatively thin shellremains. That large and deep tafoni surviveattests not only to the inherent strength of thefresh crystalline rock (e.g. DALE, 1923, p. 11;KESSLER et al., 1940) but also to that of thecase hardened zones (figure 3a). The physical

    weakness of argillaceous rocks explains why,though they are susceptible to weathering,tafoni infrequently develop in such lithologi-cal environments.

    Figure 3. (a) Tafoni developed inboulders on Yarwondutta Rock,

    near Minnipa, Eyre Peninsula.Note considerable mass of granitethat remains, in each instancemore than half the original mass.The larger boulder amounts toover 20 m3, the smaller some 8m3. Each block is supported bylegs of which the total area in con-tact with the underlying platformis approximately 110 cm2. Someof the supporting rock is laminat-ed but all is indurated on the outer

    face. (b) Tafoni in massive sheetstructure, Pearson Islands, easternGreat Australian Bight.

    (a)

    (b)

  • 8/11/2019 Twidale Caves Granite

    10/23

    In some tafoni the walls display mamilla-tion (figure 3c) and also books of flakes (fig-ure 3d) which become detached at the touch,and flakes and fragments of flakes are scat-tered on the floor. Clearly these tafoni are stillactive and are extending. Eventually the visoris breached from the inside at some sites toproduce the apertures that have given theforms their generic name (figure 3e). In otherexamples, however, and in some areas at adja-cent sites, the walls are solid with at most afew loose crystals or fragments. What deter-mines the activity or stability of the interior

    surfaces is not known. Whether the hollowsare associated with initial centres of seepageand salt precipitation is not known, but booksof laminae are well developed even, or per-haps most obviously, on the intervening risesof projections. Tafoni are favoured dens forkangaroos and wallabies and their droppings(excrement) encourage plant growth and thedevelopment of siliceous speleothems atsome tafoni sites as well as shelters andcrevices there are veritable forests of tiny cen-timetre-high speleothems (VIDAL ROMANIet al., 2003).

    CAD. LAB. XEOL. LAXE 33 (2008)44 Twidale and Bourne

    Figure 3. (c) Mamillated ceil-ing of tafone, RemarkableRocks, Kangaroo Island. (d)Books of flakes or laminatedgranite in ceiling of sheettafone on Ucontitchie Hill,northwestern Eyre Peninsula.

    (c)

    (d)

  • 8/11/2019 Twidale Caves Granite

    11/23

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 45

    Figure 3. (e) Granite boulder with breached visor, on The Humps, near Hyden.

  • 8/11/2019 Twidale Caves Granite

    12/23

    The hollows develop from below, typical-ly being initiated on the underside of the boul-der, block or sheeting slab or parting. Such

    sites are moist or moister than adjacentexposed surfaces because they are not asexposed to the sun and wind. Moreover,organic and biotic detritus that retains mois-ture accumulates in them. Frequently, weath-ering of the upper surface of the basal joint orsurface is matched by weathering of the

    CAD. LAB. XEOL. LAXE 33 (2008)46 Twidale and Bourne

    underside or surface on which rests the boul-der or block (figure 3f) so that some incipienttafoni are twinned with a shallow saucer-

    shaped depression that will become a basin orgnamma of one type or another (TWIDALEand CORBIN, 1963). It is significant that onthe platform beneath the larger tafoni shownin figure 3a is preserved a remnant of pittedgranite, indicative of the former presence ofsoil and moisture.

    Figure 3. (f) Base of boulder showing basal hollow and matching basin in platform on which the boulder stands.

    The growth of tafoni has been attributed towind erosion, frost action, microclimatic con-ditions, exfoliation, and haloclasty, the lattercategory embracing crystal growth, hydrationexpansion, and osmotic pressure (EVANS,1969; WINKLER, 1975; GOUDIE andVILES, 1997). Blackwelder (1929), forinstance, suggested that hydration of feldsparsto clays induced exfoliation, that is, scaling,

    flaking or lamination (cf. HUTTON et al.,1977). The range of temperature and humidity

    is less inside tafoni than outside(DRAGOVICH, 1967). Most tafoni are soenclosed by visors that wind cannot effective-ly penetrate into the interior space, and so on:most suggested explanations fail to satisfy thefield evidence in some degree or another.

    However, there is a measure of agreementthat salt weathering by crystal growth (seeTHOMSON, 1863) is largely responsible not

    only for many alveoles but also tafoni, and aswith alveoles there is much supporting evi-

  • 8/11/2019 Twidale Caves Granite

    13/23

    dence. Both forms are commonly developedin arid and coastal environments in which saltoccurs at the land surface in crystal form, sug-

    gesting that the same process may have a rolein the formation of the larger as well as thesmaller hollows (e.g. WINKLER, 1975;BRADLEY et al., 1978). Efflorescences ofhalite have been observed on the walls andceilings of some relatively open tafoni onnorthwestern Eyre Peninsula. Halite is hygro-scopic and water accelerates the rates of vari-ous forms of chemical weathering (e.g.McINNIS and WHITING, 1979; WINKLER,

    1981). That salt crystallisation and hydrationexert a force sufficient to overcome the tensilestrength and rupture fresh rocks includinggranite has been demonstrated experimentally(WINKLER and SINGER, 1972; WINKLER,1975; KNACKE and ERDBERG, 1975). Inaddition, however, the effects of hydration,and especially the production of hydrophilicclays from the reaction of feldspars and micaswith water (see e.g. BLACKWELDER, 1929;LARSEN, 1946; HUTTON et al., 1977), can-not be overlooked.

    Even so, haloclasty and associatedprocesses (see GOUDIE and VILES, 1997,pp. 123 et seq.) present problems in the con-text of tafoni development. One is to ascertainthe origin of the salt and explain how it cameto be in the tafoni walls developed on outcropsdistant from the sea or major salinas. Anotheris to explain why the visor has resisted weath-ering and erosion while tafoni are hollowed

    out in the fresh rock beneath.As to the first problem, weathering of theminerals that form granite produces no haliteor gypsum. Gneiss and schist may containminerals with chlorite (e.g. scapolite) butthere is no correlation between chlorine-bear-ing minerals and the distribution of tafoni.

    Salts are carried on the wind and arewashed from the atmosphere by rain (HUT-TON, 1976). Thus, the availability of salts incoastal zones and in arid lands is not in ques-tion. The difficulty is to explain how the saltscome to be at the base of rock masses, and

    later in walls and ceilings of tafoni. Giventime, saline solutions may percolate throughconsiderable thicknesses of rock, even rocks

    of low permeability such as granite, for allinclude small fissures. In addition, the releaseof lithostatic pressure consequent on the ero-sional unloading implied by the surface expo-sure of granite may have allowed separation atcrystal boundaries and along cleavages(BAIN, 1931), so that in time salts depositedin rainwater or, on the coast, in spray, mayhave infiltrated through the rock mass to thelower surface where the salts are precipitated,

    and where the rock is shattered. On the otherhand, weathering may be facilitated also bycrystal strain consequent on stresses inducedby tectonism or by gravity by blocks of freshrock resting on one another which may com-pensate and more than compensate for anyrelaxation associated with erosional unloading(RUSSELL, 1935; TURNER and VER-HOOGEN, 1960, p. 476; VIDAL ROMANI,1989).

    The formation of the visor has beenexplained in terms of a slight concentration ofsuch minerals as iron, manganese, and silicaby lichens, mosses and algae (e.g. FRY, 1926;SCOTT, 1967; though the expansion and con-traction of hyphae on taking in and sheddingwater has been cited as a possible cause ofrock disintegration: GOUDIE and VILES,1997, p. 157). Also, exposure and compara-tive dryness may be a relevant factor; forwater with salts in solution would tend to

    gravitate to the base of the rock, and any lat-eral seepage could result not in evaporation,but in the salts being taken up by the hyphaeand roots of plants. Alternatively, the concen-tration of ferruginous salts at the water tableappears to be responsible for some encrusta-tions. On the bornhardt known as TheHumps, near Hyden, in the southwest ofWestern Australia, for example, some boul-ders display two encrustations, one externaland the second located some 20 cm deeper inthe rock mass but now exposed by weatheringand erosion (figure 4a).

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 47

  • 8/11/2019 Twidale Caves Granite

    14/23

    Some hollows have formed but their devel-opment was either unaccompanied by visordevelopment or the former outer crust has beenwholly or largely eliminated (figure 4b). Suchfeatures are not tafoni for they lack the essen-tial enclosure. They are flared slopes devel-oped on boulders and are of limited extent, butthey are of two-stage origin and were initiated

    in the shallow subsurface (TWIDALE, 1962,TWIDALE and BOURNE, 1998).Some tafoni in the form of cliff-foot caves

    appear to be extensions of flared slopes and beassociated with water concentrations, in thisinstance at the scarp foot. The waters are,however, commonly alkaline. Such a mergingof concavity and cave can be seen at KokerbinHill, on the Yilgarn Craton of the southwest ofWestern Australia. Similar occurrences areseen on the southern flank of the arkosicUluru, where, near Mutitjulu (MaggieSprings) a flare merges laterally with a cliff-

    foot cave. Flared slopes demonstrably are etchforms and at many sites minor flares and hor-izontal slots occur just above soil, slope orplain level, suggesting that they were initiatedin the subsurface and have been exposed as aresult of recent anthropogenically-inducederosion. Thus, on the southern flank of Ulurubreaks of slope at about 35 m above present

    plain level coincide with deep caverns lackinga significant visor and so shaped that thetafoni warrant being described as gaping-mouth caves. They have been interpreted asmarking the position of a former fluctuatingwater table associated with a plain that wasessentially stable and stood 35-60 m higherthan the present plain (TWIDALE, 1978).

    But no tafoni as such have been located inshallow exposures. Small zones of intenseweathering and discrete voids have beenobserved (TWIDALE and BOURNE, 1975)but whether they were voids located in the

    CAD. LAB. XEOL. LAXE 33 (2008)48 Twidale and Bourne

    Figure 4. (a) Weathered boulder with remnants of two case-hardened zones, The Humps, near Hyden.

  • 8/11/2019 Twidale Caves Granite

    15/23

    subsurface or whether they were patches ofintense weathering that have fallen away afterexposure, is not clear. No hollows with anenclosing carapace or visor have been report-ed in granitic terrains. Tafoni are evidently asubaerial development.

    Overall it can be argued that hydrationmay account for subsurface development of

    flared forms, but that haloclasty assumesgreater significance after exposure, when theoverhanging surfaces and ceilings interceptdownward percolating saline solutions andwhere evaporation causes salt precipitation.But it is the inherent strength of crystallinegranite that allows hollows developed in therock mass to be maintained.

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 49

    Figure 4. (b) Flared boulder, the concavity lacking a visor, Murphy Haystacks, west coast of Eyre Peninsula.

    SUBTERRANEAN VOIDS

    Granite caves appear to be of severaltypes. Well-jointed granite with open fracturesspaced up to a few metres apart can be differ-entiated weathered by circulating meteoricand groundwaters penetrating along partings.The rock with which it comes into contact isweathered, the corners and edges of blocksmore rapidly than plane faces, so that angularblocks are converted to rounded boulders(MacCULLOCH, 1814; de la BECHE, 1898;LOGAN, 1851). The weathered granite or

    grus can be washed away by subsurface flush-ing and underground streams, thus creatingconnected openings or caves. TheLabertouche Cave in the South GippslandUplands east of Melbourne, Victoria, is such afracture-controlled cavern some 200 m longand essentially straight, though irregular indetail (OLLIER, 1965). Caves of similar ori-gin are reported from the Banana Range incentral Queensland (SHANNON, 1975),Colorado (ARNOLD, 1980) where the caves

    are up to 15 m high though subject to boulderfalls and prone to flooding; and from Guyana

  • 8/11/2019 Twidale Caves Granite

    16/23

    whence SHAW (1980) has described theMakatau cave system. The O Folon cavenear Vigo, Galicia, appears to be of similarorigin (VAQUIERO RODRIGUEZ et al.,2006). FENIGER (1969) reported variouskarstic features including caves from acid plu-tonic rocks (quartz diorite) in Columbia andthere is anecdotal evidence of extensive linkedbetween boulder caves on the BlackMountain, a large granite nubbin nearCooktown, north Queensland.

    Second, openings resulting from the pref-erential weathering of weaker members of a

    rock sequence are reported from some areas:sideritic veins in some of the High Tatra cavesof southern Poland, and feldspar-rich zones inthe Karkonosze Mountains (WOJCIK, 1961a,1961b).

    Third, in some areas a ferruginous crustdeveloped on granite boulders has beenbreached and the underlying kaolinised rockhas been preferentially weathered and fallenaway (figure 5a). At the eastern end of HydenRock, (Western Australia) shallow voids have

    CAD. LAB. XEOL. LAXE 33 (2008)50 Twidale and Bourne

    formed as a result of subsurface flushing (cf.RUXTON, 1958) of weathered granite to adepth of a metre or so beneath a thin butdurable case-hardened skin of slightly weath-ered granite that is cemented by silica andcolonised by algae (TWIDALE andBOURNE, 2001). Various stages of develop-ment can be observed from a clean opening toa void, to an irregular surface cut in weatheredgranite with a few remnants of the former crustpreserved. Both the intense and widespreadweathering of the granite, and the algal coatingthat appears to be responsible for the formation

    of the crust, and hence for the subterraneanvoids, may be associated with the proximity ofthe eastern sector of the inselberg to the formerCamm River, one of many palaeodrainagechannels preserved in the southern YilgarnCraton and adjacent areas (VAN de GRAAFFet al., 1977). The piedmont plain adjacent tothe eastern edge of Hyden Rock sounds hollowwhen stamped upon, suggesting that cavernousdevelopments extend beneath the presentplains surface.

    Figure 5. (a) Weathered granite with remnants of cemented crust protected by algae enclosing substantial voids,eastern sector of Hyden Rock, Western Australia.

  • 8/11/2019 Twidale Caves Granite

    17/23

    What may be an early stage in such sub-surface weathering combined with the forma-tion of a thin crust can be seen at GrahamRock, a few kilometres east of Hyden Rock

    and like it, immediately adjacent to the Cammdrainage. Patches of the granite slopes displaysinuous narrow apertures and shallow planarhollows beneath the crust (figure 5b).

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 51

    Figure 5. (b) Shallow and narrow breaches and subsurface voids (worm tubes) on Graham Rock, near Hyden.

    Fourth, voids and openings of various kindsoccur where blocks and slabs have been dis-placed either under gravity or as a result ofearth tremors (figure 5c) or as a consequenceresult of preferential weathering, for instancealong fractures though such voids are rarelycovered. The formation of blisters and A-tents, or pop-ups (figure 5d), as a result ofcompressive stress has enclosed triangularspaces (TWIDALE and SVED, 1978; WAL-LACH et al., 1993). Whether they can be con-sidered subterranean is dubious, though theyare located at depths as shallow as the voidsdiscussed above, and many are separated fromspace by slabs thicker than the crusts enclosingthe voids featured in figures 5b and 5c. Many

    A-tents have formed instantaneously as a resultof earthquakes (TWIDALE and BOURNE,

    2000) but others evidently have formed gradu-ally and are still in the process of conversionfrom blisters, continued compression produc-ing the crestal crack that distinguishes the twoforms (TWIDALE and BOURNE, 2003).

    Fifth, caves have been formed as a result ofweathering along sheet fractures and particular-ly the preferential weathering of triangularslabs like those exposed on the flanks of someinselbergs (TWIDALE and VIDAL ROMANI,2005, p. 40) but located within the hills. Theslabs are associated with sheet fractures that areplanes of shear activated by lateral compres-sion and slippage along sheet fractures of dif-ferent radii (TWIDALE et al., 1996). The well-known cave within Enchanted Rock (figure

    5e), in the Llano of central Texas, USA, is ofthis type (KASTNING, 1976).

  • 8/11/2019 Twidale Caves Granite

    18/23

    CAD. LAB. XEOL. LAXE 33 (2008)52 Twidale and Bourne

    Figure 5. (c) Voids possibly caused by disturbance ofblocks by earth tremor at Devils Marbles, NorthernTerritory, Australia. (d) A-tent, Kokerbin Hill, south-western Yilgarn Craton. (e) Part of cave withinEnchanted Rock, Texas, USA (E Kastning).

    (c)

    (d)

    (e)

  • 8/11/2019 Twidale Caves Granite

    19/23

    DISCUSSION AND CONCLUSIONS

    Caves of several shapes, sizes, and originsare found developed in granitic rocks. Most ifnot all have their congeners in other rocktypes. Though understood in general termsmost still pose problems.

    Tafoni are particularly well developed ingranite blocks, boulders and sheet struc-tures. This can be attributed partly to thecontrasted susceptibility of granite in dry

    and wet microenvironments; partly to theease with which dry granite can be furtherreinforced by biotic impregnation by ironoxides and silica; partly to the susceptibilityof granite to hydration and haloclasty, result-ing in flaking and the gradual growth of hol-lows, including their upward extension; andpartly to the inherent strength of fresh gran-ite which allows hollows to form withoutinducing the collapse of the entire rockmass.

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 53

  • 8/11/2019 Twidale Caves Granite

    20/23

    REFERENCES

    ARNOLD, .J. (1980). Death in the Rockies.NSS News (September), p. 209-215.

    BAIN, G. W. (1931). Spontaneous rockexpansion. The Journal of Geology, 39:715-735.

    BARTRUM, J. D. (1936). Honeyconb weath-ering near the shore line. New ZealandJournal of Science and Technology, 18:593-600.

    BATES, R. L. and JACKSON, J. A. (1987).

    Glossary of Geology. (3rd edition).American Geological Institute,Alexandria, VA.

    BECHE, H. T. de la (1839). Report on theGeology of Cornwall, Devon and West

    Somerset (Geological Survey of Englandand Wales). Longman, Orme, Brown,Green and Longmans, London, 648 pp.

    BLACKWELDER, E. (1929). Cavernousrock surfaces of the desert. AmericanJournal of Science, 17: 393-399.

    BOURNE, J. A. and TWIDALE, C. R. (2002).Morphology and origin of three bornhardtinselbergs near Lake Johnston, DundasShire, Western Australia. Journal of theRoyal Society of Western Australia, 85:83102.

    BRADLEY, W. C., HUTTON, J. T. andTWIDALE, C. R. (1978). Role of salts in

    the development of granitic tafoni, SouthAustralia. The Journal of Geology, 86:647-654.

    DALE, T. N. (1923). The commercial granitesof New England. United States GeologicalSurvey Bulletin, 738, 488 pp.

    DRAGOVICH, D. J. (1967). Flaking, aweathering process operating on cav-ernous rock surfaces. Geological Societyof America Bulletin, 78: 801-804.

    EVANS, I. S. (1969). Salt crystallization androck weathering; a review. Revue de

    Gomorphologie Dynamique, 19: 155-177.

    FEINIGER, T. (1969). Pseudokarst in quartzdiorite, Columbia. Zeitschrift frGeomorphologie, 13: 287-296.

    FRY, E. J. (1926). The mechanical action ofcorticolous lichens.Annals of. Botany, 40:397-417.

    GILL, E. D., SEGNIT, E. R. and McNEILL,N. H. (1981). Rate of formation of honey-comb weathering features (small scaletafoni) on the Otway coast, S.E. Australia.

    Proceedings of the Royal Society ofVictoria, 92: 149-152.

    GOUDIE, A. and VILES, H. (199). SaltWeathering Hazards. Wiley, Chichester,241 pp.

    GRISEZ, L. (1969). Alveolisation littorale deschistes metamorphiques. Revue deGomorphologie Dynamique, 11: 164-177.

    HALE, H. M. and TINDALE, N. B. (1930).

    Notes on some human remains in thelower Murray valley.Records of the SouthAustralian Museum, IV: 145-218.

    HANKS, P. (Editor) (1986). CollinsDictionary of the English Language.Second Edition. Collins, Sydney, 1771 pp.

    HUTTON, J. T. (1976). Chloride in rainfall inrelation to distance from the ocean.Search, 7: 207-208.

    HUTTON, J. T., LINDSAY, D. S. and

    TWIDALE, C. R., (1977), The weatheringof norite at Black Hill, South Australia.Journal of the Geological Society of

    Australia, 24: 37-50.

    JENNINGS, J. N. (1976). A test of the impor-tance of cliff-foot caves in tower karstdevelopment. Zeitschrift frGeomorphologie SupplementBand, 26:92-97.

    KASTNING, E. (1976). Granitic karst andpseudokarst, Llano County, Texas withspecial reference to Enchanted Rock

    CAD. LAB. XEOL. LAXE 33 (2008)54 Twidale and Bourne

  • 8/11/2019 Twidale Caves Granite

    21/23

    Cave. Proceedings of the 1976 NationalSpeleological Society Annual Convention,p. 43-45.

    KESSLER, D. W., INSLEY, H. and SLIGH,W. H. (1940). Physical, mineralogical anddurability studies on the building andmonumental granites of the United States.Journal of Research of the National

    Bureau of Standards, 24: 161-206.

    KNACKE, O. and ERDBERG, R. (1975). Thecrystallisation pressure of sodium sulphatedecahydrate. Berichte der Bunsen-Gesellschaft fr Physikalische Chemice,

    79: 653557.

    LARSEN, E. S. (1948). Batholith and associ-ated rocks of Corona, Elsinore and SanLuis Rey quadrangles, southernCalifornia. Geological Society of AmericaMemoir, 29, 113 pp.

    LOGAN, J. R. (1851). Notices of the geologyof the straits of Singapore. QuarterlyJournal of the Geological Society of

    London, 7: 310-344.MacCULLOCH, J. (1814). On the granite torsof Cornwall. Transactions of theGeological Society, 2: 66-78.

    McINNIS, C. and WHITING, J. D. (1979).The frost resistance of concrete subjectedto a deicing agent. Cement and ConcreteResearch, 9: 325-336.

    MULVANEY, D. J., LAWTON, G. H. andTWIDALE, C. R. (1964). Archaeological

    excavation at Rock Shelter No 6, FrommsLanding, South Australia. Proceedings ofthe Royal Society of Victoria, 77: 487-516.

    MUSTOE, G. E. (1982). The origin of honey-comb weathering. Geological Society ofAmerica Bulletin, 93: 108-115.

    OLLIER, C. D. (1965). Some features ofgranite weathering in Australia.Zeitschriftfr Geomorphologie, 9: 285-304.

    PATON, J. R. (1964). The origin of the lime-stone hills of Malaya.Journal of TropicalGeography, 18: 138-147.

    RUSSELL, G. A. (1935). Crystal growth insolution under local stress. AmericanMineralogist, 20: 733-737.

    RUXTON, B. P. (1958). Weathering and sub-surface erosion in granite at the piedmontangle, Balos, Sudan. GeologicalMagazine, 45: 353-377.

    SCOTT, G. D. (1967). Studies of the lichensymbiosis. 3. The water relations oflichens on granite kopjes in central Africa.The Lichenologist, 3: 368-385.

    SCRIVENOR, J. B. (1928). The Geology of

    Malayan Ore-Deposits. Macmillan,London, 216 pp.

    SHANNON, C. H. C. (1975). Pseudokarstcaves in duricrust/granite terrain, BananaRange, central Queensland. Proceedings

    of the 10th Biennial Conference,

    Australian Speleological Federation

    (Brisbane), p. 20-24.

    SHAW, P. (1980). Cave development on agranite inselberg, south Rupununi savan-

    nas, Guyana. Zeitschrift frGeomorphologie, 24: 68-76.

    TATE, R. (1884). Notes on the physical andgeological features of the basin of theLower Murray River. Transactions of theRoyal Society of South Australia, 7: 24-46.

    THOMSON, J. (1863). On the disintegrationof stones exposed in buildings and other-wise exposed to atmospheric influences.

    Report of the 23rdMeeting of the British

    Association for the Advancement ofScience. Notices and Abstracts. Murray,London.

    TURNER, F. J. and VERHOOGEN, J. (1960).Igneous and Metamorphic Petrology.McGraw-Hill, New York, 694 pp.

    TWIDALE, C.R. (1962). Steepened marginsof inselbergs from north-western EyrePeninsula, South Australia. Zeitschrift frGeomorphologie, 6: 51-69.

    TWIDALE, C. R. (1964). Effect of variationsin rate of sediment accumulation on a

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 55

  • 8/11/2019 Twidale Caves Granite

    22/23

    bedrock slope at Fromms Landing, SouthAustralia.Zeitschrift fr GeomorphologieSupplementBand, 5: 177-191.

    TWIDALE, C. R. (1978). On the origin ofAyers Rock, central Australia. Zeitschriftfr Geomorphologie SupplementBand, 31:177-206.

    TWIDALE, C. R. (2000). Scarp retreat, slopestability, and the evolution of piedmontassemblages. South African GeographicalJournal, 82: 54-63.

    TWIDALE, C. R. (2006). Origin and develop-

    ment of Karstinselberge, with particularreference to some South East Asian evi-dence. Geological Society of MalaysiaBulletin, 49: 145-155.

    TWIDALE, C. R. and BOURNE, J. A. (1975).The subsurface initation of some minorgranite landforms. Journal of theGeological Society of Australia, 22: 457-484.

    TWIDALE, C. R. and BOURNE, J. A. (1976).

    Origin and significance of pitting on gran-ite rocks.Zeitschrift fr Geomorphologie,20: 405406.

    TWIDALE, C. R. and BOURNE, J. A.(1998). Flared slopes revisted. PhysicalGeography, 19: 110-133.

    TWIDALE C. R. and BOURNE, J. A. (2000).Rock bursts and associated neotectonicforms at Minnipa Hill, northwestern EyrePeninsula, South Australia.Environmental

    and Engineering Geoscience, 6: 129-140.TWIDALE, C. R. and BOURNE, J. A. (2001).

    A Field Guide to Hyden Rock, Western

    Australia. Wave Rock Management P/L,Hyden, 68 pp.

    TWIDALE, C. R. and BOURNE, J. A. (2003).Active dislocations in granitic terrains ofthe Gawler and Yilgarn cratons, Australia,and some implications. South AfricanJournal of Geology, 106: 71-84.

    TWIDALE, C. R. and BOURNE, J. A. (2008).Haloclasty on the bed of Lake Gairdner,

    South Australia. Cadernos do LaboratorioXeolxico de Laxe. In press this volume.

    TWIDALE, C. R. and CORBIN, E. M.(1963). Gnammas. Revue deGomorphologie Dynamique, 14: 1-20.

    TWIDALE, C. R. and SVED, G. (1978). Minorgranite landforms associated with therelease of compressive stress. AustralianGeographical Studies, 16: 161-174.

    TWIDALE, C. R. and VIDAL ROMAN, J.R. (2005). Landforms and Geology ofGranite Terrains. Balkema Leiden, 351pp.

    TWIDALE, C. R., VIDAL ROMAN, J. R.,CAMPBELL, E. M. and CENTENO, J. D.(1996). Sheet fractures: response to ero-sional offloading or to tectonic stress?Zeitschrift fr Geomorphologie

    SupplementBand, 106: 1-24.

    VAN de GRAAFF, W. J. E., CROWE, R. W.A., BUNTING, J. A. and JACKSON, M. J.(1977). Relict Early Cainozoic drainagesin arid Western Australia. Zeitschrift frGeomorphologie, 21, 379-400.

    VAQUEIRO RODRGUEZ, M., BARRIEROBEN, B., COSTAS VZQUEZ, R.,SUREZ PREZ, R. and GROBAGONZLEZ, X. (2006). Relation betweenstructure and morphology in the develop-ment of the granite cave of O Foln (Vigo,Galicia - Spain). Cadernos do LaboratorioXeolxico de Laxe, 31: 87-103.

    VIDAL ROMAN, J. R. (1989). Granite mor-phology in Galicia (NW Spain).Cuadernos do Laboratorio Xeolxico de

    Laxe, 13: 89-163.

    VIDAL ROMAN, J. R., BOURNE, J. A.,TWIDALE, C. R. and CAMPBELL, E. M.(2003). Siliceous cylindrical speleothemsin granitoids in warm semiarid and humidclimates.Zeitschrift fr Geomorphologie,47: 417-437.

    WALLACH, J. L., MOHAJER, A. A.,McFALL, G. H., BOWLBY, J. R.,PEARCE, M. and McKAY, D. A. (1993).

    CAD. LAB. XEOL. LAXE 33 (2008)56 Twidale and Bourne

  • 8/11/2019 Twidale Caves Granite

    23/23

    Pop-ups as geological indicators of earth-quake-prone areas in intraplate easternNorth America. In: Neotectonics: RecentAdvances. L. A. Owen, I. Stewart and C.Vita-Finzi (Eds). Quaternary ProceedingsNo. 3, Quaternary Research Association,Cambridge, U.K., p. 67-83.

    WELLMAN, H. W. and WILSON, A. T.(1965). Salt weathering: a neglected geo-logical erosive agent in coastal and aridenvironments.Nature, 205: 1097-1098.

    WINKLER, E. M. (1975). Stone: Properties,Durability in Mans Environment.Springer, New York, 230 pp.

    WINKLER, E. M. (1981). Salt weathering bysodium chloride in the Saudi Arabian

    Desert.American Journal of Science, 281:12441245.

    WINKLER, E. M. and SINGER, P. C. (1972).Crystallisation pressure of salts in stoneand concrete. Geological Society ofAmerica Bulletin, 83: 3509-3514.

    WOJCIK, Z. (1961a). Caves in granite in theTatra Mountains. Actes 3re CongresInternationale de Spleologie (Salzburg)A, p. 43-44.

    WOJCIK, Z. (1961b). Karst phenomena andcaves in the Karkonosze granites. DieHhle, 12: 76.

    YOUNG, A. R. M. (1987). Salt as an agent inthe development of cavernous weathering.Geology, 15: 962-966.

    CAD. LAB. XEOL. LAXE 33 (2008) Caves in granitic rocks 57