Rivers going underground, great springs

emerging from the ground, independent

hollows and basins instead of connecting
valleys, deep potholes and vast caves,
isolated towerlike hills reminiscent of the
unbelievably steep peaks depicted in
Chinese paintings—these are some of the
distinctive features of karst, the name given
to the kinds of country that owe their
special characteristics to the unusual degree
of solubility of their component rocks in
natural waters.
The special nature of karst is not only
intrinsically interesting; it affects many
aspects of life in the areas where it is
found—water supplies, agriculture, engin­
eering construction, tourism. There are,
then, practical as well as scientific reasons
for its study.
The dramatic quality of karst landforms
has caught the imagination of specialists
and laymen alike. This book contains much
to stimulate and inform the general reader
as well as the undergraduate and high
school student for whom it is written. It
will be of particular interest to the
hydrologist, the speleologist, and the sport­
ing caver and potholer.
 This book was published by ANU Press between 1965–1991.
This republication is part of the digitisation project being carried
out by Scholarly Information Services/Library and ANU Press.
This project aims to make past scholarly works published
by The Australian National University available to
a global audience under its open-access policy.
AN INTRODUCTION TO SYSTEMATIC GEOMORPHOLOGY
The following volumes have been published:

3 Landforms of Cold Climates J. L. Davies

4 Coasts E. C. F. Bird
5 Structural Landforms C. R. Twidale
6 Volcanoes Cliff Ollier
 An Introduction to Systematic Geomorphology
VOLUME SEVEN

Karst
J. N. JENNINGS

Professorial Fellow in Geomorphology,

The Australian National University

1971
AUSTRALIAN NATIONAL UNIVERSITY PRESS
CANBERRA
 © Joseph Newell Jennings 1971

This book is copyright. Apart from any fair dealing for the
purposes of private study, research, criticism, or review, as
permitted under the Copyright Act, no part may be reproduced by
any process without written permission. Inquiries should be made
to the publisher.

National Library of Australia Card no.

and ISBN 0 7081 0324 3
Library of Congress Catalog Card no. 75-161236,

P RI NT ED IN AUSTRALIA

Registered in Australia for transmission by post as a book

 INTRODUCTION TO THE SERIES

This series is conceived as a systematic geomorphology at univer­

sity level. It will have a role also in high school education and it is
hoped the books will appeal as well to many in the community at
large who find an interest in the why and wherefore of the natural
scenery around them.
The point of view adopted by the authors is that the central
themes of geomorphology are the characterisation, origin, and
evolution of landforms. The study of processes that make land­
scapes is properly a part of geomorphology, but within the present
framework process will be dealt with only in so far as it elucidates
the nature and history of the landforms under discussion. Certain
other fields such as submarine geomorphology and a survey of
general principles and methods are also not covered in the volumes
as yet planned. Some knowledge of the elements of geology is
presumed.
Four volumes will approach landforms as parts of systems in
which the interacting processes are almost completely motored by
solar energy. In humid climates (Volume One) rivers dominate
the systems. Fluvial action, operating differently in some ways, is
largely responsible for the landscapes of deserts and savannas also
(Volume Two), though winds can become preponderant in some
deserts. In cold climates, snow, glacier ice, and ground ice come
to the fore in morphogenesis (Volume Three). On coasts (Volume
Four) waves, currents, and wind are the prime agents in the
complex of processes fashioning the edge of the land.
Three further volumes will consider the parts played passively
by the attributes of the earth’s crust and actively by processes
deriving energy from its interior. Under structural landforms
(Volume Five), features immediately consequent on earth move­
ments and those resulting from tectonic and lithologic guidance of
denudation are considered. Landforms directly the product of
volcanic activity and those created by erosion working on volcanic
vi In tro d u c tio n to the Series

materials are sufficiently distinctive to warrant separate treatment

(Volume Six). Though karst is undoubtedly delimited lithologic­
ally, it is fashioned by a special combination of processes centred
on solution so that the seventh volume partakes also of the
character of the first group of volumes.

J. N. Jennings
General Editor
 PREFACE

This small book perforce omits many kinds of karst features and
even more of the ideas karst has prompted. Because of personal
experience, Australia, New Zealand, New Guinea, and Malaysia
loom larger than their place in karst literature might direct but an
outlook other than European or North American is part of the
object of this series. I have preferred enlarging on the elements to
pursuing the complex, and regional ‘personality’ has been omitted
altogether. Recent morphometric experimentation and process
study have been given more space than product from these
approaches probably warrants at the moment, yet the growing
points deserve some favour. M y hope is that this restricted selection
of examples and ideas w ill entice readers to more thorough books
on the subject and above all to karst itself where one cannot fail
to be seized by the endless round of rock and water, rock consumed
by water, water creating rock.

The book has benefited greatly from the suggestions and

criticism of my friends M r C. D. Ollier and Dr P. W. Williams to
whom I am particularly grateful. However little it will satisfy him,
Chapter V I I I is in debt to M r R. M. Frank without his formal
help. I have consulted D r K. Crook on limestone classification,
Dr R. Rosich on the chemistry of limestone solution, and M r L.
M ilton on piping. I am indebted to them all for aid, though the
inadequacies of the book remain mine. M r K. Fitchett helped in
innumerable ways in both field and laboratory. M r ‘ Ian’ Heyward,
who drew the figures, was both patient and constructive. I thank
both of them and all those who contributed photographs and on
whose work the line figures are based. Last but not least I must
acknowledge my memory-filling debt to the goodly company from
several countries, who have shown me caves or gone caving
with me.
Canberra J.N.J.
1970
vii
 CONTENTS
page
Introduction to the Series v

Preface vii

I THE NATURE OF KARST AND ITS STUDY 1

The meaning of karst 1
Types of karst 3
Pseudokarst 5
History of karst study 5
Karst literature 7
Terminology 8
Methods of investigation 8

II KARST ROCKS 10
Limestone 10
Reef facies 14
Dolomite 16
Evaporites 16
Pores and planes of weakness: permeability and
strength 17

III KARST PROCESSES 23

Solution and precipitation 23
Piping 30
Subsidence 32
Collapse 33
Other weatheringprocesses 38
ix
X Contents
papc
IV MINOR SOLUTION SCULPTURE 39
Factors affecting minor solution sculpture 39
Types of minor solution forms on limestone 42
Minor solution features on other karstrocks 52
Assemblages of minor solution features 54
Rates of solution in superficial zones 54

V DRAINAGE 61
Infiltration, overland flow, and throughflow 61
Surface rivers 64
River regimes 66
Sinking of rivers 69
Springs 74
Regimes of karst springs 79
Relations of surface and undergrounddrainage 84
Theory of karst hydrology 88

VI SURFACE LANDFORMS 98
Gorges 98
Meander caves 100
Natural bridges 101
Constructional action of rivers 106
Semiblind valleys 109
Blind valleys 110
Steepheads 112
Dry valleys 114
Dolines and cockpits 120
Uvalas 135
Poljes 136
Karst margin plains 143

VII KARST CAVES 144

General characteristics 145
Cave formation 147
Morphometric analysis of caves 165
Underground solution 168
 C ontents xi
page
V III CAVE DEPOSITS 171
Speleothems 171
Cave ice 174
Biogenic deposits 174
Clastic sediments 175
Entrance facies 177

IX INFLUENCE OF C LIM A TE AND

RATES OF DENUDATION 179
Rates of karst denudation 179
The arid extreme 181
The cold extreme 184
The ‘botanic hothouse’ extreme: tropical humid
karst 186
Further karst climato-morphogenic systems 194

X INFLUENCE OF GEOLOGICAL
STRUCTURE 196
Depositional relief in reef and dune limestone karsts 196
Structural relief in reef karst 199
Holokarst and merokarst 200
Structure in tropical humid karst 201
Structure within caves 202
Hydrogeologie systems 209

XI HISTORICAL GEOMORPHOLOGY
OF KARST 210
The karst cycle 210
Climatic change 213
Tectonic movements and changes of base levels 215
Complex evolution 221

X II PRESENT STATE OF KARST

GEOMORPHOLOGY AND ITS V A LU E 224

Bibliography 229
Index 243
 FIGURES
page
1 Carbonate rocks 11
2 Major kinds of limestone 12
3 Reef facies and present relief, Napier Range, West
Kimberley, Australia 15
4 Saturation equilibrium curves for calcium carbonate 26
5 Mixing corrosion 27
6 Stress effects on cave cross-sections, (a) at considerable
depth, (b) at shallow depth 35
7 Collapse dome formation, (a) theoretical sequence,
(b) in Easter Cave, Augusta, Western Australia 37
8 Solution sculpture and moraine cover, Clare, Ireland 50
9 Solution pipes in gypsum, Walkenried, Harz, Germany 53
10 Runnel and grike relationships 55
11 Solution pit in calcareous moraine, Austrian Alps 58
12 Glacial erratic on limestone pedestal, Norber Crags,
Craven, England 58
13 Seasonal discharge of a karst and a normal river 67
14 Discharge, hardness and limestone removal relationships
in certain rivers and springs 68
15 Streamsinks, River Manifold, Peak District, England 72
16 Ordering of streamsinks 73
17 Types of springs 75
18 Artesian spring, Ras el ’Ain, Syria 80
19 Blue Waterholes, Cooleman Plain, N.S.W. 82
20 Cave, stream and dry valley relationships, Clare. Ireland 85
xii
 Figures xiii
page
21 Effects of structure on surface and underground drainage
relationships 87
22 Karst hydrologic zones on watertable concept 89
23 Surface and underground drainage, Malham Cove, Craven,
England 91
24 Karst hydrologic zones on independent karst conduit
concept 92
25 Effect of cross-section and hydrostatic pressure on water
levels 94
26 Underground drainage in (a) Dachstein plateau, Austria,
(b) in Chalk, Northern France 95
27 Deep phreatic water movement with (a) uniform inputs,
(b) irregular inputs 96
28 Meander cave, Boree Creek, Borenore, N.S.W. 101
29 Natural bridge due to (a) cave roof collapse, (b) river
piracy, Cedar Creek, Virginia 103
30 Plitvice Lakes, River Korana, Yugoslavia 108
31 Semiblind valley, Cooleman Plain, N.S.W. 110
32 Blind valley, Yarrangobilly, N.S.W. 111
33 Dry valley, Pazinski Potok, Istria, Yugoslavia 115
34 Sadeng dry valley, Southern Java 116
35 Dry valleys, Peak District, England 117
36 Collapse and solution dolines 121
37 Solution dolinc, Grandes-Chaumilles, Swiss Jura 123
38 Collapse dolinc and possible solution dolinc, The Punch­
bowl, Wee Jasper, N.S.W. 124
39 Other types of dolines and dolinc asymmetry 127
40 (a) Cenote, Hells Hole, Mt Gambier, South Australia
(h) Subjacent karst collapse dolinc, Braidwood Big Hole,
N.S.W. 128
41 Dolinc karst near Mammoth Cave, Kentucky 129
42 Dolinc morphometry, Craigmorc, South Canterbury, New
Zealand 132
43 Tropical cockpit morphometry, Mt Kaijende, Australian
New Guinea 134
xiv Figures
page
44 Kupres polje, Bosnia, Yugoslavia 137
45 Relationships of poljes to structure 140
46 Polje sections and block diagram 141
47 Longitudinal section, Punchbowl-Signature Caves, Wee
Jasper, N.S.W. 147
48 (a) Pic St-Loup pothole, Montpellier, France
(b) Tectonic cave diagram
(c) Shelves in caves due to differential solution 148
49 (a) Blind shafts, Kentucky
(b) Absorption pothole, Gouffre Henne Morte, Pyrenees,
France
(c) Barber Cave profile, Cooleman Plain, N.S.W. 152
50 Branchwork (a) and network (b)caves 154
51 Current markings 155
52 Cave levels and river terraces,Tiliff,Belgium 159
53 Types of passage cross-section 160
54 (a) Solution hollowing, Picznice Cave, Hungary
(b) Resurgence pothole, La Luirc, Vcrcors, France
(c) Potholes of Grand Plan de Canjuers, Provence,
France 163
55 Solution patterns, (a) Clare, Ireland, (b)Mendip, England 169
56 (a) Straw stalactite
(b) Angular helictite
(c) Curving helictite 172
57 Conekarst, northern Puerto Rico 188
58 Cockpit karst evolution 189
59 Tower karst, Kinta valley, West Malaysia 190
60 (a) Cliff-foot cave sections
(b) Mogote asymmetry 191
61 Types of emerged reef relief 197
62 Types of limestone tower 202
63 Mendip cave development in steep dip 206
64 Mair’s Cave, Flinders Ranges, South Australia, in vertical
beds 206
65 Hydrogeologie systems in low to moderate karst relief 209
 Figures xv
page
66 Karst cycle of Grund 210
67 Karst cycle of Cvijic 212
68 (a) Karst evolution, Craven, England
(b) Cave at rejuvenation head, Craven, England 216
69 Storeying in tower karst, Baisha, Kweilin, China 218
 PLATES
page
1 Weathering of Oligocene limestone, King Country, New
Zealand 21
2 Gypsum karst, Bas Queyras, French Alps 31
3 Collapse dome, Koonalda Cave, Nullarbor Plain, South
Australia 36
4 Solution ripples, Cooleman Plain, N.S.W. 43
5 Solution flutes and bevel, Wee Jasper, N.S.W. 44
6 Solution funnel steps, Mt Arthur, New Zealand 45
7 Meandering solution runnels, Nullarbor Plain, Western
Australia 46
8 Solution notch, Central Highlands, New Guinea 49
9 Limestone pavement, Craven, England 51
10 Exposed subsoil solution features. Central Highlands, New
Guinea 52
1 1 Underground breach of limestone ridge, Central Highlands,
New Guinea 62
12 Inflow cave, River Rak, Slovenia 70
13 Outflow cave, River Lagaip, Central Highlands, New
Guinea 76
14 Boiling spring, Piu Piu, Takaka, New Zealand 77
15 Bungonia Gorge, N.S.W. 99
16 Natural bridge, London Bridge, Burra. N.S.W. 104-5
17 Tufa dams, McKinstry's Canyon, New Mexico, U.S.A. 107
18 Steephead, Sogöksu, Turkey 113
19 Dry valley, Cooleman Plain, N.S.W. 114
xv i
 Plates xvii
page
20 Solution doline, Velebit Mountains, Yugoslavia 122
21 Cenote, Konya, Turkey 126
22 Doline field, Craigmore, New Zealand 130
23 Uvala, Lost World, King Country, New Zealand 136
24 Planina Polje, Slovenia 138
25 Polje, Napier Range, Western Australia 138
26 Waterfall and plunge pool, Coolagh River Cave, Clare,
Ireland 150
27 Spongework, Tunnel Cave, Borenore, N.S.W. 156
28 Dry passage, Barber Cave, Cooleman Plain, N.S.W. 158
29 Keyhole passage, Metro Cave, Charleston, New Zealand 161
30 Speleothems, Metro Cave 162
31 Arid karst, Nullarbor Plain, Western Australia 182
32 Conekarst, Iehi, River Kikori, New Guinea 183
33 Snowy mountain karst, Mt Arthur, New Zealand 185
34 Tower karst, Perlis, West Malaysia 190
35 Doline and arete karst, Mt Kaijende, New Guinea 194
36 Gorge in dune limestone, Deepdene, Western Australia 199
37 Karst corridor, Kinta valley. West Malaysia 204
38 Blockpile, Exit Cave, Tasmania 205
39 Drowned tower karst, Langkawi Islands, West Malaysia 220
40 Subsidence doline, May River, Konya, Turkey 227
 TABLES
pope
1 Permeability of common rocks 17
2 Compressive strength of common rocks 18
3 Carbonate concentrations of superficial waters 57
4 Surface lowering rates from limestone pedestals 59
5 Analysis of drainage densities on different rocks in Clinch
Mountains area, Virginia and Tennessee 65
6 Rates of karst denudation 181
7 Hydrogeologie systems in low to moderate karst relief 208

xviii
 I

THE NATURE OF KARST AND ITS STUDY

THE MEANING OF KARST

Derived from the Slav word Krs, meaning crag or stone and also
the geographical name of limestone country in western Slovenia,
the German word Karst has passed into international usage in an
extended and imprecise way. It is therefore necessary to explain
how the term will be used in this book and to comment on other
usage. This discussion of the meaning of ‘karst’ will also serve to
introduce its study.
Karst signifies terrain with distinctive characteristics of relief
and drainage arising primarily from a higher degree of rock
solubility in natural waters than is found elsewhere. The word is
also used adjectivally to refer to rock, water, streams, caves and
other features making up such landscape.
Karst is frequently marked by intermittent stream flow and by
valleys without stream channels— dry valleys— yet it cannot be
adequately defined in terms of replacement of surface by under­
ground drainage. Much distinctive tropical karst has predominantly
surface drainage; pumice and loess areas are often virtually with­
out it yet fail to develop other attributes common in karst.
Continuous systems of slopes and channels taking water to the
sea typical of fluvial relief give way in karst to apparently dis­
organised, even chaotic relief in which valleys are frequently
interrupted and where there are many types of closed depression.
Again, however, some kinds of karst lack these characters which
conversely are found in country other than karst. In a dry climate,
for instance, river action may not be persistent enough to over­
come the disruptive effects of earth movements on valley systems
or to offset the power of wind to hollow out basins. In dune fields
the primary hollows resulting from aeolian deposition may remain
1
2 K arst

closed for a long time, even in humid regimes, because rainfall

soaks easily into the sand and streams cannot integrate the fields
of hollows into valley networks.
Nor can definition be solely in terms of the morphology of
limestone since similar relief and drainage are developed on
other rocks. The topography associated with the full range of
carbonate rocks is a better basis since dolomite is the next most
widespread rock type giving rise to karst. But in addition to
carbonate rocks, very soluble evaporites such as halite, gypsum,
and anhydrite induce karst, especially in dry climate as with the
Castile Formation of northwest Texas. At the opposite extreme in
very humid climate some karst characteristics develop in less
soluble rocks such as quartz diorite in northern Colombia
(Feininger 1969) and eclogite in the Owen Stanley Range of New
Guinea.
The definition adopted in this book stresses solubility. However,
all rocks are soluble in natural waters to some extent and in karst
itself solution need not be the only or even the dominant process,
but it does play a significantly more important role here than in
other landscapes. Solution operates in more than one way to
produce karst but by far the most important effect lies in the
enlargement of voids in the rock. The outcome is the steady
growth of permeability. The resulting capacity to transmit large
amounts of water rapidly is responsible for the development of
underground drainage and the disruption of valley systems.
Solution of this localised kind can produce voids which arc
penetrable so caves are more frequent, more elaborate, and larger
in karst than in any other terrain. In all branches of geomorpho­
logy there is substantial involvement with subsurface data, for
example, with the thickness of superficial deposits or the nature
of buried surfaces, but only in karst can one truly speak of ‘under­
ground geomorphology’ wherein a wide variety of erosional and
depositional forms demand survey and explanation. This aspect
of karst therefore requires serious consideration in this book
though the balance of treatment between external and internal
forms must favour the former, whereas in works of speleology
the reverse is properly the case.
Precipitation from natural waters is important in several mor-
phogcnic systems; witness salt pans in deserts and laterites in
seasonally humid tropics. But deposition from solution looms
 The Nature of Karst 3

very large in karst where its significance also varies with climate.
Along the coasts of karst country solution and deposition of
carbonate make for distinctive assemblages of landforms. This
aspect of the subject will not be dealt with here and readers are
referred to the book on Coasts in this series (Bird 1968).

TYPES OF KARST

Apart from coastal karst very many kinds of karst have been
recognised, some of which will be discussed later. However,
certain distinctions are so general that it will help to touch on
them now.
The most common contrast made is that between bare and
covered karst where bedrock is largely exposed to the atmos­
phere in the former (PI. 33), modestly or not at all in the latter
(PI. 22). Clearly there will be every stage of transition between
the two. The cover may be residual soil and organic matter
accumulating in place, or transported unconsolidated deposits of
aeolian, glacial, fluvial and other genesis. J. F. Quinlan (unpub­
lished paper) proposes subsoil and mantled karst to distinguish
the two, though it is essential to retain the general term because
it is often difficult to discriminate between them and they cer­
tainly occur together like a patchwork quilt. Karst features
beneath transported materials may be older or younger than their
cover whereas those associated with residual soils are syngenetic,
developing as insoluble and organic fractions accumulate.
Karst features, which may or may not affect the surface as
mentioned above, also develop beneath other bedrock formations
(Penck 1924). The term ‘covered karst’ is not normally employed
for this condition, for which the term subjacent karst following
Martin (1965) is adopted here. The overlying rocks may be other
marine beds and the contact may be conformable or unconform-
able. Alternatively they may be terrestrial outpourings of lava, for
example. With conformable relationships there is no doubt that the
karst features are younger than the deposition of the overlying
strata. This is essentially true, for example, of the karst of the
Carboniferous Limestone on the northern flank of the South Wales
Coalfield, in which large cave systems have developed under the
4 Karst

overlying Millstone Grit. The surface of the Millstone Grit has

also acquired a karst landscape as a result (Thomas 1963).
When the strata are unconformable, however, there is always
the possibility that some features may be inherited from the time
before the overlying rocks were laid down. Surface karst forms
may be overlain by later geological formations and caves com­
pletely filled by this process. This buried karst is also termed
fossil karst or palaeokarst. Quarrying exposed a cave completely
filled with Devonian rocks near Devonport, Tasmania (Burns
1964) and by exploring old lead mines Ford and King (1966)
have shown that a portion of the large Golconda Cave in the
Peak District, England, had been filled with Permo-Triassic
mineralised breccia and Tertiary quartz sands. Infilled closed
surface depressions may come to light through quarrying and
their fillings can range over long periods of geological time
(Gilewska 1964; Ford and King 1969). Natural erosion also can
expose fossil karst features. This gives rise to exhumed karst-, it
is more likely to happen and to be more readily detected with
projecting, rather than recessive forms. In central Manus Island,
New Guinea, conekarst (p. 184) is being uncovered by removal
of Pliocene lavas (P. W. Williams, pers. comm.). Whether the
reef knolls of parts of the Lower Carboniferous karsts of the
British Isles belong here is debatable. Most of these knolls repre­
sent submarine biohermal masses of limestone, that is construc­
tional features, subsequently stripped of the weaker rocks which
buried them, for example argillaceous limestone in east central
Ireland, whereas some which underlie an unconformity are
probably subaerially eroded features that have been similarly
uncovered and represent true exhumed karst, for example the
knolls of the western Peak District such as Chrome Hill.
Without having been buried, some karst forms are reliet from
previous morphogenic conditions no longer operative; thus
Gavrilovic (1969) maintains that in southern Yugoslavia there is
karst inherited from Tertiary tropical climate. The names palaeo­
karst and fossil karst have been applied in such cases also. The
former agrees with usage of the wider term, palaeoform, but the
latter is less satisfactory in that ‘fossil’ usually implies interment.
Another fundamental distinction lies between karst that is
completely surrounded by impervious rocks and that which can
drain directly to the sea without the intervention of different
hydrogeological systems. The French call the former karst harre,
 The Nature of Karst 5

which can be rendered impounded karst as opposed to free karst.

The impounded condition has most consequences when the karst
area is small, and very much affected by runoff from surrounding
non-karst terrain.

PSEUDOKARST

There are also landforms and even landform assemblages which

resemble karst forms but are the product of different processes.
This pseudokarst is found in sediments such as the finer pyro-
clastics and loess as a result of piping. Here subsurface mechanical
eluviation of clay particles leads to cavitation and to subsidence.
Small caves may result such as those on Officer’s Cave Ridge,
Oregon, in altered tuff.
More obviously pseudokarstic are larger caves of several types
and collapse depressions which occur in volcanic regions through
liquid segregation in layered lava and so also are the caves, natural
bridges, and melt dolines in glaciers formed by melting of ice
(glacier thaw karst). Thermokarst, in which melting of ground ice
(permafrost) in unconsolidated sediments occasions subsidence
and other forms, is less readily distinguished from true covered
karst. These different kinds of pseudokarst are discussed in Ollier
(1969) and in Davies (1969) in this series.
This label of pseudokarst has been applied to rounded grooves
in granite, reaching largest dimensions in the humid tropics. The
exact nature of the processes fashioning these Silikatkarren is not
known but despite the low solubility of granite it is probable that
they are due to solution. Even chemically inert rocks can develop
minor features convergent with those of karst as exemplified in
quartzite in the Wonderland of the Grampian Ranges, Victoria,
and in the Carrao basin, Venezuela (White and others 1966).
Priesnitz (1969) makes a strong case for the inclusion of all
these in karst proper.

HISTORY OF KARST STUDY

However far back in time one traces embryonic geomorphological

notions, karst participates throughout the historical development
of the subject. A bronze engraving records graphically the results
of the first known expedition to investigate karst phenomena sent
6 Karst

by the Assyrian King Salmanassar III to the springs and caves at

the source of the Tigris River; the writings of several Greek and
Roman classical poets, philosophers, and natural historians des­
cribe karst forms from the limestone which abounds in the
Mediterranean. As in most matters the Middle Ages added little
to the knowledge of the ancients although inscriptions in the
famous Postojna Cave in Slovenia date from the thirteenth
century onwards. With the Renaissance direct observation of
nature came to supplement quotations from the classics, though it
was the economic search for minerals in caves which led to most
of the new knowledge. The seventeenth century, the century of
the first scientific societies, saw also the appearance of books
devoted to karst, the first being that of Jacques Gaffarel of Paris
in 1654, of which little survives unfortunately. In 1689 the
impressive work of J.W.F. von Valvasor on the Slovenian karst
appeared with many maps of caves. The next developments in
karst study lay in the less central direction of investigating the
fossil content of cave deposits, culminating in the writing of
G. L. Cuvier and the Reliquiae diluvianae of R. Buckland. The
first really systematic geological and hydrological investigation of
a karst area may be attributed to Adolf Schmidl on the Postojna
region about the middle of the nineteenth century.
The classical period in the development of geomorphology of
the second half of the nineteenth century and the beginning of the
twentieth was equally fruitful in karst investigation and in the
emergence of scientific societies and caving clubs to prosecute it.
In this time Edouard A. Martel’s immense field and literary efforts
brought France alongside Austria in the forefront of karst studies
where it has since remained, and he also stimulated the English
into serious activity. Franz Kraus’s Höhlenkunde (1894) marked
an epoch in the German literature and in 1893 appeared Das
Karstphänonien of Jovan Cvijic, the greatest of the Yugoslavian
karst geomorphologists. A. Grund’s Die Karsthydrographie
(1903) provoked the most bitter phase of controversy over
underground water circulation in karst, which had older roots and
which persists in milder form to this day. In Czechoslovakia,
K. Absolon is chiefly remembered for the stimulus he gave to
Moravian karst investigation, whereas J. V. Danes from Bohemia
was the first karst specialist to range the world in this pursuit, and
pioneered the study of tropical karst. In Belgium, E. van der
Broek, in Romania, E. G. Racovitza, and in Italy, O. Marinelli,
 The Nature of Karst 7

were in this period the most significant contributors in their

respective countries to this flowering of the subject.
After World War I karst studies greatly intensified in coun­
tries where they were long established and spread to many others
so that it becomes even more invidious to mention individuals.
Institutes solely for karst and cave research came into being now
—at Vienna, with which the name of Georg Kyrie, the author of
Theoretische Speläologie (1923) and perhaps the first university
professor of speleology in the world, is closely linked, at Cluj in
Romania founded by Racovitza, and at Postojna, at this time part
of Italy, where G. A. Perko was director. Only since World War
II has there been a real spread of scientific study of karst away
from the main centres of Western learning. International con­
gresses of speleology began in 1953 and their proceedings are
publications of rapidly growing importance. Between 1954 and
1964 the International Geographical Union had a standing Com­
mission on Karst Phenomena under the chairmanship of Herbert
Lehmann of Frankfurt am Main, who had fostered the climatic
morphology approach to karst since 1936 with his study of Javan
limestone relief. The Frenchman, Jean Corbel, was also an inno­
vator in this aspect, particularly in the measurement of rates of
limestone removal in varying climates. During and since World
War II there has been great elaboration of the techniques of cave
exploration, one of the necessary bases for the understanding of
karst. The major stimulus here came from the French with the
name of R. de Joly outstanding.

KARST LITERATURE

This brief history will have made evident one of the several
difficulties about karst literature for the English speaker, namely,
that the contribution of the Anglo-Americans has not loomed as
large in this field as in practically every other aspect of geomor­
phology. In recent decades a surge of interest in the scientific side
of karst has enlivened the older sporting interest in Britain and the
United States, and has begun to redress this balance. However, it
is still true that very many of the important books and serials are
not in English. Moreover the really significant literature ranges
beyond French and German into languages less well known to
English speakers to a degree not common in science at large.
8 Karst

In particular, by reason of the richness and of the long history of

investigation of the Dinaric karst, an important sector of writing
in the south Slav languages lies behind a very real language
barrier. Fortunately J. Cvijic’s mature view of the subject is avail­
able to us in a posthumous work (1960) in French and a most
helpful critical review of the Yugoslav contribution for most of
the time since Cvijic’s death has been prepared by A. Blanc
(1958). A third difficulty arises from the fact that a great propor­
tion of important results has appeared and is appearing, outside
the geological and geographical serials familiar to the geomor­
phologist, in speleological literature, less readily available in
libraries, often local in approach and one in which many journals
have suffered great variations in effectiveness and too frequent
early demise. Books on speleology are usually easier of access and
good introductions to karst are found within their pages, e.g.
Cullingford 1962; Trombe 1952; Moore and Nicholas 1964;
Geze 1965; Trimmel 1968.

TERMINOLOGY

Additionally it must be admitted that no branch of geomor­

phology is so troubled by terminology as this. No other kind of
terrain seems to have bred such a multiplicity of regional terms
so that reaching agreement on a common vocabulary within the
frontiers of one country or one language is a major task in itself
prior to that of securing international equivalence. However,
vocabularies thrashed out by representative committees have been
published for German terms (H. Trimmel, ed., Speläologisches
Fachwörterbuch, 1965) and for French (P. Fenelon, ed., Voca-
bulaire frangais des phenomenes karstiques, 1968). In English the
situation is less formalised but there is a useful glossary in British
Caving (Cullingford 1962) and in a more convenient form we
now have A Glossary of Karst Terminology by W. H. Monroe
(1970). FAO is preparing a glossary of karst hydrogeological
terms.

METHODS OF INVESTIGATION

Most field and laboratory methods in geomorphology find appli­

cation in the realm of karst. Additional techniques are commonly
employed in the study of caves and karst hydrology.
 The Nature of Karst 9

The survey and cartography of caves requires some adaptation

of surface methods (Butcher and Railton 1966; Symposium on
Cave Surveying 1970). Radiolocation has come into use as a
check on cave surveying from the surface, and geophysical
methods such as soil resistivity and gravimetry may help in the
discovery of new caves. Cave diving is also a specialised tech­
nique, more valuable for the way it permits examination of com­
pletely waterfilled caverns where morphological development goes
on in different fashion than for the access it gives to ordinary
caves beyond water barriers.
The tracing of underground drainage connections is special to
karst study and a variety of means of tagging water has been
employed— salts, dyestuffs such as fluorescein, rhodamine and
pyridine, and exotic spores (Drew and Smith 1969; Symposium
on Cave Hydrology and Water Tracing 1968). The use of radio­
active tracers presents health hazards but naturally occurring
‘environmental isotopes’ are now being put to use. Determination
of other natural chemical and physical parameters of karst waters
has long been used comparatively to ascertain underground links,
but is now more valuable in establishing the rate and distribution
of solution (Douglas 1969). Artificial flood pulses as well as
natural flood pulses are now being monitored as a tool in making
inferences about unknown systems behind karst springs.
 II

KARST ROCKS

Many unsolved problems in karst geomorphology may arise from

inadequate appreciation of differences between the rocks involved
(Sweeting 1968). There is more variety under ‘limestone’ than
any other common rock name. The usual definition is that at least
half of the rock is made of carbonate minerals, mainly in the form
of calcite (CaCOQ. This is satisfactory for karst study since the
60 per cent of CaCO;{ estimated by Corbel (1957) to be neces­
sary for any significant karst development is probably of the right
order. He also considered that a purity of 90 per cent or better is
necessary for its full development, though this does not always
result as with chalk with its 95 per cent content of CaCO;}.
These points indicate the need for more detailed knowledge of
lithology than is usually accorded karst studies.
Besides calcite, the minerals common in carbonate rocks are
aragonite, dolomite, magnesium carbonate, and the impurities—
chalcedonic silica (chert and flint), detrital and authigenic quartz,
authigenic feldspar, and clay minerals (chiefly illite and kaolinite).
Lesser impurities include siderite and derived oxides of iron,
glauconite, collophane, pyrite, and bituminous matter. Quartz and
clay are the commonest impurities. The limits adopted to define
mixed rocks vary; those of Leighton and Pendexter (1962) are
set out in Fig. 1. Marl is an ambiguous term used both for cal­
careous clays and very pure lacustrine limestones. Magnesian
limestone is generally taken to mean limestone with an unusually
high content of magnesium carbonate (though still never more
than a few per cent).

LIMESTONE
Although terrestrial and freshwater limestones are locally
10
 K a rs t R o c k s 11

IMPURITIES
100%

In place of (impure) use

quartzose (sandy), / MIXED \
argillacepus (marly), / ROCKS N
WITH SOME
glauconitic, CARBONATE
cherty, etc
(eg calcareous '
sandstone dolomitic shale)

A / (IMPURE) (IMPURE)

CALCAREOUS DOLOMITIC

"<P/ DOLOMITE LIMESTONE

CALCAREOUS DOLOMITIC
DOLOMITE LIMESTONE

DOLOMITE % CALCITE
Dolomite as % of carbonates

1 L im its f o r m ixed carbonate rocks o f L e ig h to n and Pendexter 1962

important and in Australia loom large even on the continental

scale in the form of aeolian calcarenite (calcareous dune lime­
stone), most limestone is of marine origin. Many materials—
detrital, chemical, and organic— go to make them up. They
frequently undergo much change after deposition at low tempera­
ture and pressure (diagenesis) and may also be altered by high
temperature and pressure (metamorphism) to marble, which is
made up of large, clear calcite grains. Because of this com­
plexity, there have been many attempts to classify them. Modern
classifications hinge mainly on texture and as this has consider­
able bearing on response to karst processes, the following brief
account rests on those of Folk (1959) and Dunham (1962). The
ideal procedure would be to devise a special classification solely
for geomorphic purposes.
A few limestones are the result of organisms growing together
so that their calcareous skeletons are bound into a rigid frame­
work. The voids between may be filled later on by materials washed
in or by chemical precipitates. These autochthonous fossiliferous
limestones are called biolithites by Folk (his Type IV , Fig. 2) and
boundstones by Dunham. According to their dominant life form,
they may be named algal, coral, stromatoporal, or bryozoal. In
addition they assume two major dispositions— compact massive
12 K arst

IntracLasts Oolites Fossils Pellets

\
ALLOCHEM GRAINS
/ /
In tra s p rite
In traspa rru dite
B iom icrite Oo sp arite
B io m ic ru d ite O o sp arr udite
P elm icrite B io s p a r ite
B iosparrud ite
P elsparite

M ic rite s
MICROCRYSTALLINE ^ ________ 5 0 % _________10% SPARRY CALCITE
CALCITE MATRIX CEMENT

I n t r a c la s t ic O o li ti c Fossilif erous P e l le t a l

S parry
ALlochemical
rocks
I j iffi
In tra s p a rite & O o s p a ri te & B iosparite & P e L s p a rite
In tra s p a rru d ite O o s p a r r u d i te B io s p a r ru d ite

M icro crystalline
c a l c i t e ooze
n
M ic r o c r y s ta llin e
S parry ALlochemical
v -^nX v I c a l c i t e cement rocks
B io m ic rite & P e lm ic rite
B io m icru d ite

IV
M ic r o - Autoch -
c ry s ta llin e thonous
rocks rocks
M icrite D is m ic rite B io lith ite
( B io h e r m ,
B iostro m e )
2 M a jo r kinds o f lim estone according to F o lk 1959

mounds or thick lenses which are called bioherms, and extensive

tabular deposits or biostromes.
A t an opposite extreme, there arc limestones formed entirely
or nearly so of lime mud or ooze, accumulating more or less in
 Karst Rocks 13

place on calm, shallow sea floors. These are called micrites by

Folk (Type III, Fig. 2) and lime mudstones by Dunham. These
microcrystalline rocks are the aphanitic or lithographic lime­
stones, and calcilutites of older classifications. The lime mud may
be disturbed whilst still soft by slumping, and by boring animals;
voids so created get filled with precipitated sparry calcite. These
rocks are Folk’s dismicrites.
Very many limestones are more complex than these end
members and consist of a framework of larger grains with a matrix
in the intergranular voids, either of dull lime mud or of sparry
(i.e. clear and lustrous) calcite cement.
The larger grains are transported or secreted aggregates, Folk’s
allochems, of which he recognises four important classes— intra­
clasts, oolites, fossils, and pellets. Intraclasts are detrital frag­
ments, eroded from weakly consolidated carbonate sediments and
included after transport in almost contemporaneously forming
limestone. These fragments are usually rounded and range from
sand to boulders in size. Oolites are spheroidal precipitates of
concentric or radial structure around a foreign nucleus, forming
where bottom currents are strong. Fossils in these rocks are not
bonded in growth but consist of discrete skeletal components such
as foraminifera, sponge spicules, corals, bryozoans, brachiopods,
and molluscs. Algae such as Lithothamnion are the most impor­
tant of all. Many of these organisms introduce aragonite into the
rocks but it soon inverts to calcite over geological time. Pellets,
well sorted and rounded aggregates of microcrystalline calcite, are
thought to be faeces of worms and other invertebrates.
Where bottom currents were weak or impersistent, lime mud
accumulated along with the larger framework constituents. In
these rocks (Folk’s Type II, Fig. 2) there is a great range in the
proportion of allochems to muddy matrix. In some the grains are
so far apart that they float in the mud, they are mud-supported;
these are the lime wackestones of Dunham’s classification. When
the grains are abundant enough to be in contact and support one
another, the rock is the lime packstone of Dunham.
Elsewhere currents were strong enough to pile the intraclasts,
oolites, pellets or fossils together and to winnow out the ooze.
These are all grain-supported rocks but the voids between may be
partly or fully filled in later by the precipitation of clear calcite.
14 Karst

These rocks comprise Folk’s Type I (Fig. 2) and Dunham’s lime

grainstones4
Both the cleanly washed and the muddy allochemica! lime­
stone groups can be subdivided on the basis of their dominant
framework constituents— intraclastic (detrital), oolitic, biogenic
(skeletal), and pelletal. Also they can be categorised in terms of
the dominant size of their large grains into the sand-sized cal-
carenites and the gravel- and boulder-sized calcirudites.
To cover all possible combinations concisely, Folk has intro­
duced a set of terms compounded of two or three elements. The
first element (intra-, oo-, bio-, pel-) depends on the proportions
of the framework constituents; the second element refers to the
nature of the material filling the voids in between (-micr-, if it is
lime mud; -spar-, if it is sparry calcite); the third element (-rud-)
is only used when the framework grains are larger than 2 mm,
otherwise a calcarenitic texture is implied. All the terms end in
-ite. Thus a clastic limestone consisting of oolites cemented by
sparry calcite is an oosparite, whilst a limestone composed of
bivalve shells of the order of 1 cm set in a matrix of lime mud is a
biomicrudite. The dominant fossil type may be added where
appropriate, e.g. crinoidal biosparite.
Folk does not include in that part of his classification lime­
stones formed of clasts of much older, consolidated calcareous
rocks; he designates these calclithites but thinks they are very rare.
Recrystallisation of both mud matrix and framework con­
stituents complicates the classification set out and indeed may
have proceeded so far that original depositional textures are no
longer recognisable, except for some ‘ghosts’ of the allochems.
Dunham terms these simply crystalline limestones. Marble is often
loosely used for such unmetamorphosed rocks.

REEF FACIES

Many marine limestones have been laid down in association with

‘coral’ reefs— wave-resistant projections from the sea floor built

1 The definitions of Folk’s sparry allochemical limestones and Dunham’s

lime grainstones do not coincide exactly. Neither authority isolates lime­
stones consisting entirely of allochems and voids. Yet these are highly porous
and likely to behave in a special way in karst. Coquinas—shelly lime­
stones—fall into this category.
 Karst Rocks 15

3 (a) Devonian reef facies in Napier Range, West Kimberley, Australia

After Playford and Lowry 1966.
(b) Present relief.

by colonial organisms, and characteristic facies patterns of

different lithologies develop in such contexts (Playford and Lowry
1966).
The reefs themselves, whether atolls in oceanic situations, linear
and patch reefs in barrier complexes or fringing reefs, are massive
bioherms with varying fossiliferous composition (Fig. 3). How­
ever, the bioherms often form only a low percentage of the whole
reef structure. Cavities in them are numerous initially but these
tend to fill with intraclasts, ooze, fossils, or terrigenous sands.
Recrystallisation is common and so is replacement of the aragonite
of many organic skeletons by calcite. Dolomitisation is also
characteristic but irregularly patchy.
The steep seaward flanks are usually subject to violent wave
action and here submarine talus accumulates in forereef beds with
steep primary or depositional dips up to 30-35°. Characteristic­
ally this forereef facies consists of intrasparrites and intrasparru-
dites. The steep submarine slopes are unstable and subject to
slumping when still only partly lithified. This process gives rise to
the megabreccias of Playford and Lowry (1966)— intrasparru-
dites with many large blocks greater than 3 m across.
16 Karst

On the protected sides of the reefs, particularly in lagoons, the

backreef facies accumulates as nearly horizontal, well-bedded
deposits. Close to the reef, clastic material from it may be washed
over to accumulate at modest dips as intramicrites. Biostromes are
especially characteristic of the backreef facies and there are bio-
micrites, oomicrites, and micrites through the prevalent accumula­
tion of calcareous oozes in the quiet conditions.
These different facies in reef formations can be important in
karst areas as is evident from the Permian reef complex of the
Guadalupe Range in New Mexico, U.S.A., which includes Carls­
bad Cavern, and in the Devonian reef complexes of the Lime­
stone Ranges of West Kimberley, Australia.

DOLOMITE

Most dolomites appear to be secondary in origin through the

replacement of limestone; consequently they are dominantly
crystalline. They are best classified on the basis of the mode of
crystal size and of the nature of the surviving determinable
constituents from the limestone, which are usually the ‘ghosts’ of
intraclasts, fossils, oolites, or pellets. It is impossible to determine
the original proportions of ooze and sparry calcite. Dolomites with
crystals finer than 10 p,, i.e. dolomicrites, lacking relict limestone
textures and associated with evaporites, are thought to be of
primary origin, probably forming in very warm, hypersaline
lagoons or embayments.

EVAPORITES

Similar in origin to primary dolomite are the evaporites. There are

many evaporite minerals and these occur in complex mixtures as
rocks. However, easily the most important are halite (rock salt),
anhydrite, and gypsum.
Halite (NaCl) is a massive, coarsely crystalline rock, lacking
joints. It yields plastically at low pressure and temperature; conse­
quently it frequently occurs in the form of diapiric ‘salt domes’.
Clay and anhydrite impurities are usual.
Anhydrite (C aS04) is more frequently finely granular than
crystalline and occurs in thick beds or may be thinly laminated.
 Karst R ocks 17

Gypsum (CaS04.2Hi;0 ) is formed by hydration of anhydrite,

during which it swells 30-50 per cent and much crumpling may
result. It is usually finely granular, may be thin bedded or very
massive and jointless.
In the Permian Castile Formation of New Mexico and Texas,
all three occur together, accumulating to an exceptional thickness
of over 1200 m.

T ab le 1 Permeability of common rocks

Range o f permeability coefficients

R ock ( l/s e c /m ‘ with hydraulic gradient
o f 1 in 1)
U nconsolidated:
Clay 0-35 — 35
Sand 350,000 — 350,000,000
G ravel 350,000,000 — 350,000,000,000
In d u ra ted :
Shale 00035 — 3-5
Sandstone 35 — 3,500,000
Conglom erate 35 — 3,500,000
Limestone 3-5 — 350,000
Basalt 3-5 — 35,000
G ranite 00035 — 0-35
Gneiss 00035 — 3-5
Tuff 0-35 — 35,000
Based on J. P. Walz, G round water. Pp. 259-67 in Water, Earth and
Man, ed. R. J. Chorley. London, 1969.

PORES AND PLANES OF WEAKNESS: PERMEABILITY AND

STRENGTH

The response of karst rocks to geomorphic processes depends

very much on their permeability and mechanical strength as well
as their purity, and these in turn hinge in large measure on poro­
sity and planes of weakness. These interrelated factors will be
discussed together for convenience and brevity.
Porosity refers to all the voids in a rock expressed as a fraction
or percentage of the bulk volume and is determined on specimens
in the laboratory. Permeability on the other hand is the capacity
 18 Karst

T able 2 Compressive strength of common rocks

Uniaxial Uniaxial
compressive compressive
Karst rocks Other rocks
strength strength
(bars) (bars)
Limestone (excluding 340- 3310 Shale 360-2310
chalk and coral Sandstone 120-2350
rock) Conglomerate“ 1660
Limestone, reef 6 0 - 350 Tuff 350-2620
breccia Basalt 810-3590
Marble 460-2380 Diorite 1550-3350
Dolomite 620 - 3600 Granite 1590-2940
Anhydrite“ 410 Gneiss 1530-2510
Quartzite 1460-6290
Suggested limits for terms
Very weak < 350
Weak 350 - 700 1Single rock sample
Strong 700 - 1750
Very strong > 1750
Based on D. F. Coates, Rock Mechanics Principles. Mines Branch
Monograph 874, Ottawa, 1967; J. C. Jaeger and N. G. W. Cook,
Fundamentals of Rock Mechanics. London, 1969; L. Obert and
W. I. Duvall, Rock Mechanics and the Design of Structure in Rock.
New York, 1967.

of the rock to transmit water and it is determined in the field

from bores and wells as well as in the laboratory. Table 1 sets
out the coefficients of permeability for some common rocks. These
measures vary in relation to the size and continuity of the voids,
and limestones range most drastically in both the controlling
factors and permeability.
The mechanical strength of rocks can be determined in various
ways in the laboratory. Table 2 sets out the uniaxial compressive
strength of some common rocks; this is the maximum stress
cylinders of rock can withstand under compression in a single
direction without rupture or bending. Tensile strength, the
maximum tensional force rocks can undergo without deformation,
is much harder to measure and varies much more from one
specimen to another. However, uniaxial compressive strength is
usually a reasonable guide to variation in other strength properties
of rocks, including the tensile strength, which is, of course, many
times less than it. It will be seen from Table 2 that karst rocks,
 Karst Rocks 19

especially limestones, are exceedingly variable in mechanical

strength, varying from very weak to very strong. But the
behaviour of large masses of rock in nature is even more variable
than these laboratory determinations suggest because of larger
scale inhomogeneities, especially the planes of weakness traversing
them.
Intergranular porosity (sometimes called ‘primary permeability’)
is a textural characteristic of the rock closely related to its sedi­
mentary origins and diagenesis, and any later metamorphism. Bio-
lithites, skeletal and detrital limestones of many kinds can be
extremely porous because of the irregular shapes of many fossils
and some intraclasts, whereas micrites are usually devoid of pores.
However, initial pores may be filled completely with lime mud or
calcite cement, and very many limestones have virtually no
porosity of this type. Dunham (1962) points out that grain-
supported limestones are more likely than mud-supported lime­
stones to acquire intergranular porosity after deposition as a
result of differential solution and compaction in several ways. If
pore infilling does not take place during deposition or diagenesis,
it may well occur subsequently through recrystallisation (tending
to crystalline carbonates) so that the older the limestone the more
likely it is to be non-porous. Martel (1921) gives an intergranular
porosity range of 009 to 2 55 per cent for the ‘limestones’ of
France and 144 to 439 per cent for the chalks (in northern
England chalk can reach 46 per cent, Kendall and Wroot 1924).
The mechanical strength of limestones is closely related to
their porosity and consolidation; it greatly affects surface slope
development in karst and even more the possibility of significant
cave development. Weak limestones may collapse into incipient
caves and this may in turn affect the nature of the hydrologic
system. This factor probably dominates the evolution of chalk
country. Chalk is mainly a pure biomicrite, very porous and of
earthy nature, lacking strength. It develops into gently rounded
relief, with very rare crags at the points of pronounced river or
marine attack, characterised by extensive dry valley systems and
few caves, though some are known in both France and England.
Water infiltrates rapidly into chalk surfaces but does not escape
so readily; underground circulation is dependent on joints as in
most limestones (Brown 1969; Ineson 1962).
The fewer the intergranular voids, the more cemented and
20 Karst

compacted the limestone, the more important become its planes

of weakness, which promote permeability through their solutional
enlargement (‘secondary permeability’ of some). In fact the
nature, frequency, and pattern of planes of weakness probably
constitute the most important single factor of structure (in the
geomorphic sense) in karst. Indeed Tricart (1968) has argued
that a combination of high intergranular porosity and poverty in
joints is responsible for poor karst development in Barbados
despite very soluble rock, heavy rainfall, and lush vegetation,
through failure to canalise underground circulation.
Bedding in limestone may involve (1) dividing surfaces only,
(2) variation from top to bottom of beds, with greater purity at
the bottom, or (3) lithologic differences between succeeding strata.
Cyclic changes in the beds are common in limestone sequences.
Thin bedding is usually unfavourable to karst because it is often
accompanied by insoluble terrigenous constituents in the parting
planes and by shale or clay interbeds, which may block incipient
cave and underground drainage development. Moreover, it may
also lead to reduced strength with equivalent effect. With thick
beds, the bedding planes are conducive to cave development. In
the English Peak District, according to Warwick (1953), caves
are preferentially fashioned in massive bioherms.
Joints are planes of weakness favourable to karst, though
rewelding of joints by solute precipitation to minimise their
geomorphic effect is common in carbonate rocks. In these rocks,
joints are usually due to release of strain energy, residual from
earlier compression, during later uplift which permits extension
under load (Price 1959; Hancock 1968). Because of this domi­
nant origin, most joints are normal to bedding planes. Minor joints
lie within a single bed whilst major joints cut several. They usually
occur in parallel sets and frequently two sets intersect, commonly
at about 60°, in a conjugate joint system (PI. 9 ). Joints are (1)
latent— thin and hairlike, only apparent when the rock is broken
up, (2) closed—visible and allowing descent of capillary water,
and (3) open— allowing free water movement.
Joints may be too close set for adequate rock strength for cave
formation. This may be particularly true of cleavage joints, close-
set planes of weakness at right angles to compressional forces,
which usually make the rocks too weak mechanically for good
cave development as at Shatter Cave in the Goodradigbee River
valley, New South Wales, so named because of its instability.
 Karst R ocks 21

1 Stylolites weathered out to give appearance of flaggy bedding. Lower

surface smoothed by subsoil solution and exposed by soil removal.
Oligocene limestone, King Country, New Zealand.
22 Karst

Opposite views have been expressed about the role of faults in

karst development (Stringfield and LeGrand 1969a). They are
frequently accompanied by much cleavage; mineralisation is
another accompaniment which will influence their geomorphic
role. In soft chalk, plastic material injected into a fault zone can
make it a partial barrier to groundwater flow (Ineson 1962).
Factors such as these may be responsible for their varying effects
on underground drainage, doline and cave development.
Stylolites may be so prevalent in limestones and dolomites as to
affect their relief expression significantly. These are sutures in the
rock where pressure solution has taken place, often leaving thin
laminae of insolubles. The solution may have taken place along
bedding planes or elsewhere in the rock. Recementation takes
place often enough and the stylolites may become the resistant
parts of the rock (Pluhar and Ford 1970). The widespread
Oligocene shell calcarenites of New Zealand have many residual
seams of this nature as a rule (Barrett 1963), and this gives rise
to a ribbing in outcrop, which appears geomorphically to resemble
lensy, flaggy bedding, with the bedding planes projecting (PI. 1).
 Ill

KARST PROCESSES

Karst as defined in this book— namely as a type of landscape with

distinctive landforms that arise primarily from abnormally high
solubility of the bedrock— includes terrains in which processes
such as mechanical action of rivers and frost shattering have
played significant, even dominant parts, though these processes
are in no way special to karst. By employing such a definition,
unrealistic standpoints are avoided such as that of Panos and
Stelcl (1968) who excluded areas such as the Sierra de los
Organos, Cuba, from karst, even though solution played a par­
ticularly great role in producing distinctive landscape there. Never­
theless, only those processes or aspects of processes assuming
peculiar importance in karst will be discussed here.

SOLUTION AND PRECIPITATION

Limestone
Since limestone is the most widespread karst rock, its solution
and deposition are the most important processes to consider. As
calcite, it is only modestly soluble in pure water, at saturation the
amount varies from about 13 mg/1 at 16°C to about 15 mg/1
at 25°C (Frear and Johnston 1929). Aragonite, a much rarer
mineral, is about 16 per cent more soluble.
However, much greater concentrations than these are common
in natural waters. Other solutes are responsible for this and varied
lines of evidence mustered by many authors have shown that
usually the most important is carbonic acid (e.g. Adams and
Swinnerton 1937; Smith and Mead 1962; Gross 1964; Pitty
1966).
23
24 Karst

Much less work has been done on the role of organic acids
resulting from rotting of vegetation such as lactic acids. Yet there
can be no doubt about the importance of organic acids regionally
as Williams (1970) has pointed out in relation to the widespread
contact of peaty water with limestone in Ireland, and perhaps
generally (Muxart and others 1968). Chelates are other organic
compounds that have an action similar to that of acids on lime­
stone (Keller 1957).
Of more localised importance, sulphuric acid produced by the
weathering of sulphide minerals such as pyrite and marcasite,
sometimes from interbedded shales, is a powerful solvent of lime­
stone (Morehouse 1968; Pohl and White 1965). Another source
of sulphuric acid may reside in the physiology of certain iron­
fixing bacteria (e.g. Crenothrix, Callionella). Nitric acid in rain­
water as a result of lightning is apparently negligible as a lime­
stone solvent.
Because of the paramount part played by carbonic acid, our
main concern is with the carbon dioxide-water-calcium carbonate
system. This will be set out here in a simplified way though it has
a much more complicated physical chemistry than is evident from
most geomorphology texts. It is not just a matter of a single
reaction to produce more soluble bicarbonate from less soluble
carbonate but of a series of reversible reactions and ionic dis­
sociations each governed by different equilibria (Bögli 1960;
Roques 1964, 1969; Thrailkill 1968).
Dissolved calcium carbonate is in an ionic state:
CaC 03 ( s o l i d ) Ca-+ ( h y d r a t e d ) "j- (hydrated) • • • (1)

and the product of the two kinds of ion is a constant. The warmer
the water and the more agitated its motion the less time solution
takes to reach the saturation equilibrium values already given;
the nature of the limestone surface will also govern the rate of
solution. For greater amounts of limestone to be taken into
solution than the equilibrium for the temperature, a further
chemical reaction has to take place with carbonic acid.
Natural waters have some carbon dioxide in solution; the
equilibrium amount increases with rising partial pressure of
carbon dioxide (Pco2) in the air in contact with the water and
falls with rising water temperature. As temperature rises from 0°
to 35°C, the amount at saturation drops between one-third and
two-thirds. These relationships constitute Henry’s Law. A small
 Karst Processes 25

part of the dissolved CCL reacts very rapidly with the water to
produce carbonic acid which is always in an ionic state:
CCL (dissolved) H^O H + -j- HCO" . . . (2)
Carbonate ions from the dissolved limestone react instantaneously
with the hydrogen ions to produce bicarbonate ions:
CO2-3
+ H+ — HCO2- 3
. . . (3)
This last reaction upsets the equilibria of (1) and (2). More
limestone goes into solution to keep the product [Ca2+ ] [CO2 -]
constant and more dissolved carbon dioxide reacts with water to
produce more carbonic acid. This phase is rapid but it only
achieves a modest additional solution of about 8 mg/1 in normal
atmosphere.
However, this last phase produces a disequilibrium between
the carbon dioxide partial pressures of the air and of the water.
There follows a diffusion of C 0 3 from the air to the water, in
turn permitting further solution of limestone through the chain of
reactions. The complete process of limestone corrosion may there­
fore be simplified in the following way, reminding us of the various
stages:
C aC 03 (so lid ) + H 2 O -f- C O 2 (d isso lv ed ) ^ Ca_^ -)- 2HCO^_
u
C O 2 (a ir )
Diffusion of C 0 2 through water is a slow process compared
with the earlier stages. It varies with temperature, proceeding
faster the higher the temperature, but more important is the speed
and turbulence of water movement (Weyl 1958). These are thus
the chief controls of the rate at which limestone solution takes
place. However, this final phase of C 0 2 diffusion can only happen
where the water is in contact with air (an open system); it
cannot do so where water is confined between rock alone (a closed
system).
The total amount of limestone which can be dissolved at
saturation equilibrium per unit volume of water is overall a direct
function of the carbon dioxide partial pressure of the air with
which the water is in contact and an inverse function of the water
temperature because of the latter’s control of the dissolved C 0 2
saturation equilibrium (Henry’s Law), see Fig. 4. The tempera­
ture effect is, however, relatively modest— about a threefold range
results, whereas the partial pressure effect is much greater— at
least a hundredfold range (Ek 1969). In the free atmosphere, the
26 K a rs t

450 -i

SUPERSATURATED

350 -

300-

E 250-

UNDERSATURATED
( Aggressive )

75 ToÖ 125 150 17?

Equilibrium C02 ( m g / l )

4 Saturation e q u ilib riu m curves f o r so lu tio n o f calcium carbonate at

d iffe re n t tem peratures as a fu n c tio n o f e q u ilib riu m carbon d io xid e in
solution. A fte r T ro m be 1952.
I f saturated w ater at 250 m g/1 C a C O j and 50 m g/1 C O : is cooled
(dire ctio n C), it can dissolve m ore lim estone; if it is w arm ed (W ), it w ill
th ro w some out o f so lu tio n . I f it comes in to contact w ith a ir w ith less
C O :(L ), it w ill lose some C O : and pre cip ita te C aC O g if it encounters a ir
w ith m ore C O -(H ), it w ill take in m ore C O : and dissolve m ore lim e ­
stone.

mean figure for Pco2 is quite small, 3 X 10~4 bar (0 03 per cent
of volume). Most cave air has similar values but it does vary signi­
ficantly (Holland and others 1964; Ek and others 1968). The air
in snow has about 0T per cent C 0 2 because the smaller oxygen
and nitrogen molecules diffuse out of the voids faster than the
larger C 0 2 molecules. Theoretically therefore snow and glacier ice
meltwater could be richer in C 0 2 than rainwater. But the really
significant amounts of C 0 2 are found in soil air, including that in
 K arst Processes 27

Vegetation litter. Values of 1-2 per cent are usual, but very much
higher values occur. Extreme quantities of 20-25 per cent have
been reported from poorly ventilated tropical soils. Root respira­
tion and bacterial decay of organic matter seem to be chiefly
responsible for this.
‘Biogenic’ carbon dioxide is therefore regarded by most investi­
gators as the prime control of ultimate limestone solution per
unit volume of water and it in turn is chiefly dependent on tem­
perature and rainfall which promote vegetation growth. Smith and
Mead (1962) have shown that the springs of the Mendip Hills
in southwest England have taken an average of 280 mg/1 of
calcium carbonate into solution and this agrees with saturation
equilibrium at prevailing temperatures for the average figure of
1-6 per cent carbon dioxide in the soil air beneath the grass
pasture of these hills.
The carbon dioxide-water-calcium carbonate system so central
to an understanding of karst has some important corollaries. If
water has reached saturation with respect to limestone, either at
the surface or at shallow depth underground, descends further
and is cooled in the process, as may happen in summertime in
particular, it becomes capable of dissolving more carbon dioxide

SUPERSATURATED
^ M 1:3

— 240-
? 200 -

f M 30'-1 UNDERSATURATED
W1

24 32 4Ö"
Equilibrium C 0 2 (rng/l)

5 Mixing corrosion. After Bögli 1964a and Thrailkill 1968.

Mixing of saturated waters W1 and W2 results in aggressive water.
M30:l (30 parts low Pco2 W1 and 1 part high Pco2 W2) is more under­
saturated than M l:l, an equal mixture. No mixture of supersaturated
waters W3 and W4 will be aggressive. Ml:3 (1 part saturated W1 with
3 parts supersaturated W4) remains supersaturated but M 3:l, a converse
mixture, becomes aggressive.
28 Karst

from cave air and thus of becoming aggressive to limestone once

more. Bögli (1964a) has termed this cooling corrosion.
However, he regards as much more important another
mechanism which he terms mixing corrosion (Fig. 5). Because
the relationship between C 0 2 partial pressure and calcium car­
bonate saturation equilibrium is an exponential one, the mixing
of two saturated bodies of water with different calcium bicar­
bonate concentrations produces water which is undersaturated.
It then becomes capable of renewed attack on limestone with
which it is in contact. This mixing is most effective in reviving
aggressiveness when a large body of water saturated in response
to low Pco2 mixes with a small body of water saturated at high
PcOo. This may well be the case when vadose seepage and vadose
streamflow (see Chapter V) mix. However, though the physical
reality of mixing corrosion is not in doubt, it may not be as
important geomorphically as Bögli thinks. For instance, one or
both of the waters mixing may be supersaturated and little or no
resumption of aggressiveness may result (Thrailkill 1968).
It is also misleading to ascribe too wide an importance to the
saturation equilibrium of the overall solution process since very
often natural waters will not equilibrate before escaping from
many geomorphic situations. The amount of limestone dissolved
is the product of the volume of water passing in a given time
through any karst situation and its actual carbonate concentration,
not the theoretical amount which could possibly be dissolved in
those conditions. The water volume is ultimately governed by the
surplus of precipitation over evapotranspiration and the CaCO:{
concentration usually depends on
1. the nature of the limestone surface;
2. the velocity and turbulence of the water flow;
3. the direct temperature effect on the rate of chemical reac­
tions (though this factor is only significant when comparing
karsts in climates very different thermally).
Less frequently will the Pco2 and temperature controls of satura­
tion equilibria be the arbiters from the geomorphic point of view.
Saturation comes into its own in the process of calcium car­
bonate precipitation. Many textbooks have stressed evaporation in
this connection, which undoubtedly does raise concentrations above
saturation equilibria to bring about deposition. But relative
 Karst Processes 29

humidity in most caves is so high and air movement so slow that

evaporation is there at a minimum. However, the various reactions
of the C 0 2-H20 -C a C 0 :i system are all reversible without the
intervention of evaporation. If water has acquired high carbonate
content in response to high Pco2, for example as a result
of passing through soil and thence through tight rock fissures
without air, it will diffuse the gas back into normal air if it
resumes contact with the latter. This brings about carbonate super­
saturation and precipitation of calcite. There can be little doubt
that this is the dominant process in many contexts, especially in
caves. Holland and others (1964) have shown this to be the case
in Luray Caverns, Virginia, where saturated seepage water
depositing calcite declines in calcium content whilst doing this but
maintains its magnesium concentration. If evaporation caused the
deposition, both magnesium and calcium would decline. A
further but less important cause of deposition is the warming of
water which has achieved saturation at lower temperatures;
saturation equilibrium for C 0 2 solution is lowered by this,
diffusion of C 0 2 occurs and calcium carbonate is precipitated.
Reprecipitation of C aC 02 is important not only in giving rise
to independent deposits and thus constructional landforms but
can be equally significant in case-hardening limestones and
calcareous sands (Monroe 1966; Jennings 1968). Induration by
reprecipitation of calcite in voids near the surface gives calcrete
or kankar which can be a vital factor in landscape development.

Dolomite
Dolomite rock behaves in natural waters in an essentially
similar way to that of limestone, though the various equilibria in
the solution of the double carbonate mineral, dolomite, which is
its chief constituent, have been less fully investigated (Holland
and others 1964). Under normal air and water interface condi­
tions, it is claimed that dolomite is usually the less soluble, as is
witnessed by recessive weathering of calcite veins in dolomite.
But the situation is more complicated than this (Douglas 1965).
With very high Pco2 more magnesium than calcium is dissolved
from dolomite as is also the case at very low Pco2, whereas at the
pressures frequent in soil voids and rock crevices, they are
dissolved about equally. The responses of subaerial and subsoil
surfaces of dolomite should therefore vary. Magnesium carbonate
30 Karst

which often occurs in carbonate rocks behaves in a complex way

also because it occurs in three mineral forms, each with its own
characteristic behaviour.
These complexities introduced by the presence of magnesium
make study of the geomorphic role of solution in karst difficult
because limestones are often partly dolomitic or magnesian. This
can be illustrated by the attempts which have been made to infer
from the Ca/Mg ratio of spring waters the nature of the rocks
feeding the spring (Hem 1959; Jennings and Sweeting 1963).
As we have seen this ratio will vary with the Pco2 of the
dissolving waters as well as with the lithology. Moreover the ratio
also varies as a result of differential deposition, calcite preceding
magnesium minerals in precipitation, whether this be the result
of evaporation or of carbon dioxide diffusion to the air (Douglas
1965). Impurities such as salt (NaCl) and metal trace elements
affect the carbon dioxide-water-carbonate system significantly in
natural waters and this may explain many difficulties found in
interpreting field measurements of carbonate content, tempera­
ture, and pH (Roques 1969).

Evaporites
The third most important karst rock, gypsum (PI. 2), is much
more soluble than either limestone or dolomite (Trombe 1952);
C 0 2 is not involved since it cannot react with either ion of
CaS04. Gypsum is most soluble at 37°C, a temperature above
that of most natural waters. Therefore what matters geomorphic-
ally is gypsum’s increasing solubility with rising temperature up
to that maximum. Conversely deposition results from the cooling
of natural waters and by evaporation causing supersaturation.
Essentially the same relationships apply with the chloride rocks
such as halite but these are more soluble still so that they can
seldom persist in the surface landscape.

PIPING

In Chapter I reference was made to piping as a cause of pseudo­

karst; it needs mention here because it also occurs in karst proper.
Piping or tunnelling occurs in clastic sediments and soils where
percolating water transports clay and silt fractions internally,
leaving underground conduits (Parker 1964). It is thus related
 Karst Processes 31

2 Steep gypsum karst below Pointe de la Jaquette, Bas Queyras, French

Alps. Soluble rock yields active gullies and crumbling spires.
32 Karst

more to eluviation than to solution, although the solution of

soluble grains in a soil or sediment may assist piping.
It can occur in material of any grain size. In coarser materials
the open fabric provides the necessary permeability. In fine­
grained materials, extensive cracking and/or fine dispersion are
necessary. It follows that the presence of clay minerals with high
swelling capacities (e.g. montmorillonite), and a high dominance
of Na and Mg over Ca in the exchange complex, are both con­
ducive to piping in fine-grained materials.
Therefore piping occurs in various soils and superficial deposits
on karst rocks. Indeed piping may be specially promoted by
solution beneath the covers providing openings for removal of
fines as well as by solution of residual rock fragments in some of
them. Piping is accompanied by subsidence1 with the development
of shafts and surface depressions. Since solution and collapse1
induce subsidence in covers, it is difficult to assess just how
important piping is in karst but it must not be overlooked.

SUBSIDENCE

Karst is favourable in some ways and unfavourable in others to

the mass movement of residual and transported mantles, and of
interstitial bodies of soil and sediment. Antagonistic is the
tendency of these materials to be drier than on impervious rocks;
other things being equal, downward percolation into the bedrock
allows covers to dry out more quickly and pronouncedly than on
other rocks where this will depend more on runoff and lateral
percolation. For this reason the more lubricated kinds of mass
movement—block (rotational) slump, debris slide, debris
avalanche, and debris flow (terminology of Varnes 1958)— are
proportionately less active in modifying karst landscapes than
others in the same climatic conditions. Because of the low per­
centage of insolubles in karst rocks, residual soils tend to be
shallower on the average than on many other rocks, minimising

1 In this book a distinction will be made between collapse— sudden mass

movement of the karst bedrock— and subsidence— mass movement, often
gradual, of soils, weathering mantles, and superficial deposits. With many
rocks, this would be a highly artificial distinction but with karst rocks the
dominance of solution usually ensures a sharp division in profile between
bedrock and regolith.
 Karst Processes 33

superficial mass movement. Furthermore the ready reversal of

the solution process leads to frequent cementation, reducing
mobility, e.g. in talus.
Working against these factors is the effectiveness and wide­
spread action of solution in removal of support in all kinds of
unconsolidated materials. Removal of lateral support is clearly
evident at springheads and at streamsinks, for example, though
the infrequency of surface streams reduces the opportunities for
undercutting of channel banks. Removal of underlying support is,
of course, the outstanding characteristic of karst in this context,
occurring where solution or collapse of the bedrock has under­
mined the covers at central points and induced centripetal mass
movement. This frequently takes place as creep, block slump,
and debris slide; debris avalanche, debris flow, and mud flow are
proportionately less important for the reasons given above. Dry
soilfall and slow earthflow on the other hand can assume unusual
importance. For example, dry soilfall appears to be the dominant
process in the Maiden’s Tresses Chamber of Easter Cave near
Augusta, Western Australia, where sandy loam soil has been fed
vertically from a choked solution pipe in virtually single grain
state to build a conical pile below. Very widespread is the gradual
descent of surface materials in slow flow or slide down joints and
other solution-widened planes in the bedrock, usually with a
modest degree of lubrication. Surface clays and loams gradually
incorporate blocks of the bedrock on their way down, producing
gash-breccia, a very important category of cave sediment. There
may be little or no surface evidence of the whole of this process.
Desiccation of covers is not entirely a restraining factor in
mass movement because with swelling clays it readily creates deep
cracking. Rapid infiltration from heavy rainfall down such cracks
is a common trigger of block slumping.

COLLAPSE

In no morphogenic system is collapse—essentially rockfall, block

slide and rock slide (Varnes 1958)— as significant as in karst for
several reasons. Bare rock slopes and cliffs, where these operate,
are more frequent since soils are thin. Solution acts as freely
laterally as downwards to a degree not true of other erosive
processes so that undercutting by streams is more effective in
34 Karst

karst than in less soluble terrain. Moreover solution produces

caves where roofs as well as walls are subject to fall. Neverthe­
less geomorphologists have shown little interest in this aspect of
karst dynamics until recently (Davies 1949, 1960; Montoriol-
Pous 1951, 1954; Renault 1967). Renault has ventured notably
into the difficult task of applying the laws of rock mechanics to
the complexity of natural situations in karst. In engineering
laboratories and even in tunnel or cutting construction, stresses
are applied instantaneously in terms of geological time whereas the
deepening or widening of a valley or the enlargement of a cave
proceed infinitesimally slowly in comparison; findings may not be
easily transferable from the one domain to the other.
Rocks are not uniform lithologically nor are they isotropic
structurally; bedding, jointing, and cleavage intervene. Therefore
shear strength varies complexly. Nor is gravity the only force
involved when a slope is undercut or a cave passage created; there
are permanent tectonic stresses in the rocks as well. Also decom­
pression by erosion opens up joints, which permit rockfall on the
surface and roof collapse underground. Changes in the surface
form or cave shape through solution, corrasion, and removal of
cave deposits gradually upset a previously established equilibrium
of forces until breakdown occurs to produce a fresh temporary
equilibrium. Draining of a waterfilled cave system will reduce
support quickly and this has been considered by Davis (1930)
and others as a potent factor in cave evolution.
Stresses in cliffs in the open and on cave walls and roofs tend
to drive them into space; falling and gliding of masses of rock
result. In the confined condition of a cave, there is the difference
that the continuity from wall to wall results in a local compression.
Tensional and shear components in this compression tend to
produce dome or arch forms, though how perfectly this will be
achieved depends on the consolidation, planes of weakness and
attitude of the rock and the pressure of the overlying rock, which
depends on depth below the surface. Below a critical depth, the
mechanical response of the rock will be different, a plastic regime
taking over from an elastic one. Here an elliptical form with a
vertical major axis should develop, though if collapsed material
on the floor is not removed, this will interfere with the system
of forces. This approximates to the interpretation Trimmel (1950,
1951) placed on the shape of certain Austrian caves in limestones
 Karst Processes 35

6 Effects of stress on cave cross-section

(a) Cave enlargement through spalling under high rock pressure at
depth. After Trimmel 1968.
(b) Stress zone around a shallow cave. After Davies 1960.

of great mechanical strength (Fig. 6a). However, the critical depth

is considerably more than 150 m postulated by Davies (1960) and
plasticity will not be relevant in most caves (Renault 1967;
Schoemaker 1948).
Halite and gypsum have greater plasticity than most carbonate
rocks so that collapse will occur more readily in these rocks
(Messines 1948). Chalk also can behave more plastically than
most limestones. Whether this helps to explain comparative
absence of caves in these rocks is not certain.
At shallow depths, arches are certainly produced by piecemeal
collapse as in the chalky bryozoan limestone of the Wilson Bluff
Limestone of the Nullarbor Plain, Australia (PI. 3), and in weaker
rocks still such as semi-consolidated dune limestone where domes
form, with the space they enclose remaining almost full of
collapse material (Fig. 7). With very strong rocks near the
surface, the rock structure intervenes pronouncedly and collapse
results in quadrangular cross-section with very low (McEachern
Cave, Victoria: Ollier 1964) or very high dips (PI. 38), and
triangular forms with intermediate dips. When a cave is very near
the surface, the roof behaves like a beam, which tends to sag into
the opening and eventually collapse occurs, opening the cave to
the sky (Figs. 6b, 38). Theoretically, prior to this release of
forces, there should also be an upward compression of the floor
of the cave. This has not been demonstrated in nature because
3 Roof dome and blockpile formed by breakdown in horizontally bedded
Eocene limestone. Brackish water piped to surface for sheep. Koonalda
Cave, Nullarbor Plain, South Australia. Photo by H. Fairlie-Cuninghame.

bedrock floors of any extent are rarely free from rockfall and
other cave fill.
In terms of observable process, the patterns of tensional and
shear stresses and of shear strength in cliffs and caves are for
practical purposes permanent but the incidence of rockfall and
rock slide is intermittent. Triggering actions are involved but little
study has been made of them. Additional transitory stresses at
the moment of earthquakes are obvious in this connection but
there is probably no reason to think they play a role more
important underground than on the surface or in karst than in
other terrain. Rock blasting in quarries or for road construction
or heavy vehicular traffic near caves could have this effect, though
the Dark Cave, Bukit Batu, Kuala Lumpur, Malaya, shows no
signs of collapse in the part exposed to regular tremor from a
quarry crushing plant close by. Hydrostatic pressure of water in
cracks may increase after rainstorms and act in this way. Other
possible triggering forces are the swelling on wetting of colloids
in slightly opened planes of weakness or of shale interbeds, and
the force of crystallisation of minerals, especially of the
commonest, calcite. Conversely it has been suggested that the
drying out of clay interbeds or breccia fills in fissures may reduce
 K arst Processes 37

U n s u p p o r te d diam eter

Tens ion c r a c k

0 Wt. of
Cave 1 ro of

1 Dome on p o in t o f collapse 2 Tension s ta g e of c o l l a p s e

Increased shear
s t r e s s due to
Shear Tension S hea r fa ilu r e of support
in ce n tr e

Collapse dome

3 Shear stage of collapse

Subsidence
doline
• Sandy
soiL •

K a n ka r

ENTRANCE
CHAMBER

F a l le n blocks
Aeolian
c a l c a r e n i te

15 M e tr es
50 Feet

7 Collapse dome formation in weak aeolian calcarenite

(a) Sequence of collapse after A. L. H ill {in Jennings 1968). S = shear;
T = tension.
(b) Collapse dome in Easter Cave, Augusta, Western Australia. Soil
depths from Lowry 1967a.
38 Karst

cohesion and shear strength. With caves near valley walls, retreat
of the latter decompresses the rocks and joints can open up as a
result and promote cave collapse. However, the probability is
that the commonest trigger will be the simple progress of solution
by seepage water along planes of weakness and stream flow under­
cutting walls.

OTHER WEATHERING PROCESSES

There is no need here to consider surface weathering processes in

karst other than the dominant and characteristic one of solution.
Since many weathering processes are dependent on large and
repeated changes in atmospheric conditions and on biological
activity, the equable atmosphere and the comparatively abiotic
conditions of most caves imply their absence from cave walls and
roofs in large measure.
Departures from this generalisation may be important locally,
however. Thaw-freeze action can penetrate into the forward parts
of caves, acting directly on their surfaces, and its products can
penetrate further still under gravity. Some caves are dry enough
for crystallisation by evaporation and there salt weathering can
play a major role in wall and ceiling sculpture, e.g. in some caves
of the Nullarbor Plain (Lowry 1964; Jennings 1967c and its Fig.
2:12). In biological weathering, there are doubtful claims about
bats scratching sizeable bell holes in cave roofs (King-Webster
and Kenny 1958; contrast Hooper 1958) but there is the
undoubted fact of corrosion of both bedrock and secondary cave
deposits by bat guano and possibly urine (Jennings 1963). More­
over recent recognition of the ubiquity of bacteria has led to the
idea that they may have both corrosional and depositional actions
in caves (Smyk and Drzal 1964), for example, in the formation
of moonmilk (Caumartin and Renault 1958).
 IV

MINOR SOLUTION SCULPTURE

Weathering etches minor forms of diverse nature in many rocks,

but the greatest variety is undoubtedly found in karst because of
the susceptibility of karst rocks to solution. In German, Karren
has come to be used as a comprehensive term for small-scale
solutional sculpture (Bögli 1960) and for some French investi­
gators lapiez has the same wide sense. Though there are transi­
tional and compound forms, the multiplicity of distinct types has
required further terminology such as Bögli has employed. Unfor­
tunately in English there is no general term equivalent to Karren,
and confusion and inadequacy rule amongst specific terms. A
partly new nomenclature will be adopted here since to use a
whole foreign terminology untranslated is neither euphonious nor
readily assimilable.

FACTORS AFFECTING MINOR SOLUTION SCULPTURE

Many factors interact to control the nature and pattern of small-

scale sculpture on karst rocks, and consequently that sculpture is
far from being properly understood. Perhaps the most important
control is the presence or absence of cover (soil, plant litter,
superficial deposits, and vegetation itself) because the conditions
affecting solution are very different in these two circumstances.
Nearly all of the limestone of the Peak District of Derbyshire,
England, is covered chiefly by aeolian mantles (Pigott 1962), so
there is virtually only sculpture at the subsoil interface. On the
very similar Carboniferous Limestone of the nearby Craven
district of Yorkshire, there is much more rock exposed and much
more obvious solutional activity. This is partly due to the last
glaciation stripping earlier weathering mantles from Craven
39
40 Karst

whereas the Peak District was not overridden by this ice sheet
(cf. a similar contrast within the Jura Mountains; Aubert 1969).
However, the country rock itself may inhibit the formation of
many kinds of detailed sculpture. Thus the porous, weak Creta­
ceous Chalk of southeast England and northern France is
unfavourable to nearly all such landforms and there is not such a
sharp interface between rock and soil with it as with most lime­
stones. The same effect is found from different causes on the
contact-metamorphosed Silurian limestone of the eastern side of
Cooleman Plain, New South Wales. It is a hard, coarsely
crystalline rock and weathers to rounded forms in outcrop, closely
resembling those of glacial abrasion. Joints generally find little
surface expression and any protected recesses carry a calcite gravel
skeletal soil. The readiness with which the rock disaggregates
operates against solutional sculpture.
Rapid changes of lithology or closely spaced bedding, jointing,
or cleavage planes can also prevent the development of many
forms. Increasing percentage of impurities affects the nature of
the forms that do develop, especially producing more rounded
edges and ribs.
In cold climates, thaw-freeze replaces solutional sculpture
altogether (Corbel 1954). Thus the marble of Mt Arthur, Nelson,
New Zealand, has an interesting range of solutional forms on out­
crops above the treeline, but in the final one to two hundred
metres below the summit there are only angular corners and platy
surfaces on bedrock, with active scree formation. Temperature
also exerts an influence through its effect on the chemical reactions
of the water films. A probable consequence is that many forms are
much longer downslope in tropical climes than in temperate ones.
The nature of precipitation enters into the matter, certain forms
being associated with snow cover, which must act rather like a
temporary (but biologically sterile) soil. The amounts, duration,
and intensity of rainfall critically affect solution of bare rock and.
through soil water, of subsoil surfaces. The most obvious precipi­
tation effect is the extreme one of aridity, which minimises and
deflects solutional activity. Solutional forms are few, for example,
in the Nullarbor Plain of Australia (Jennings 1967a). The
dominant result may be one of case-hardening of the exposed rock
surface through evaporation, leading to reprecipitation of much
of the modest amount dissolved. Accompanying this crust forma-
 Minor Solution Sculpture 41

tion there may be the development of tafoni as in the Hadhramaut,

Arabia (Wissmann 1957). The complexity of the matter is illus­
trated by the fact that, in a comparatively humid climate in
various areas of Canterbury, New Zealand, Oligocene limestone,
chiefly a porous shelly and foraminiferal biomicrite, also weathers
to a rounded, hardened outer surface, broken by large tafoni; this
is in part a lithological effect.
Historical factors involving climatic change have been inade­
quately evaluated, but may be important in some areas. In the
Limestone Ranges, West Kimberley, Australia (Jennings 1969),
Devonian reef rocks exhibit a tremendous array of bare karst
solutional sculpture, yet the rainfall is a modest 450-700 mm
annually with evaporation of the order of 2500 mm. Though the
minor forms are so sharp they must still be developing, it may be
unrealistic to attribute them wholly to a short season of a few, if
intense falls of rain. The question arises whether Pleistocene
pluvials (of which there is as yet no local evidence) were respons­
ible for their creation, with solution since then sufficient only to
prevent their degradation by other processes.
A factor which explains many patterns of sculptural forms as a
result of exposure of subsoil surfaces through soil erosion is the
destruction of forests by man and his grazing animals, loss of
plant litter by oxidation on exposure, and loss of soil and litter
by accelerated erosion consequent on agricultural activities. These
effects can be seen to be operative today in areas where defores­
tation for grazing has recently been carried out or is still going on,
e.g. Mole Creek, Tasmania, and Te Kuiti, New Zealand. But they
are startlingly obvious where dense gardening populations fell
and burn tropical rainforest, e.g. in the Central Highlands of
New Guinea. With shifting agriculture, fresh subsurface forms,
starkly blanched, emerge from litter and soil season by season
(PI. 10). Once this process is seen in action, the reconstruction
of prehistoric forest clearance’s denudative effects on limestone
surfaces by palynologists such as that by Oldfield (1960) for the
karst areas around Morecambe Bay in northwest England
carries complete conviction. There can be no doubt that very
many Karren in central and western Europe must be regarded as
stripped. This was realised earlier in the Mediterranean lands.
Indeed the certainty of this kind of history leads conversely to
inferences about length and intensity of human occupation from
42 Karst

the distribution of sculptural forms. Within the Central Highlands

of Australian New Guinea, it has been postulated on other grounds
that the eastern parts were settled earlier than the western. That
at very least high density of population and intensity of land use
were achieved much sooner in the east than the west is suggested
by the presence of bare karst sculptural forms in the east and
their absence from the west except on sheer precipices. There
have been at least 2000 years of agriculture during which rock out­
crops artificially exposed to the elements in the east could be so
modified.
Three groups of factors are therefore responsible for a very
complex overall Karren pattern: the passive factors of rock
petrology, porosity, bedding and jointing, and presence or absence
of covers; the active factors of quantities, temperature and acidity
of rain and soil water, plant growth, glacial preparation; and the
historical factors of changes over time of the preceding controls.

TYPES OF MINOR SOLUTION FORMS ON LIMESTONE

Some of the more common forms of small-scale solutional sculp­

ture on limestone will now be described. The classification is
based on that of Bögli (1960, 1961a) with some omissions and
additions. Littoral forms will not be discussed and related cave
features will be mentioned in Chapter VII.

I. Forms developed on bare karst

These forms develop with free movement of water uninterrupted

by mosses, liverworts, soil or sediment. Additionally it seems
necessary that the surface should not be so overhung by taller
vegetation that the fall of rain is significantly affected.
A. With areal wetting. The simplest effect of rain falling on
bare rocks is to produce small pitting, each rainpit being usually
less than 3 cm across and 2 cm deep. Rainpits form on gentle
surfaces rather than steep ones and may be separated by original
surface or become so close-set as to have only sharp rims between
them. Then the surface has an irregular, carious appearance. In
the tropics, solution flutes (see below) are sometimes interrupted
at intervals by rainpits along their length.
 Minor Solution Sculpture 43

4 Horizontal solution ripples in Silurian limestone, Cooleman Plain,

N.S. W. Scale 6 inches long.

Another very simple form is solution rippling (PI. 4). This is

found on steep (20-30°) to vertical surfaces and consists of
practically horizontal, very shallow solution ripples, extending
laterally over tens of cm but with each ripple 2-3 cm high. The
edges between them are sub-rounded and the surfaces are much
smoother than with pitting. These are well developed on Cooleman
Plain, N.S.W. and the ‘crinkled' weathering of Craven, England
(Sweeting 1966, p. 197 and PI. 22) seems to be related. The
solution ripples of Wall and Wilford (1966) found in the tropical
humid karst of Sarawak differ in that they occur on underhangs
and are narrower and deeper with sharp ribs between. Moreover
the water trickling over them is not fresh rainwater but has
already had much contact with rock, humus, and soil. However,
both are wavelike forms transverse to downward water movement
under gravity, implying a definite rhythm in flow or periodicity of
chemical reaction about which nothing is known at present.
More widespread and more striking are solution flutes (Ger.
Rillenkarren) , again found on fairly steep to vertical surfaces
 5 Solution flutes and bevel in Devonian limestone, Wee Jasper, N.S.W.

(PI. 5). These are longitudinal hollows, running in sets straight

down the steepest inclination with sharp ribs between. There is
a strong modality in cross-section, with a width of 2-4 cm and
depth of 1-2 cm maintained uniformly. Length is much more
variable, measured in tens of cm in temperate latitudes but often
in metres in hotter climates. Their length and frequency are also
related directly to rainfall. Where flutes develop on opposed sides
of a block, a serrated crest results with a herring-bone pattern
seen from above.
Rippling may occur in combination with fluting to give a netted
appearance to the rock.
More frequently associated with solution flutes on moderately
sloping surfaces are solution bevels (Ger. Ausgleichsflächen;
PI. 5). These are very smooth, nearly flat elements, forming
micro-treads backed by steeper, fluted risers, which may curve in
arcs round the upper margin of the bevel. In snow climates, the
risers at the back of bevels are steeper and smoother; when these
are arcuate in plan, the result is the solution funnel step (Ger.
Trittkarren, Trichterkarren-, PI. 6). On vertical surfaces, beyond
a certain length the ribs between two or three neighbouring
solution flutes die out, leaving rain solution runnels (Ger.
Regenrinnenkarren) of the second type of Bögli (1960) (PI. 33).
These forms are about two or three times as big as the solution
flutes themselves, but like them retain the same cross-section
throughout their length.
B. With concentrations of runoff. Solution runnels proper (Ger.
Rinnenkarren) differ from the forms just described especially in
 Minor Solution Sculpture 45

that they increase in width and depth downstream through gather­

ing volumes of runoff water. The ribs between neighbouring ones
though substantial may be rather sharp. Bögli states that runnels
are the product of slower solution than produced the features
already described. This solution is dependent on diffusion of CO-

6 Shallow solution funnel steps in Palaeozoic marble, Mt Arthur, New

Zealand. In rear, grikes, rain solution runnels, and solution ripples.

from the air into the water, and is the stage at which large
amounts of limestone are dissolved.
Not only does water collected from areas of bare rock act in
this way (it is common for sets of solution flutes to feed into
solution runnels), but so also does water running from moss
polsters, from snow patches and from soil and humus covers;
furthermore seepage water re-emerging from bedding planes and
joints may participate. So it operates in many different circum­
stances. We can recognise therefore related forms such as
meandering runnels (Ger. Mäanderkarren) which wind over
flattish surfaces (PI. 7), and wall solution runnels (Ger. Wand-
karren) which are conversely very straight runnels due to water
pouring down vertical faces. The latter are deeper and less
regular than rain solution runnels, nor are they separated by
narrow, sharp ribs.
46 Karst

Most important and widespread are grikes (Amer. solution

slots; Ger. Kluftkarren) which are solution widened joints or
cleavage planes (PI. 9). These inherent planes of weakness
canalise flow and so promote their widening. On flat surfaces
especially, they may form by the merging of a series of vertical
holes, lens-like in horizontal section, arranged along the joints.
Grikes can be extremely straight, deep and long, and often occur
in networks. Enlargement and rounding may take place at joint
intersections to produce cylindrical pits several metres deep, known
as karst wells. These are related both to solution pipes (p. 50)
and potholes (p. 151).
When rocks are steeply dipping or vertically disposed, bedding
planes are likely to be enlarged in the same way; they may be
termed bedding grikes (Ger. Schichtfugenkarren). The strike ribs
left between them may, however, differ from the residual blocks

7 Meandering solution runnels in calcreted Tertiary limestone, Nullarbor

Plain, Western Australia
 Minor Solution Sculpture 47

left between joint grikes. These ribs often break up into pinnacles
(Ger. Spitzkarren) and beehives decorated by solution flutes.
When grikes develop in horizontal, thin-bedded limestone, the
uppermost bedding planes are also likely to be opened up by
seepage. This results in freeing the intervening tabular blocks
known as clints (PI. 9). These get shifted about as solution
beneath disturbs their equilibrium and eventually they break up
into irregular smaller fragments strewn about. Shillow, a term
from the north of England, deserves employment for this latter
condition (PI. 33). It approximates to the German Trümmer­
karren and Scherbenkarst, though these imply the likelihood of
frost wedging. It would be better to eschew this connotation for
the general term and use a qualifying adjective defining the process
where appropriate.

II. Forms developed on partly covered karst

Some areas are patchily covered by soil, sediment, humus, moss

polsters, or plant litter. It is important to bear in mind that this
patchiness may be a result either of removal of a formerly more
continuous cover or of progress in formation of such cover.
Certain forms are characteristic of these conditions because water
is retained in the patches of cover which usually supply biogenic
C 0 2 and, spongelike, permit sustained and substantial solution
over the whole interface. Thus grikes commonly have soil or
humus along their bottoms, with plants growing in the protected
habitat.
To be classed here although they themselves may provide the
only cover, are solution pans (Amer. tinajitas, etched potholes;
Slav. Kamenitsa). These are basin or dish shaped depressions,
usually covered by a thin layer of algal remains, silt, clay, or
gravel. The organic matter supplies additional CO> and the fine
elastics may seal off the flattest, lowest parts of the bottom from
corrosion. In this way corrosion is concentrated on the steep to
overhanging sides of the pan and extends the flat bottom. Indeed
the most favourable locus of solution is the rock-air-water line;
here attack will be most continuous through the renewal of C 0 2
precisely where sealing is least likely. However, it must be noted
that Williams (1968) found higher acidity in the organic and
inorganic material at the bottom of a pan than in the pool of
48 Karst

water itself. Centripetal patterns of solution flutes may surround

the pan above the highest water levels. Under forest, solution pans
can be preferentially enlarged through dripping from tree branches;
here they are not accompanied by flutes. Related but deeper forms
have been called solution cups by Zotov (1941) who infers they
have formed with mosses in them since deforestation last century.
However, such forms are found beneath forest and the problematic
cases are those without any trace of organic or inorganic matter
on their bottoms.
Closely related forms are undercut solution runnels (Ger.
Hohlkarren) with bag-like cross-sections. These are attributed by
Bögli to the filling of solution runnels on gently sloping surfaces
with humus, soil or litter, whereby the sides and bottom are
persistently wetted by water rich in biogenic C 0 2 and thereby
recessed with respect to their uncovered top.
Solution notches (Ger. Korrosionskehlen) are larger but
usually still minor landforms due to particularly active solution in
the top few decimetres of the interface between soil and projecting
rock. This results in curved incuts at soil level up to a metre high
which may run for many metres laterally. They are best developed
in tropical climates (PI. 8). Lowering of the soil surface can
expose the whole of such features as in PI. 8 though this event is
not always a result of anthropogenic soil erosion.
Swamp slots (Wilford and Wall 1965) may be restricted to
humid tropical conditions. They are horizontal smooth grooves
with very flat roofs, a few centimetres to a few decimetres in
height and up to about a metre or two deep. They can be found
with swamp water and organic muck in them at the foot of cliffs
or may occur in series high and dry above this level. In the latter
case they are inferred to have formed when the swamp surface
lay higher.

III. Forms developed on covered karst

Soil or sediment bears like an acidulated sponge on the under­
lying surface and produces its own array of forms. These will
only be seen in excavations unless drastic denudation has removed
unconsolidated inorganic and organic materials which have acted
in this way (PI. 1). However, in special circumstances where
calcareous moraine covers the limestone, subsoil sculpture may
be prevented and glacial striae may even survive, as Williams
8 Solution notch in Eocene limestone, Muriraga, Central Highlands, New
Guinea. Soil removal in abandoned gardens, exposing smooth incurve
formed beneath soil cover.

(1966a) found in Ireland. In such cases the solvent capacity of

percolating water has been entirely neutralised within the drift.
Figure 8 shows how solution forms gradually disappear when
traced from a bare limestone pavement to a glacial moraine cover
gradually exposing rock. Ice erosion has removed former solution
sculpture and subsequently moraine has protected it. Aubert
(1969) describes a similar case in the Jura.
Movement of soil water is not very free and so the predominant
effect is for movement to be strongly guided by the rock structure.
Therefore grikes are produced very effectively in these circum­
stances. Very many of the grikes of Craven, England, where the
term originates, evolved with such fill but are now devoid of soil
or humus, the change being due to interference with the geomor-
phic system by farmers from Neolithic times onwards (PI. 9).
In the United States, the term ‘cutters’ is commonly applied to
soil-filled grikes exposed in quarries (Howard 1963). Below a
certain depth, water movement is very slow indeed and the micro­
structures of the rock are etched out on the walls of open joints
and bedding planes.
Within a metre or so of the surface, however, soil water move­
ment is more strongly influenced by gravity and probably the rate
of solution is faster. In these circumstances the minutiae of rock
structure are overborne by corrosion and particularly smooth
surfaces are developed. If any hollow begins to develop, it gathers
50 Karst

Glacial moraine

Tr^vpnipnn']f r-1-20
Freshly Shallow pans 3m
exposed up to 15cm Pans up to 30cm
bedrock deep and deep Runnels drain Uniform development
nearly runnels up to to grikes up to deep grikes up to
smooth 30cm deep. 1m deep 2m deep
Few grikes
30 Metres
100 Feet

8 Relationship of solution sculpture to moraine cover in Clare, Ireland.

After Williams 1966a.

more water to itself and promotes its own growth (Tricart and
Silva 1960). Thus gravity controlled longitudinal hollows develop
beneath the covers; they resemble solution runnels except that the
ribs between them get rounded by the omnipresent moist blanket.
These are the Rundkarren of the German literature and may be
called rounded solution runnels. Again they are more familiar to
us when stripped by subsequent erosion. The rounded runnels
which form dendritic patterns on the nearly horizontal surface of
dints in Craven (PI. 9) are of this nature since they can be traced
beneath nearby glacial drift (Sweeting 1966). However, the
possibility of forest extending in recent times over bare rock
surfaces on which ordinary solution runnels have developed
previously must be borne in mind; these may have had their
keels rounded and have been converted to the form characteristic
beneath complete covers under forest litter. This sequence could
clearly occur during Postglacial climatic amelioration following
glacial erosion, and Bögli maintains this has happened around the
Bödmeren area, Muototal, Switzerland.
In this zone close to the soil surface much less systematised
smoothed surfaces also develop rounded dimpling and deeper
hollowing which may penetrate as tubes through projections. This
is the cavernous subsoil weathering of the German literature,
kavornosen Karren. It is well exhibited in Highlands New Guinea
gardens on limestone (PI. 10).
Subsoil features which may become very large and pass over
into major forms are vertical solution pipes (Ger. geologischen
Orgeln). However, many of these are quite small cylindrical or
9 Limestone pavement, Craven, England. Former soil cover eroded to
expose grikes and rounded solution runnels. First row of dints about
1 metre wide.

conical holes, a few decimetres in diameter and less than a metre

in length, particularly when they develop in series along enlarged
joint planes (Fig. 9). In the Chalk of northwest Europe they can
develop in more isolated fashion unrelated to joints (Kirkaldy
1950). In aeolian calcarenites in southwestern Australia sections
through earth filled pipes show that solution is not the only
process involved in this special context of induration of dune
sands into limestone (Jennings 1968). The pipes have a shell of
calcrete around them through reprecipitation of calcite; the
developing pipes are centres of induration. Roots are common in
earth fills and pass down beyond into the calcarenite. Again there
is an autocatalytic or positive feedback relationship. Root exudates
and root respiration help the solutional deepening, and enlarge­
ment of the pipe promotes plant growth. Taproots of trees may
completely fill the lower parts of pipes. On emerged coral reefs
pipe development may be promoted by guano which corrodes the
limestone.
Root activity is not restricted to pipes. Wall and Wilford
(1966) have described root grooves of varying size and pattern
in joint planes in Sarawak karst. In the Highlands of New Guinea
roots can penetrate massive limestone without the help of planes
of weakness and even vertical faces carry vegetation which riddles
the surface with grooves and holes. Tricart and Silva (1960) and
52 Karst

Jennings and Sweeting (1963) draw attention to the importance

of tree roots in seasonally wet tropical climate in Bahia, Brazil
and northwest Australia respectively. Even in the dry climate of
the Nullarbor Plain, Australia, intense perforation of the lime­
stone close to the surface is, according to Lowry (1969), also due
to tree roots.

MINOR SOLUTION FEATURES ON OTHER KARST ROCKS

The processes producing small solution features in dolomite are

fundamentally similar to those in limestone (Chapter III). A
detailed study of the features characteristic of dolomite outcrops
of the Niagara escarpment has been made by Pluhar and Ford
(1970). The dominant forms are controlled by rock structure,
chiefly grikes along joints, often rectangular in pattern. Gravity
controlled forms are few though there are some solution runnels
down the vertical sides of grikes. There are also horizontal grooves
along these sides but these are controlled by stylolite seams. The

10 Dimpled and pocketed, yet smoothed solution forms in process of

exposure by removal of soil in gardening at Mainomo, Central High­
lands, New Guinea
 Minor Solution Sculpture 53

9 Solution pipes in gypsum at Walkenried, Harz, Germany. After Penck

1924.

grooves are developed in zones of coarser crystal size and higher

porosity due to weathering between the stylolite seams, which
project as blunt round ribs. Variations in porosity also seem to
localise fairly numerous solution pits (deep solution pans).
In the seasonally dry tropics of northern Australia, dolomites
as at Camooweal, Queensland, also show the dominance of struc­
tural control, grikes and dints being characteristic. Their bare
surfaces show no flutes or runnels, having an overall carious
aspect through tiny pitting.
Sculpture on non-carbonate karst rocks such as gypsum,
anhydrite, and rocksalt is governed by physical solution without
the interaction of solvents such as carbonic acid (Priesnitz 1969).
In humid climate, the sheer rate of solution and mechanical weak­
ness of gypsum produce unstable slopes in which Karren have
little chance to develop, as below Pointe de Jacquette, Bas
Queyras, in the French Alps (PI. 2). In drier climate, this is not
so and, even on the more soluble rocksalt, surfaces become intri­
cately cut up with flutes and runnels as at Slanic in Romania
(Krejci-Graf 1935) and at Djelfa in Algeria (Würm 1953).
54 Karst

ASSEMBLAGES OF MINOR SOLUTION FEATURES

Where extensive areas of bare or partly covered karst occur,

various types of sculptural form may be found in association,
sometimes systematically. Their relationships may permit their
order of development to be determined (Williams 1966a): Fig.
10 shows how the relative ages of runnels and grikes which inter­
sect one another can be inferred.
In Craven, England, the patterns of associated grikes and dints
vary with the spacing of joints, the thickness of beds and their dip
(Sweeting 1966). With moderately thick beds and widely spaced
joints, dints are better developed and are more likely to carry
solution pans and rounded solution runnels than with very thin
beds and very close jointing. With increasing dip, pans disappear
and runnels adopt parallel courses down the dints, rather than
dendritic ones.

RATES OF SOLUTION IN SUPERFICIAL ZONES

Study of the rates at which solutional sculpturing goes on has been

a major activity in quantitative karst geomorphology during the
last two decades. There are many practical problems such as that
of measuring what goes on at the soil/rock interface. Moreover the
circumstances are so full of variety that results are often not
properly comparable with one another.
In 1947 Sweeting (1966) stripped moraine from a glaciated
pavement in Craven, England and by 1960 the glacial striae
exposed had been removed by corrosion, with a surface lowering
as much as 3-5 cm in places. From another area, peat was removed
and peaty water directed to flow on to this stripped rock surface;
this cut runnels 7-15 cm deep in the same period. Experiments of
this nature, though valuable, upset natural conditions and results
obtained may not have a great deal of application elsewhere.
Installation of stainless steel pegs as reference points causes much
less disturbance (Hodgkin 1964) and future measurements with
their help should have more significance.
Measurements of solutes in waters forming these solutional
features today are not readily interpreted because of varying length
and intensity of precipitation, varying length, rate and duration of
surface flow, thickness of film, and presence of ions complicating
the main solution process. Sampling water flowing over limestone
 Minor Solution Sculpture 55

Runnels which
Captured upper post-date
runnel grike formation

Runnel which
pre-dates grike
formation

10 Time relations of runnels and grikes. After Williams 1966a.

on the Dachstein plateau, Austria, Bauer (1964) found that the

concentration of CaCOs varied from a maximum of 40 mg/1 at
the start of a rainstorm to a minimum of 13 mg/1 shortly after
the greatest rate of rainfall when rate of flow outbalanced the
solution rate, rising again to 25 mg/1 during the dying phase. He
also compared the concentrations achieved in flows down a smooth
slope (14-63 mg/1), a shallow runnel (15-76 mg/1), and a deep
one (30-96 mg/1). Converting the mean values to long-term rates
of surface removal by applying them to the mean annual rainfall,
he arrived at rates of horizontal surface lowering of 0 9 c m /1000
years, 10 c m /1000 y and 13 c m /1000 y respectively. This shows
how runnels promote their own development (autocatalysis).
At the same time Bauer’s figures indicate that small numbers of
somewhat haphazard samples from different kinds of karst are not
56 Karst

likely to provide very meaningful comparisons. Table 3 sets out

some of the available data. Williams’s large number of samples
from solution pans from Clare, Ireland, provides a reliable mean
of 66 mg/1 which for the prevailing temperature and normal
atmospheric Pco2 corresponds with the theoretical saturation
equilibrium. Because of very free interchange with the atmos­
phere, the algae in the pools can only increase the speed with
which saturation is reached and do not enable a greater equilibrium
concentration to be achieved here. The moss polster and the soil/
rock interface samples have generally higher concentrations than
the bare rock samples, though there is no significant difference in
the Yugoslavian data and there is a reverse relationship in the
Lapland case.
Because the soil provides the bulk of the biogenic COo, a great
deal of solution takes place inside the soil as long as carbonate
fragments remain and at the surface of contact between the soil
and the bedrock in place. Measurements of CaCO:{ concentrations
in the Jura have shown that this superficial solution is equivalent
to the removal of a layer of limestone 0 05 mm thick each year
and it is thought that this is about 60 per cent of the total amount
going into solution in the area (Aubert 1969).
Investigators have generally found their highest concentrations
of C aC 02 in waters passing through unconsolidated deposits
containing many limestone clasts. Thus on the Dachstein plateau
Bauer (1964) determined the rate of calcium carbonate loss from
a glacial moraine through the amount removed to create small
depressions in its surface overlying tongues of decalcification (Fig.
11). The rate of loss of rock equivalent was 3 6 c m /1000 y com­
pared with 2-8 cm at the forest humus-rock interface and the much
lower rates from bare surfaces mentioned earlier.
This last determination avoids the danger in extrapolating back­
wards in time from observations of the conditions at the present
moment. Thus Sweeting (1966) points to the great reactiveness
of smoke from industrial areas of northeast England which reaches
limestone areas of Craven, for example, and it is well established
that climate changes significantly over quite short periods of time.
Therefore a great deal of interest attaches to surface losses
attested by the pedestals (Fig. 12) which have formed beneath
dints and glacial erratics on limestone pavements eroded by ice
of the last Pleistocene glaciation (Table 4).
 M inor Solution Sculpture 57

NO'S
a3 £ J
8 CQ
00 GO 3
£ - cn io m is - m .S
•*-» ac
o c © <N vo O in oo vn <D n
03 — o (N O vo >n vo rn w
C/3
bO
_C
•an CaCOi concentration in m g/1 froi

c
• r\ 1

SCQ VO
On

OV ON © S /—v
rf f*n *T)
3 -tö
C'S
Q -o
II ’S
J&S
J/S

S
IW

CQ 3
S

o X)
cv 00 CQ
C
a m m m r*i 8»
E vo
a. —3 m
1vo • V3 ON 150 ON
W VO 00 r- o
Sb
:0 .5 «
ffl u U
<u -c
II * o
C/3
EH
on S QQ CQ < s J « T3 ’S
ON ^ ^ VO /—N do rt
(n| VO VO <N Tf
VO « « Tt o | l
(N m r~ Pi ON
or .5 s
jy
§H
° II
1 -s „ i ^
Bshw
Amw

Amw

& l I H £ £ s£ X ) «4-»
X3 a*—,
X3 <+_<
X> aX* )_
Bsh

«4H
Aw

Afi

:o C WQQQ U U U Ü U
* 'S 1$ be
C
II ~ 'S
11 <U <u
X3
£
a 5 ÖC
Sl on

Ct T3 VO
TD 1 = </> av
0) i—I w)
S om erset I., N . Cane

(w et season)
W . K im berley, A ustr;

Ja m aica (d ry season)
N u lla rb o r P lain , A us

bo
a - c
B isistal, Sw itzerland
D ach stein , A u stria

E .to ‘c
L ap lan d , N o rw a y

6C/3 cC
C raven, E n gland

s
S arthe, F ra n c e

£3
S.E. A u stra lia
C lare, Irelan d

a
Ju ra, F ran ce

P u e rto R ico

*
Indo nesia

E .2 00
on = o^j
is ^ ”
C u ba

I "I
O > ccj
03 !> •a
58 Karst

S u r f a c e depression due to solution

Area of depression
Weight of clay soil tongue
Moraine to yield this (93 4 %
carbonate, 6 6 % insolubles)— 7485 kg
Volume for this weight
at S.G. 1 5 6 --------------------------4 785m3
Expressed as depth of moraine lost
over area of depression-------- 65-3cm
Expressed as depth of limestone rock
lost over a r e a of depression—35cm Clay soil plug
Average loss of Limestone rock over Rendzina on
10,000 years since g laciation—3-6cm/1000y calcareous moraine

11 Determination of limestone solution from pit in rendzina on calcareous

moraine in the Austrian Alps. After Bauer 1964.

The good agreement between the results achieved by the two

methods is therefore surprising but very useful. The conclusion
follows that very many surface solution features have formed in
the Holocene. This agrees with the observations that they are
commonly found in areas which were strongly abraded and plucked

12 Glacial erratic of Silurian sandstone on a Carboniferous limestone

pedestal due to Postglacial solution, Norber Crags, Craven, England.
Sketch by C. D. Ollier.
 T able 4 Surface lowering rates from limestone pedestals

£
£
•5

a
"§■

§
Average ht of Time since from runoff & Proportion of

£
pedestals ice disappearat precipitation total limestone

o
cm years B.P. c m /1000 y removal from
area

o
o

s
Ö
O

S
in

'1
in

-a

N
C^*

on

:S5
<D

c<U
1-51

1
^

. -

S
&
-H

CO ON
VO
$
O

u
m
N
in

S
T+
<N
1 2 , 0 0 0 40
M inor Solution Sculpture

|
^ C o

r. ' S
w .S

CÜ 'w * '
(N
&

T3
15,000 10 0-5-1-5
60 Karst

by glacier ice in the final phase of the last glaciation (Williams

1966a; Haserodt 1969). Nevertheless it must not be assumed that
all minor solution features, especially deep grikes, formed in such
a short period.
 V

DRAINAGE

Surface drainage in karst is liable to be intermittent, disrupted,

widely spaced, or absent. Marked permeability has enabled under­
ground drainage to take over the task of moving water to the sea
or to surrounding country on other rocks to varying degrees in
space and time (PI. 11). In this chapter the passage of water
through karst will be discussed separately from the resulting land-
forms as far as possible so that the functioning of karst hydro-
logic systems will stand out more clearly.

INFILTRATION, OVERLAND FLOW, AND THROUGHFLOW

Karst is marked by rapid and substantial infiltration into rock out­

crops and soil, restricted overland flow which rarely reaches stream
channels, and modest throughflow in the sense of lateral movement
through soil pore space, though these generalisations need some
elaboration and qualification.
Falling on karst rock outcrops, rainwater usually flows over the
surface very short distances before infiltrating. The more intense
the rainfall and the steeper the surface the longer will overland
flow be, as witness the long solution flutes and runnels on vertical
faces in the perennially or seasonally humid tropics. The coarse
porosity of emerged coral reefs usually causes immediate absorp­
tion so that lack of surface water can be extreme even with mean
annual precipitations of 2000 mm or more. Rapid infiltration is
also characteristic of chalk despite its micritic texture and very
fine intergranular porosity. With compact rocks lacking inter­
granular porosity, water flows over the surface till it encounters
planes of weakness in them. The frequency, openness, and con­
tinuity of these planes controls infiltration. Aubert (1969) contrasts
61
62 Karst

11 River Angabara, Central Highlands, New Guinea, penetrating through

Miocene limestone strike ridge. Flat-floored closed depression beyond
the ridge is partly developed on impure limestone.

the numbers and openness of joints, particularly strike joints,

along anticlines in the Jura Mountains with their paucity and tight­
ness in the synclines; very fine joints, especially characteristic of
argillaceous limestones, hinder infiltration as a result of surface
tension and blockage by weathering products and precipitates.
Joints wide enough to permit infiltration are liable to be enlarged
by solution so that any initial permeability accentuates itself. Close
fields of grikes bring about virtually immediate loss of rain under­
ground. In thick beds lacking porosity, solution pans and larger
solution hollows, such as the rockholes of the Nullarbor Plain
(Jennings 1967c), can hold water until it is entirely lost by
evaporation.
With covered karst, the nature of the soil or superficial deposits
is critical. Many residual soils on karst rocks allow high infiltra­
tion rates. This is true of rendzinas, common in mid-latitude
karsts; these soils are dark, alkaline, shallow loamy soils, with
 Drainage 63

crumb structure and including residual rock fragments. Such soils

also permit throughflow but this will only be significant on steep
slopes since infiltration water tends to pass rapidly into the under­
lying bedrock with but modest lateral movement through the soil
pores or at the base of the soil.
However, other residual and transported soils in karst behave
very differently. Prolonged leaching of the weathering mantle can
result in acid, dense clay soils which consist of clay mineral and
iron sesquioxide residues from the bedrock. Some at least of the
‘terra rossa’ commonly found in Mediterranean karstlands is of
this nature. Soils like these restrain infiltration, especially when
they are cleared of forest, grazed or cultivated; moreover they
tend to block joints in the underlying rock. Longer overland flows
prior to infiltration or to feeding surface streams result. Where the
karst is only partly covered, less permeable soils of this kind
generally occupy hollows and deflect water laterally towards out­
crops. Heavy clay soils act as seals in the bottoms of depressions;
hydromorphic soils, swamps, and even standing water may
develop.
Superficial deposits are as various as the soils. Colluvia derived
from limestone or dolomite are likely to include much coarse
material and be very permeable, promoting both infiltration and
throughflow on the appreciable slopes associated with them.
Whether originating within or without the karst, alluvia will
chiefly vary according to texture, the finer the texture the less the
infiltration and throughflow and the greater the overland flow.
Glacial moraine derived from the karst itself is commonly per­
meable and overland flow modest over it. Loess covers behave
similarly and so also those of volcanic ash until pedogenesis in
humid climates has made substantial progress, by which time they
are also likely to have been stripped from higher relief and concen­
trated in the lower parts of the landscape. The bauxitic clays in
the Jamaican karst cockpits, of low infiltration capacity, may be of
this latter origin. This list could be made very long and it is
important to recognise that there can be sharp variations in the
relative importance of infiltration and runoff both within and
between karst regions.
Vegetation also has a great influence on infiltration, largely
through varying loss by transpiration. Holmes and Colville (1970
a, b) have shown that on the same Gambier Limestone and the
same soil in southeastern South Australia, pine forest causes twice
64 Karst

as much loss to the atmosphere as grassland and this reduces

infiltration into the limestone to nil.

SURFACE RIVERS

Surface streams are usually few in karst. In the whole Dinaric

karst of Yugoslavia, only four rivers cross it to reach the sea— the
Krka, Cetina, Moraca, and Neretva— and of these only the last
crosses the whole width of the main limestone belt. One-quarter of
the whole area drains directly into the sea without surface stream
flow at all (Gams 1969). In semiarid and arid karst, rivers are
completely absent and this is rarely true of desert country on imper­
meable rocks. The Nullarbor Plain, Australia, a free karst like the
Dinaric, feeds all its meagre water underground to the ocean.
In very humid climates or on less pure karst rocks, more
elaborate river patterns are found, but generally the drainage
density is less than on other rocks in the same conditions. With
well developed karst this attribute is self-evident but in other
circumstances it may need careful demonstration. Such an instance
is provided by Miller’s analysis of drainage basin characteristics
in three lithologies in the Clinch Mountains area of Virginia and
Tennessee, employing Horton’s measure of drainage density of
length of channel per unit area of catchment (Miller 1953).
Dolomite is the bedrock in the Copper Ridge area whereas thick,
coarse sandstones alternate with shales in the neighbourhood of
Blountville, with the Pennington district providing the greatest
contrast with thin, fine-grained sandstones and siltstones inter-
bedded with a dominant body of shale. All have similar local
relief and dendritic drainage but streams are longer and basins
larger in accordance with degree of perviousness (Table 5). As a
result drainage density is least in Copper Ridge and greatest in
Pennington in both first and second order streams. The differences
between the dolomite area and the dominantly shale area are
significant at the 95 per cent confidence level. The Blountville
area with its thick pervious sandstones as well as thick shales is
only significantly different from the Pennington shales in drainage
density of first order streams.
The four rivers which cross the Dinaric karst are allogenic
(allochthonous) in that their headwaters are on impervious rocks.
They have practically no surface tributaries along their entire
 Drainage

c
o
VJO
+
1
I
a.

1
«
'S
o l 'S «Ö
CQ c 'S
I ^3
C o «50

1 jo
+
>
•Sg
>>

(3
9
<3* ll
s
O

JS
*Ö cd
9 “
u ^ vo <N Wo O VO
.s s (N r~ O <N t " vo - o
\ WO rb VÖ sb b - to C

5 |X to [X to
.s & X
V)
Ov
J2
£ S c
'S s
c
uc
2 I | + +
£I
■e £ ft.
1
BS «c
® 9
■J. 'S
«50
I # 1 + "55
1 CI§ 0
• S 'S
CQ JOo
•e

f t
«
1
I*,
’S 3
WJ.S2
S .2
+ ” 'S
•i £ § «
<U ,,
S* E 11

.a
|X +
00 00 0\ o <n r-~
►> S WO Tf 0 \ <N ■'t '3 -
1
•J>£ WO (N Wo (N t ' ' <N
CO
<u
9 O
◄
s IX to IX to o
X <u
W
to
>-1 Q
<
H

<D
00
"O ^
!§S
J=, H
C ~
■Ms? T3 9
5 £
«
93
T3
C S ’g-S g -
Ö s
D. O
>
C
3
ü
«

rt
fills
& o
ü 3
O
TT
CQ
J3
w
CO Mill
66 Karst

lengths through the karst, though springs do supply them with

water. Allogenic rivers are very common in karst regions. How­
ever, there are rivers which begin their courses on karst rocks,
i.e. they are autogenic (autochthonous). They do so in springs,
often large ones, so that autogenic karst rivers are liable to be born
adult, as it were. The circumstances which give rise to springs are
discussed below.
To cross karst on the surface, both kinds of river are frequently
dependent on alluvium which seals off the permeable country rock.
Some tropical karsts consist largely of broad, alluvial plains on
limestone, with perennial rivers meandering over them as in the
Kinta valley in Malaya. In Guadeloupe dendritic valley systems
occupy much of the karst with intermittent surface streams along
their bottoms where terra rossa has accumulated thickly (Lasserre
1954). Although the Central Lowland of Ireland is karstic, it has
many surface streams and lakes; this is undoubtedly partly due to a
mantle of Pleistocene glacial deposits. But it is probably also a
consequence of preglacial reduction by solution of the limestone
to a plain close to sea level on which residual red clay cover had
developed (Williams 1970).

RIVER REGIMES

The more underground drainage participates in a hydrological

system the more efficient it is and the more of the precipitation
can act geomorphically. Rapid infiltration means that water
escapes the heat, wind, and low relative humidities of the surface
sooner and so evaporation is reduced. It is true that plant roots
probably reach more deeply as another consequence. Tree roots
20-30 m underground are common; in Lake Cave, southwestern
Australia, living roots occur about 60 m down. Nevertheless
biological productivity is not generally higher in karst than else­
where, and is often less, so that increased transpiration does not
balance reduced evaporation. Because of this, limestone areas dis­
charge a higher proportion of the water they receive as precipita­
tion than does impervious terrain. Parde (1965) estimates that
evapotranspirational loss, determined by the precipitation-dis­
charge deficit, in the Nera, Aniene, and Pescara headwaters in the
calcareous Apennines is about 500 mm whereas impervious catch­
ments in these relief and climatic conditions lose 600 mm. This
 Drainage 67

represents a 15 per cent saving. Equivalent figures for the Tarn

in the Grandes Causses, France, are about 300-310 and 425-450
mm respectively, making it even more efficient with a saving of
30 per cent.
Nera River at Torre Orsina Tiber River at Ponte Nuovo

13 Graphs of mean monthly discharges as fractions of mean annual dis­

charge of a karst river (River Nera) and a normal river (River Tiber).
After Parde 1965.

The water storage capacity of karst also affects river regimes

and so modifies their geomorphic behaviour. Parde (1965)
estimates the percentage of voids at 10-20 per cent for small karst
masses but at only 3-5 per cent for whole plateaux. However, the
voids are virtually all effective in terms of permeability and
storage. The result is that karst rivers sustain higher base flows
longer and have only moderate flood peaks compared with rivers
on impervious country. In Italy, for example, the mean monthly
discharges of the River Nera at Torre Orsina only vary between
0 83 and T06 of the mean annual discharge, whereas the Tiber
at Ponte Nuovo has corresponding figures of 018 and T95 (Fig.
13). The extreme minimum discharge of the Nera of 035 com­
pares very favourably with the 0 07 of the Tiber. The maximum
flood recorded on the Nera is only about 500 m3/sec but 3000-
4000 m 3/ s e c could be expected if it were on impervious terrain
like the Tiber. White and Reich (1970) have similarly shown
that mean annual flood is low in carbonate basins compared with
that in basins without karst rocks with data from Pennsylvania.
This kind of flow regime might be expected to minimise river
erosion in karst since it is accepted that rivers perform most of
their work during floods, whether these are exceptional in magni­
tude or more numerous ones of lesser magnitude. Diminishment
of power must apply to mechanical action by karst rivers. How­
ever, corrosion is much more important than corrasion in this
context. Correlation of discharge and carbonate solute concentra­
tion in karst rivers generally shows an inverse relationship (Fig.
68 K a rs t
a b
Green River at MunfordviLle 1956 - 1957 C liff Foot Rising
y = 183 9 1 7 - 4 0 176 L o g 10,000 X <200 Blue Waterholes
r = -0 -9 1 8 Cooleman Plain
i/i
a>
£
° 100 ° 100
o < ... o

0-001 0-01 0-1 10 0-01 0-1 1-0

Discharge (m 3/sec) Discharge (m 3/s e c )

Fergus River at Ennis Fergus River Basin

in 10, 000-1
OJ
% at
O a;
Log Y = 0 -6 7 3
+ 0 9 8 4 Log X
o 100
o
It' . <:

1
o-J-------------------- ,---------------------,
0-001 0-01 0-1
1
Discharge (m 3/s e c ) Discharge at Ennis (m3/s e c )

Shannon River at Killaloe Shannon River Basin

Log Y = 2 4 8 6 4 + 0 3 1 7 5 Log^X

100,000 -, Shannon Basin

10,000
N = 22
r = 0 941

100 1000 100 200 300 400 5CW

Discharge (m 3/sec) Flows equal to and exceedina
( m3/sec'

14 (a) Inverse relatio n sh ip o f hardness and discharge o f a karst rive r,

G reen R ive r, K e n tu cky. A fte r D o u g la s 1968.
(b) Same relatio n sh ip in a karst spring, Blue W aterholes, N .S.W .
(c) L a ck o f co rre la tio n o f hardness and discharge o f a karst rive r,
Fergus R ive r, Ire la n d . A fte r W illia m s 1968.
(d) D ire c t relatio n sh ip o f lim estone re m o va l and discharge, Fergus
R iver. A fte r W illia m s 1968.
(e) Same relatio n sh ip in Shannon R ive r, Ire la n d . A fte r W illia m s 1970.
(f) R e la tionship o f lim estone rem o va l rates to flo w frequencies, Shannon
R ive r. A fte r W illia m s 1970.
 Drainage 69

14a), though Douglas (1968) and Williams (1968) have shown

that in the case of the Hull River, England, and the Fergus River,
Ireland, there is no significant change of hardness with discharge
(Fig. 14c). The River Thames near Oxford, England, even shows
a direct relationship between carbonate hardness and volume of
flow, but this is exceptional. Thus reductions of high stage in
karst rivers and consequent diminishments of velocity are not such
as to involve loss of solutional capacity as well; turbulence must
generally be sufficient for the moving water to dissolve as much
of the karst rock as other circumstances permit. It must be noted
that the higher stages will still remove more limestone in unit
time and for the Fergus, Williams (1968) found that in the
winter month of December, with its higher discharge, four times
more limestone was transported from the catchment than in either
June or July (Fig. 14c). In the Shannon, also in the Central
Lowland of Ireland, mean flow and lower stages account for about
68 per cent of the annual load of dissolved limestone (Fig. 14e,
f; Williams 1970). Thus karst rivers can remain geomorphically
effective with their moderated regimes because of their dominant
solutional action. This would not be the case if corrasion was the
chief way in which they deepened their valleys.
Nevertheless many limestone rivers may transport rather than
corrode. Through his studies of water chemistry in the Fergus and
Shannon catchments, Williams has shown that most solution takes
place before water reaches the surface streams where indications
of carbonate precipitation are the most prominent evidence. The
constructional action of karst rivers is discussed in Chapter VI.

SINKING OF RIVERS

Rivers entering karst are liable to lose all or part of their drainage
underground. Where the waters of a stream can be seen to enter
a cave entrance, opening laterally (PI. 12), or vertically from the
surface channel, or to go into narrow fissures in the bedrock, the
fact of karst loss is evident whatever the climate. In Slovenia there
is the classic case of the Pivka River, which is barred by Sovic
Hill, under which it passes quietly into the famous Postojna Cave.
In Craven the stream from Malham Tarn sinks into fissures in its
bed. In humid climates the loss is still evident, provided it is sub­
stantial or complete, even when it takes place gradually into
70 Karst

gravels on the bottoms and sides of river channels or in swampy

areas, and thence into the karst rock beneath. The Takaka River
in Nelson, New Zealand, is liable to disappear into its gravels
whereas other rivers in the area in impervious catchments flow
perennially over their gravel beds to the sea.

12 River Rak entering Tkalca Cave, Slovenia. Photo by P. W. Williams.

 Drainage 71

But in subhumid and semiarid climates rivers are intermittent

in flow and may only reach part way along their courses as a
normality through evaporation, percolation into gravels, and
failing supply. Here additional loss into underlying pervious bed­
rock may not be readily discernible though it can be just as
important in the hydrologic regime. Wee Jasper Creek in southern
New South Wales feeds water into Dogleg Cave under Punchbowl
Hill by ways too small to identify, let alone penetrate, where it
flows along its gravel and bedrock bed along the flank of that
limestone hill. However, this channel does not differ in general
appearance from those of other streams in the area on imper­
meable rocks since all are intermittent in flow. Southwards some
55 km in the same drainage basin but at subalpine levels, the
North Branch of Cave Creek is frequently dry over a substantial
portion of its course across the limestone Cooleman Plain; this is
remarkable because it is an area where water balances normally
result in perennial drainage. Where only a modest part of the
discharge of a stream is lost into a karst rock beneath gravels or
swamps, it may be hard to detect without gauging even in a very
humid climate.
Frequently a stream will have a series of sinking points or
swallets along its course into which it loses successive fractions of
its volume. The Manifold River in the English Peak District may
fail to flow over 7 km of its bed between Wetton Mill and Ilam
Hall in dry weather (Warwick 1953). The upper course of the
Manifold runs over Namurian shales, then it flows on to Visean
limestone with interbedded shales and finally encounters reef lime­
stones (Fig. 15). There its discharge decreases through losses
into a series of streamsinks. Some are shallow, bouldery depres­
sions in the bed as at Wetton Mill itself; others are inflow cave
entrances, e.g. Redhurst Swallet. With falling stage, the most
downstream one becomes the limit of surface flow first, then
successively each upstream one in turn until the Manifold gets no
further than Wetton Mill itself. The continuity of surface flow is
usually broken for 1-3 months each year. With rising stage, the
Manifold exceeds the capacity of each successive sinking point in
turn downstream and eventually has continuous flow over the
surface once more.
Streamsinks can be ordered and analysed in much the same way
as ordinary surface streams in the Horton-Strahler manner
72 Karst

Re ef L i me s t o n e
J Be d d e d L i me s t o n e
Permanent st re am
Interm ittent stream
■ Dry val l ey
• 5 Streamsink
t r # I n f l o w cave
O O u t f l o w c ave / S

300 M e t r e s
1000 F e e t

15 Successive streamsinks at Wetton Mill on River Manifold, Peak District,

England. After Warwick 1953.

(Williams 1966b; Fig. 16). A first order stream has no tributaries,

whereas a second order stream has tributaries but these are only
first order streams. A third order stream can have first or second
order streams as tributaries, and so on with higher orders. A
streamsink takes on the order of the stream disappearing in it.
 Drainage 73

O Swallet
---- Watershed
...... Dry valley
_| Limestone
' j Impermeable
- - J rock

o t

o iii 10-
C TO
a l.

2 3 4 2 3 4
Order Order Order

16 Method of ordering streamsinks. Some relationships of streamsinks for

northern lngleborough area, Craven, England. After Williams 1966b.

Williams modifies the ordering system in certain ways. Dry valleys

on impervious rocks forming tributaries of sink systems are
included in the reckoning since they are liable to have surface
flows in wet seasons. On the other hand sinking streams fed from
large springs (which are the resurgences of surface streams) are
excluded from the analysis because they cannot be ordered
properly. Determining the catchment area of a swallet is based on
surface watersheds; this will be satisfactory as long as most of
the catchment is on impervious rock and little on the karst where
surface watersheds often do not correspond with underground ones.
A higher order streamsink catchment may enclose those of lower
order sinks; the extent of the latter must be excluded from the
former’s area.
Williams applied this system to the lngleborough part of Craven
and found that as with normal stream systems, the frequency of
streamsinks varies inversely and geometrically with sink order,
whilst the mean area of streamsink catchment varies directly and
74 Karst

geometrically with sink order. Furthermore the mean distance

apart of swallets of the same order increases geometrically with
sink order. All these relationships are to be expected but they
provide measures of organisation in streamsink systems. An index
of streamsink density can be determined by dividing the total
number of sinks by the area of karst concerned and provides one
of the means of comparing the hydrology of different karsts.
There are some streamsinks which in time of flood may reverse
their function and discharge water, acting like springs. This
happens because the cave which a given stream is feeding also
receives other feeders underground and on these occasions the
combined volume is greater than the absorbing capacity of the cave
farther in. The French name estavelle is commonly used for these
alternating orifices but they are also well known by the Greek
katavothre.

SPRINGS

In most kinds of terrain, throughflow in soil, waste mantles, and

weathered rock gives rise to small and intermittent springs. In
karst, springs are much larger and more frequently permanent
because of greater infiltration of precipitation into karst rocks and
because of the input of streamsinks fed from surrounding
impervious rocks. A useful distinction can be made between
exsurgences fed entirely by seepage waters from the karst and
resurgences supplied by the sinking of surface streams (Trombe
1952). Resurgences are nearly always fed in part by seepage also
so that there is every transition between the two. Moreover there
are usually many springs of unknown provenance. Between Cape
Naturaliste and Cape Leeuwin in southwestern Australia, dune
limestone along the coast forms higher ground than the land just
behind. Some drainage from the east enters the dune belt to feed
resurgences near the coast; this is true of a spring west of
Conolly Cave which is fed from Mammoth and Ruddock Caves,
each receiving surface drainage. Farther south there are springs
such as those at Cape Leeuwin on both sides of the dune belt which
receive no surface inflow from without; these are exsurgences.
Karst springs are very varied. Some take the form of the open
entrances to caves out of which streams flow under gravity (Fig.
17a, PI. 13). Such are common in Craven because the valleys
 D ra in a g e 75

Free surface Flood overflow

gravity spring \
spring /
Normal
flo w v
spring^

The Fontaine de Vaucluse : an inverted siphon spring

/C /y )( Overflow at
22m3/sec Lowstage
N orm al/ 7 ^ / spring
I n ... X / /

Permeable talus
accumulation -V'
/ a /\/ y / / /x
100 Metres
300 Feet

Ebbing and flowing spring Submarine spring (a)

Siphon stops ( a ) and inflow (b)
k /A iR Siphon flows (b)
AIR

17 Types o f springs
(a) G ra v ity -fe d spring issuing fro m cave m outh.
(b) F lo o d surplus spring.
(c) S p rin g risin g under h yd ro sta tic pressure, F o u n ta in o f Vaucluse,
Provence, France. A fte r M a rte l and de Cousteau.
(d) E b b in g and flo w in g spring. A fte r T rom be 1952.
(e) S ubm arine spring and in flo w as in Sea o f A rg o s to li, K e p h a llin ia ,
Greece.

there have cut down deeply into the limestone close to or into an
underlying basement of impervious rocks. Some springs such as
White Scar Cave and Austwick Beck Head are actually developed
in the unconformity between the limestone and the Precambrian
foundation. But others such as Clapham Beck Head, Douk G ill
Cave, and B irkw ith Cave lie within the limestone. Other springs
descend similarly under gravity but emerge through the interstices
in rubble which has fallen and accumulated over a cave entrance;
this is the case with C liff Cave Spring, Cooleman Plain.
76 Karst

13 River Lagaip emerging from a cave near Kepilam, Central Highlands,

New Guinea

An alternative name for a spring is a rising and this name is

literally earned by many karst springs which well upwards, quietly
or with vigour, through open cavities or narrow fissures, gravels
or swamps. There is quiet upwelling through gravels and in wide
swamps at Kirk Göz in the north of the Plain of Pamphylia by
Antalya in southern Turkey. These risings occur on the edge of an
erosion surface against the Taurus Mountains, both in Cretaceous
limestone. These are the waters of the Diidencay, which sinks
again shortly for a further underground course of 12 km beneath
Pleistocene travertine forming part of the same surface. The second
emergence lies at the rear of a lower erosion surface cut into the
travertine (Vaumas 1968). Set in a gravel plain, Piu Piu Spring
near Takaka, New Zealand, is a perpetually ‘boiling spring’, rising
with such force with a mean flow of 10 5 m3/sec as to prevent
divers from reaching the orifices of emergence which appear to be
in white limestone (PI. 14). It is probably a true ‘Vauclusian
spring’ in that the water is rising under pressure up water-filled
passages in bedrock. The Fountain of Vaucluse itself at the head
of the River Sorgue east of Avignon in southeastern France has a
bedrock passage rising at about 45° to a little round cave in
 Drainage 77

14 Pin Piu Spring near Takaka, New Zealand. ‘Boiling spring’ rising under
pressure from limestone beneath alluvial plain.

which it forms a pool, extending out into the open at high levels
when it overtops a shallow talus barrier (Fig. 17c). Normally it
percolates through the talus above bedrock to emerge about 26 m
lower down. Divers have descended over 100 m in the bedrock
passage but have not been able to reach the descending part of
the inverted siphon which provides the pressure head for this
spring.
‘Ebbing and flowing wells’, intermittent springs with a regular
period, such as that on Buckhaw Brow, Craven or the Arize River
spring in the French Pyrenees, require a true siphon for their
action (Fig. 17d). The water level in the system oscillates between
a and b. When it builds up to b, the siphon begins to function
and the level drops rapidly to a. There must be some connection
to the atmosphere behind the siphon or else air pressure effects
would intervene and the capacity of the siphon must be greater
than the inflow of the cave stream, otherwise there would simply
be a persistent small flow.
Frequently a persistent spring has associated with it a higher
spring which only functions after heavy rains. The passage to the
lower spring fills to capacity at this time, water backs up behind
78 Karst

and there is overflow through a branch passage to the higher

outlet (Fig. 17b). This is the relationship of Ingleborough Cave to
Clapham Beck Head at Clapham, Craven, and on a larger scale
of the Hölloch Cave to its perennial spring, Schleichende Brunnen,
100 m lower down in Muototal, Switzerland.
Many factors govern the location of springs. Underlying and
interbedded impervious rocks cause springs to emerge from the
base of the overlying karst rock. Impervious deposits such as
glacial moraine banked against limestone or dolomite result some­
times in springs along their upper margin, and alluvial fills form­
ing river terraces often have the same effect. Without the inter­
vention of impervious materials, springs emerge along valley
bottoms in karst or at sharp breaks of slope as with the case
mentioned earlier. These are places where the saturated zone or
rest levels in underground streams intersect the surface. Springs
also occur high up valleysides within the body of karst rock with­
out obvious topographic or lithologic cause; such is the case with
the Golling Falls 100 m above the Salzach River in Western
Austria. Here valley deepening has proceeded so fast that solu-
tional enhancement of permeability has not been able to keep
pace. The lower part of the limestone is still impervious and
throws out the underground water. This appears to be the case
also with the Efflux in Bungonia Gorge, N.S.W. Sea level is
another important control of spring location, e.g. the large Ombla
spring emerges at the shoreline near Dubrovnik on the Dinaric
coast.
Nevertheless springs may emerge beneath lake or sea surface
level where the karst rocks descend well below it (Fig. 17e). The
Bourbioz spring emerges at —80 m in Lake Annecy in the French
Alps. Vrana Lake in the island of Cres has sublacustrine springs
which appear to be fed beneath the Strait of Kvarner from the
Istrian plateau or from the Croatian plateau to the east. The
writings of classical Greece and Rome have made famous the
submarine springs of the Asia Minor coast, e.g. near Akcay in
the Gulf of Edremit, of the Dinaric coast, e.g. near Rijeka, and of
Greece itself. Submarine springs called ‘posas’ are also common
in the sea around the limestone plateau of Yucatan.
Mistardis (1968) specifies a number of submarine springs
emerging at depths of 30 to 40 m around southern Greece. Four
hundred metres off the Kynourian coast with the high Parnon
Mountain above, Anavalos spring opens on the bottom at —36 m
 Drainage 79

but a diver has descended the spring itself a further 40 m so that

the fresh water emerges at least about 75 m below sea level. Many
writers including Mistardis attribute these deep submarine springs
to cave development beneath the exposed sea floor during Pleisto­
cene glaceoeustatic low sea levels. Theoretically deep phreatic
solution (see p. 96) should also be capable of forming them
beneath the sea.
Genetically related to submarine freshwater springs are sub­
marine sinking points of sea water such as the famous whirlpool
in the Sea of Argostoli in the Ionian island of Kephallinia (Fig.
17e). If a submarine spring has two openings, it is possible that
the impetus of the freshwater current through the more direct
passage may draw sea water down the other to mix with the spring
water and emerge through the first outlet. Trombe (1952) has
shown that the density differences between the columns of sea
water and less saline mixed water can be sufficient to motor such
a system on its own once it has started. Stringfield and Legrand
(1969b) have used the same explanation for reversing flow of salt
and fresh water between Tarpon Springs on the west coast of
Florida and Lake Tarpon nearly 2 km inland.
Williams (1966b) has proposed several measures for the
analysis of the geomorphic role of springs— the karst rising
density, which is the number of springs divided by the area of the
karst, and the rising coefficient, which is the ratio of the standard
deviation of rising heights to the mean rising height expressed as
a percentage. The rising coefficient may be useful in the inter­
pretation of karst hydrology. Other related indices put forward are
the vadose index (the difference between the mean heights of
streamsinks and of risings respectively) and the sink/rising ratio.
The mean shortest distance of underground flow is derived from
the distances between each streamsink and its nearest rising below
its own level. Insufficient application has been made of these
indices yet to estimate their usefulness. The vadose index may in
fact conceal some hydrostatic rise in the saturated zone if there
are springs under pressure in the area.

REGIMES OF KARST SPRINGS

Most of the large springs of the world are karst springs though
there are also very large concentrations of spring waters in
 80 Karst

basaltic lavas (see Ollier 1969). In terms of present knowledge,

the group of springs at the source of the Manavgat east of Antalya
in southern Turkey yields the most of all with an average dis­
charge of 125-130 m'Vsec. One spring alone at low stage has a
flow of 40 m3/sec. Only 1000 km2 of the southern flank of the
limestone Taurus Mountains feed these springs which, however,
tap also 25-30,000 km2 of the northern slopes and the central
plateau.

Djebel Aaba •
El - Aaziz

— Wat er movement F F
,7 j P le is to c e n e alluvium Miocene -Oligocene
evaporites and limestone
f c . y l P le istoc e ne ba sa lt FTTvJ Oligocene marly
1 • V . 1 limestone
Miocene clays Eocene limestone

18 Artesian spring of Ras el ’Ain, Syria. After Burdon and Safadi 1963.

Just south of the Turkish-Syrian border, Ras el ’Ain with an

average discharge of 387 m3/sec is the effective head of the
Khabur River between the headwaters of the Tigris and the
Euphrates. It is a complex of thirteen springs (Burdon and Safadi
1963) in two groups, each receiving true artesian water along a
fault plane which taps aquifers in a synclinal structure (Fig. 18).
The main aquifer comprises Eocene limestones with intakes on
their own outcrop and on overlying basalts in the Turkish moun­
tains to the north. Higher in the syncline are Miocene limestones
and evaporites which probably furnish subordinate supplies from
the south. The northern group of springs has a chemical content
indicative of supply from limestones only whereas the southern
group probably get water from the evaporites also. The springs
 Drainage 81

appear as circular ponds, the basins of which may be due to col­

lapse of the surface Quaternary conglomerate into solutional
cavities in the evaporites beneath.
In Europe the largest limestone springs are those of the Stella
at Castella Sacile in Frioul, northern Italy, with a mean flow of
336 mVsec, but more famous are the sixteen springs of San
Giovanni, 25 km north of Trieste, which merge into the big River
Timavo for its short course to the sea. This is the resurgence of
the Reka River of the Slovenian karst, which sinks at Skocjanska
30 km away. The mean flow of the Timavo is 2625 mVsec. In
the United States there are the limpid but ‘boiling’ pools 10 m
deep of Silver and Blue Springs in northern Florida, each with
discharges of 14-15 mVsec.
Karst springs generally have a more even flow than karst rivers
which do not depend entirely on underground supply. The mean
monthly discharges of Ras el ’Ain vary only between 0 93 and
108 of the mean annual flow whilst the extreme minimum
recorded declines only to 0 77 of it. Flood discharges of the Silver
and Blue Springs are never more than three times the mean flow.
These characteristics are a product of the slowness of groundwater
movement compared with surface flow, the big storages under­
ground, and the tight bottlenecks which can restrict discharge.
However, not all of them are so well regulated. The Fountain
of Vaucluse itself, which is a big spring with a mean flow of
26-27 mVsec, has floods about seven times the mean discharge
and disgorges muddy water at these times. Other large springs
such as the Bourne in the Vercors plateau in the French Alps and
the Vidourle near Nimes can flood violently also. These have very
open cave systems behind them with big drops. Unreliable precipi­
tation regimes and liability to intense falls can also induce livelier
springs. In a context of modest relief in comparison with the
French examples, the Blue Waterholes on Cooleman Plain, N.S.W.
(Fig: 14b) have recorded a range of 017 mVsec to 2T9 mVsec
and muddy water is churned out in floods. Nevertheless in virtually
all circumstances karst springs damp down peak flows compared
with surface channels.
Similarly springs respond to rain more slowly than surface
streams but vary in this lag. A flood pulse can reach the Blue
Waterholes 24-48 hours after heavy rain with the most distant
streamsinks no more than 5 km away, though the peak in the
82 Karst

A Spring from bedrock

D Spring from terrace
alluvium
C Spring from river
bed sand
Contours in metres
IAlluvium
Rock outcrop

20 Metres
20 40 60 Feet

19 Blue Waterholes, Cooleman Plain, N.S. W., with springs resurging in

various manners

surface channel they feed will already have passed. Seasonal

changes in spring discharge may, however, follow the onset of a
rainier season by several weeks. The caves of northwest Clare
(Shaw and Tratman 1969) respond more quickly than the Blue
Waterholes— in a matter of hours— probably because of freer
circulation and little storage in these shallow, youthful systems. At
the other extreme with its long and deep artesian flow, Ras el ’Ain
changes its discharge very slowly and the control appears to be
surface runoff over the intake beds one year earlier.
 Drainage 83

Karst springs have dominantly carbonate waters except where

evaporites are involved as mentioned above. They vary in their
response in chemical content to seasonal and weather changes as
they do in their discharge regime. This is evident in the compari­
son Smith and his colleagues have made between risings in Mendip
and in northwest Clare (Smith and Mead 1962; Smith 1965;
Smith and others 1969). Three out of four major risings in
Mendip have remarkably constant calcium carbonate concentra­
tions around 235 mg/1. This is thought to be due to the tapping
of substantial bodies of groundwater which has reached saturation
equilibrium with high Pco2 acquired by water percolating through
grassland soil cover on Carboniferous Limestone. Inputs of
streamsink water from Devonian sandstone inliers are small in
comparison. The fourth rising behaves differently, varying in
concentration between 120 and 200 mg/1 inversely with dis­
charge; the flood pulses carried away less limestone per unit
volume. The northwest Clare risings behave in this second way with
individual risings varying as much as 120-205 mg/1 within a
month’s observations. This is attributed to a much higher propor­
tion of streamsink water from overlying shales to seepage water
from the surface of the limestone and to more direct flow from
sinks to rising with a proportionately smaller phreas.
The Blue Waterholes on Cooleman Plain behave like the Clare
risings seasonally, and again there is a high proportion of
impervious rock to limestone in the catchment serving the risings
(Fig. 14b). The Porth-yr-Ogof resurgence in South Wales (Groom
and Williams 1965) is similar, varying between 31 and 75 mg/1.
At low stage of river flow, the waters go underground from the
Ystradfellte streamsink to the resurgence, but in flood some of the
discharge proceeds beyond Ystradfellte to the entrance to Porth-
yr-Ogof and has a short underground flow through that cave
only. The River Mellte derives most of its water from Devonian
sandstone country and this does not reach saturation equilibrium
in a short course across the Carboniferous Limestone outcrop.
Despite low concentrations in flood flow, the solute load is abso­
lutely greater because of great volume. In 1960-1, 285 tons of
CaCO:{ were removed in 38 days of flood, 274 5 tons in 175 days
of normal flow, and 254-6 tons in 152 days of drought tapping
groundwater only, with high concentration nearly balancing low
discharge. In this way corrosion is evened out through the seasons
84 Karst

and varying weather much more than corrasional action which is

so strongly tied to floods.
Springs in the Jurassic limestones of the Cotswolds depend
entirely on seepage water and Smith (1965) reports that CaCO;i
concentrations remain virtually constant around 300 mg/1. Since
relief and climate are similar to those of Mendip nearby and the
limestones of both areas are very pure, the difference in equilibrium
concentrations must be sought in other lithological characters,
e.g. greater porosity of Jurassic limestones or in differences in
soil, perhaps with higher Pco^ in the Cotswold soil air.

RELATIONS OF SURFACE AND UNDERGROUND DRAINAGE

Cave exploration, water analysis, and water tracing have permitted

the relations between streamsinks and risings to be established in
many cases. Sometimes these prove to be those to be expected
from the surface conditions, including the disposition of inter­
mittent streams and of dry valleys, and at other times they are not
(Geze 1958).
The simplest case is where the underground course simply takes
over during normal and low flow the role of the surface course.
This is the case between the Owenterbolea streamsink and
St Brendans Well rising on the Poulnagallum Cave-Gowlaun
River drainage in northwest Clare (Collingridge 1969). Such
systems may well develop from a short underground course
initially, extended by the sinking taking place at successively more
upstream swallets and successively more downstream resurgences
taking over from one another in the course of time, but all remain­
ing along the line of the surface course which formerly functioned
for the whole regime.
The same close relationship can persist even where the surface
course ceases ever to take flood flow and the intervening dry valley
has become a string of closed depressions. The waters may still
emerge lower down precisely where it once flowed over the surface
or very close by. Part of the South Branch of Cave Creek on
Cooleman Plain has drought, normal, and flood sinks successively
lower down its valley, which continues meandering onwards and
deepening to join the North Branch gorge. But whereas the North
Branch flows intermittently over the full length of its valley to the
Blue Waterholes, part of the South Branch valley below its sinks
 Drainage 85

has no channel at all though it still has a continuous downward

fall. Nevertheless its waters emerge at the Blue Waterholes, some
of which are actually in the bed of Cave Creek. The Bathhouse
Cave creek of Fig. 32 emerges from Coppermine Cave in the
Yarrangobilly valley T3 km away. This outflow point is about
425 m north of the junction with the main river gorge of a dis­
organised dry valley which runs down from the Bathhouse Cave

Shale
I 4 Moraine
^----- St r e am and streamsink
------- I nt er mi t t ent s t r e a m
Dry valley
Cave, in parts unroofed
O Rising
1 Kilometre

20 Relationship of Cullaun Caves, Clare, Ireland, to surface streams and

dry valleys. After Ollier and Tratman 1969.
86 Karst

blind valley, but essentially there is still good correspondence

between the surface and the underground drainage.
However, eventually underground flow becomes independent of
the former surface flow and the structure of the limestone can very
frequently be seen to be the cause of the severance. A stage of
partial departure from the surface drainage pattern is exhibited by
the Cullaun caves on the west of the shale ridge of Poulacapple in
Clare, Ireland (Ollier and Tratman 1969). Surface streams in
valleys in the shales are continued southwestwards and westwards
to the Killeany valley and spring along dry valleys cut in glacial
moraine down to the Carboniferous Limestone (Fig. 20). After
crossing on to the limestone, streams enter shallow caves which
follow the dry valleys at first. Then they are deflected SSW along
major joints. In this way they pass under divides south of their
originating valleys and join the Killeany waters south of the
former surface drainage. Indeed the southernmost Cullaun series
4 and 5 extend beneath the shale ridge of Knockvoarheen border­
ing the Killeany lowland on the south. Where these waters rise is
not known but it is not at Killeany.
Faults sometimes provide favourable zones for the development
of underground drainage and may deflect flow along themselves
till they intersect a deep valley where resurgence can take place
at a point far removed from the original stream course. Geze
(1958) cites the example of the Buege, which sinks in a channel
still in active use part of the year to join it to the River Herault,
but a NNE-SSW fault directs the underground flow in that direc­
tion to reappear at the Source des Cent Fons in the Herault gorge
at a point 10 km downstream of the surface junction (Fig. 21a).
In folded structures, especially where other rocks are included
in them, the fold axes may deflect underground drainage away
from the surface courses. Synclinal troughs in particular act in
this way. At Mole Creek, Tasmania, the general trend of the
drainage is northeastwards from the Great Western Tiers, the
northern flank of the Central Plateau, across the axial trends of a
synclinorium in Ordovician limestone and sandstone (Fig. 21b).
Marakoopa Creek is one of these streams and has a surface
channel across a syncline in the limestone and thence through
Sensation Gorge in an anticline of underlying sandstone. However,
it only rarely passes through the gorge; more frequently it sinks
just short of it and follows an underground course SSE along the
 D ra in a g e 87

S urfa c e stream
In te r m it te n t
surface stream
Underground river
Rising
--------j S tr eam sink
I Impermeable rock
] Limestone
C li ff s and scarps
~ T - Fault
A ntic lin e
' J L . Syncline
S u r fa c e divide

21 S tru c tu ra l deflection o f u n d erg rou n d drainage fro m surface pattern

(a) By fa u lts in the Causse de la Seile, France. A fte r Geze 1958.
(b) By fo ld s at M o le Creek, Tasm ania.
88 Karst

syncline to join Sassafras Creek, which then takes the water NE

just round the nose of the pitching anticline of sandstone.

THEORY OF KARST HYDROLOGY

Although caves allow more information to be obtained about

underground water movement in karst than in other rocks, the
fact remains that the former is the less well understood. Indeed
the fundamental nature of water circulation in karst has stayed
controversial since the late nineteenth century when two conflict­
ing schools of thought developed in Europe. Since it is so basic to
the study of karst landforms, the essence of the problem will be
set out at this point, even though it must presume some of the
geomorphic data to be discussed later.
First it will be useful to outline some of the ideas which derived
from the study of permeable rocks other than those giving rise to
karst. Water, which escapes downwards from the zone of moisture
in soil, passes first through a zone of aeration where pores are
only transitorily filled with water and thence into the zone of
saturation where it displaces all the air because hydrostatic
pressure is greater than atmospheric pressure. The upper surface
of this saturated zone is the watertable in which hydrostatic
pressure and atmospheric pressure are equal. The watertable
parallels the land surface in subdued fashion and there is move­
ment of the groundwater in accordance with the slope of the
watertable, i.e. down the pressure gradient. The Frenchman.
Darcy, experimented with water movement through columns of
sand to discover his law that flow through such a permeable
medium is proportional to the pressure gradient or the loss of
‘head’.

where Q = flow
K = coefficient of
permeability
H A = cross-sectional area
Darcy’s Q = K A of flow
Law L
H loss of head between
two points
L = distance apart of
these points
 Drainage 89

The total head is the sum of the head due to gravity (elevation
head) and the head due to pressure of the water column.
Where a bed which transmits the water (an aquifer) is not
overlain by impermeable materials, the groundwater is said to be
unconfined and the watertable is everywhere free to rise and fall
with seasonal variation in the amount of water percolating
downwards.
However, the groundwater may permeate beneath an imper­
meable stratum (an aquiclude). Here the groundwater is confined
and is not free to rise indefinitely with accessions of water from
above. If the aquiclude is inclined and bores are put down through
it, water will rise up them to heights above the confined zone. This
is the artesian condition and the imaginary surface to which the
bore waters would conform is known as the piezometric surface.

Karst watertable and single aquifer

The view of karst hydrology which until recently has dominated
Anglo-American literature hinges on the concept of a watertable
and is particularly associated with the name of A. Grund (1903,
1910a), who was primarily concerned with the Dinaric karst.
J. Cvijic (1893) originally thought along the same lines but later
modified his ideas considerably (Cvijic 1918, 1960).
It is assumed that underground water in karst behaves in essen­
tially the same way as it does when moving through other pervious
materials such as sands, gravels, and sandstones. Thus below the
soil moisture zone, three relevant hydrological zones can be distin­
guished (Fig. 22).

Watershed in
watertable
Intermediate River
Vadose zone

Saturated or phreatic zone

Groundwater base
a Dry season watertable s Permanent spring
a' Wet season watertable s' Temporary spring

22 Karst hydrologic zones based on concept of a watertable and a single

aquifer. After Cvijic.
90 Karst

(a) There is an upper or vadose zone in which water moves

dominantly downwards after rains but which can be wholly or
partly dry. It is important to distinguish between vadose seepage,
which refers to water from rain or the soil percolating downwards,
often in confined fissures, and vadose stream flow where water,
gathered into concentrated runoff at the surface, moves streamwise
downwards and laterally in open passages.
(b) In the lower saturated zone all cavities are permanently
full of water and the top of the zone forms a watertable with
underground watersheds similar to watertables in other rocks but
thought to be flatter. This is the phreatic zone and this permanent
body of water is termed the phreas. Springs occur where the water-
table intersects the surface.
(c) There is an intermediate zone where the cavities are inter­
mittently flooded to capacity, a zone through which the watertable
rises and falls. This explains the behaviour of poljes (large, flat-
floored closed depressions) which flood in winter when the water-
table rises and which may be fed from the same openings that
drain them in the dry season.

Karst conduits and multiple aquifers

The very different permeabilities assigned to different rocks
(Table 1) indicate that the transfer to karst of concepts in part
derived from Darcy’s experiments with sand may be misleading.
A critical difference is that in karst the underground water is not
moving chiefly through intergranular pores (to be regarded as a
single aquifer) but through narrow fissures and large caves (to be
regarded as multiple aquifers: Thrailkill 1968). So it was specu­
lated quite early that there is no watertable in the ordinary sense
in karst. Von Katzer (1909) was one of those who maintained
this for the Dinaric karst but the strongest attacks came from
speleologists, notably E. Martel (1910, 1921).
Wells and bores put down in limestone close together often
reach water at very different levels; dry holes occur cheek by jowl
with good yielding ones. Similarly tunnels driven through limestone
reveal dry and water-filled fissures close together. Frequently
water-filled cavities overlie empty ones. The tracing of under­
ground water movements by dyes, etc., has often shown that
underground water connections can cross one another and pass
under surface streams without interference (Fig. 23). In a particu-
 Drainage 91

23 Crossing of underground and surface drainage, and other geomorpliic

features at Malltam Cove, Craven, England

lar karst area, poljes at the same level behave differently, some
flooding, some not. Sometimes a high-lying polje floods before
lower ones close by. When some poljes flood, their streamsinks do
not reverse their flow but take in more water. Polje lakes have
been drained by deflecting streams from them; this could not occur
if the lakes were due to the polje intersecting the watertable. In the
particular case of Livanjsko Polje, there are three separate flat
92 Karst

floors, each with its own spring-surface channel-streamsink

system. Artificial opening up of their streamsinks in the late nine­
teenth century had different effects on each, demonstrating their
hydrologic independence.

Solution Pothole
doline Permanent
risings

rst 'v°ter circulation

a Solution opened joints and bedding planes with seepage water
b Potholes and joints with seepage and stream flow
Cove with free surface stream .filling in floods
■■■■■ Permanently water filled caves
C Solution opened joints and bedding planes permanently
waterfilled ,
Based on Cavaille

24 Karst hydrologic system based on concept of independent karst conduits

and multiple aquifers. After Cavaille.

Therefore Katzer, Martel and others maintained there is no

watertable, only independent underground conduit systems
operating like rivers but in a three-dimensional space (Fig. 24).
Parts of these underground systems are free-surface streams, only
partly filling caves and flowing under gravity alone. Other sections
of the streams fill the caves and water can bank up behind
constrictions, developing hydrostatic head. The hydrostatic
pressure may drive water uphill in parts of the systems; thus in
the French Alps rising water currents under pressure are known
to carry pebbles up with them over heights of at least 100 m. In
La Luire pothole, water can rise with a speed comparable with
those experienced in surge shafts in hydroelectric systems.

Compromise views
Nevertheless, there remained difficulties for this contrary view
of karst hydrology. In some areas the numbers of springs are very
 Drainage 93

much fewer than points where water sinks and they lie at very
similar levels despite lack of obvious structural cause for their
equivalence. So elaboration of ideas about ‘independent karst
conduits’ followed, particularly by O. Lehmann (1932) who
proposed the following evolution in the hydrological system in
karst (cf. also Cvijic 1918).
Initially when a dense limestone is first stripped of impervious
cover or exposed to the atmosphere by uplift from beneath the
sea, it will not be permeable. So to begin with a normal surface
drainage with a river valley system develops over it. Then seepage
water opens up planes of weakness and links up any tectonic
cavities there may be; as a result the area becomes permeable.
This ultimately produces a mature karst hydrology, with many
independent but complicated, branching, and net-like systems of
passages and chambers. Free intercommunication is not estab­
lished between them and there is no single watertable. Each
conduit system has many entrances but streams join underground
to feed a single or a few outlets. Because of variation in cross-
section, the pressure flow in filled passages is complex and not
always downwards and outwards. Under hydrostatic head, water
will rise higher up shafts leading from passages of large cross-
section than up those from ones of small cross-section because of
the inevitably varying velocities (Fig. 25). This can lead to back­
ward flow at higher levels. Such systems may explain the incon-
gruent behaviour of neighbouring poljes. Thrailkill (1968),
however, calculates that this cross-section effect will be quite small.
Further karst development according to Lehmann involves
widening of passages and removal of intervening obstacles; fissures
link neighbouring caves and free-surface flow predominates in
very elaborate systems. Eventually projecting limestone masses are
reduced in size and riddled with passages. Lehmann regards this
old stage as degenerate karst hydrology, when something approxi­
mating to a karst watertable is established. To avoid begging the
question of the nature of karst hydrology, some authorities, e.g.
Sweeting (1958), write of ‘rest levels’ instead of ‘watertables’,
which have connotations that are inappropriate but hard to avoid.
The most important recent work bearing on this general
question has been that of Zötl (1957, 1965), employing spore
drift in the Austrian Calcareous Alps. He examined the under­
ground drainage of high limestone plateaux surrounded by deep
94 Karst

25 Effect of varying cross-section and hydrostatic pressure on water levels.

After Lehmann 1932.

valleys, such as the Dachstein and the Totes Mountains, which

cannot be regarded as in an old stage of geomorphic development.
Spores fed into streamsinks near the margin of the Dachstein
plateau reappeared in a few springs close together at the foot of
it on one side (Fig. 26a). But when sinking waters farther and
farther into the plateau were tagged in this way, springs on wider
and wider sectors of the margin were reached by them. A sink in
the centre of the eastern Dachstein distributed water more or less
throughout it and similar broad spreads were determined for the
Totes Mountains. On this evidence Zötl did not revert to Grund’s
simple scheme but argued that underground river systems estab­
lish interconnections earlier than Lehmann conceived. The major
underground rivers remained the dominant arteries and any spores
 D ra in a g e 95

Spore in s e r tio n po in t o
C o l l e c t i n g 1 s tre a m a.

po in t in J sprin g •
Cont ou rs in m e tre s

Dachstein P la te a u

5 K i lo m e tres
5 Miles

Dry
va lle y

/Z o n e of

D ro u g h t w a t e r t a b l e CHALK

Ground w a t e r

-= = MARL
~~JH

26 (a) U nd e rg ro u nd drainage connections d e term ined by spore d r ift in

Dachstein plateau, A u s tria . A fte r Z ö tl 1957.
(b) U n d e rg ro u nd drainage in the C h a lk o f N o rth e rn France. A fte r
P in cliem e l 1954.

I
96 Karst

put into streamsinks close to such arteries would reappear at a

single point or a few outflow points.
Zötl's interpretation of his results resembles Pinchcmel’s for the
hydrology of the Cretaceous Chalk country of northern France
and southeast England (Fig. 26b). In this very porous limestone,
evidence from bores and wells has generally been accepted as
according with a single aquifer and continuous watertable concept
(e.g. Balchin and Lewis 1938). Flowever, Lewis also commented
on the wide spacing of springs along the scarp foot of the Chalk
south of Cambridge, England, which suggested the gathering of
considerable underground drainage to feed central outflow points,
possibly along major faults or joints. Pinchemel combines the two
elements in his scheme, with a normal watertable reflecting the
surface relief but with concentrations of linear flow along joint
systems beneath the dry valleys. Ineson’s maps of permeability
in the Chalk of England (1962) support this concept but also
show greater permeability along anticlines and domes.

Seepage Stream

f I I I i 4 I t I I I

W at ertable I
-------

S inkholes Stream

Seepage Seepage

100m

10km

27 (a) Pattern of deep phreatic movement. After Davis 1930.

(h) Streamlines of underground flow with irregular inputs. After
Thrailkill 1968.

Deep phreatic flow

Another aspect of groundwater movement in pervious but non­
karst rocks is deep flow below the watertable (Davis 1930;
 Drainage 97

Rhoades and Sinacori 1941; Fig. 27a). Deep borings in lime­

stone beneath dam sites in the Tennessee Valley substantiated the
presence of water-filled cavities as much as 100 m below the river
bed (Moneymaker 1941). The theory is that water descends to
great depths beneath interfluves and rises back to the surface under
the valleys where springs are located. Glennie (1954a) termed
waters rising from such deep phreatic paths ‘artesian’; this is
inapplicable in a strict sense but it serves to remind us that com­
pacted limestone can virtually act as its own aquiclude. Cave
diving has shown that waters do rise 50-100 m in some major
springs but amongst European speleologists at least the view
prevails that they are fed by largely independent arteries.
A recent elaboration of the pattern of flow below the watertable
is that of Thrailkill (1968; Fig. 27b), allowing for an irregular
distribution of inputs of vadose seepage and vadose flow. This
pattern is based on the idea that important vadose flows more
distant from the valley lines will be displaced laterally and
vertically by seepage and flow inputs close to the outflows. This
concept seems to give insufficient weight to Zötl’s empirical
evidence.
There is, however, a great difference in conditions between the
high alpine plateaux with complex structures of Zötl’s studies and
the nearly horizontal limestone and modest relief of Kentucky
where Thrailkill derived much of his experience. Nor are these
two examples sufficient to indicate the wide range of relief,
geological, climatic and historic contexts encountered. For
example, in the Nullarbor Plain, with its horizontal limestone, low,
flat relief and a highly porous limestone at the levels at which
water rests, nearly motionless lakes in caves suggest a single
aquifer and an almost horizontal watertable very little above sea
level at distances of 20-30 miles inland in a free karst situation.
Nevertheless the occurrence of widely spaced, big cave systems
which possess these lakes again points to concentrated lines of
flow. In this instance they may be virtually inactive today and
inherited from Pleistocene periods of more effective, if not greater
precipitation. It is a case which does not fit simply into the
different schemes set out so far. White (1969) has proposed seven
kinds of karst hydrogeologic systems for regions of low to
moderate relief based on the areas he knows best in eastern and
central United States (see p. 209 and Fig. 65). It is clear that a
complete body of theory in this subject cannot be expected yet.
 VI

SURFACE LANDFORMS

The larger surface landforms of karst are the subject of this

chapter, in particular the special attributes which valleys take on
there and the closed depressions which are the most characteristic
features of all in karst. Tectonism, vulcanicity, glacial erosion and
deposition, deflation and aeolian deposition, glacier and ground
ice melting also fashion closed depressions but not with as much
variety or as commonly. Usually these other origins are readily
distinguishable but composite forms can be problematic.

GORGES

Rivers traversing karst areas and those rising within them develop
gorges more frequently than do those in other rocks, given similar
general relief and climatic conditions. The Grandes Causses of the
Massif Central of France are divided into four separate plateaux
by the gorges of the Lot, Tarn, Jonte, and Dourbie, 300-500 m
deep. In New South Wales, Bungonia Creek descends about 450
m in some 6 km to join the rejuvenated Shoalhaven River. In so
doing, it runs initially along Devonian volcanics, including toscanite
lavas, then turns at right angles to cross the line of strike, first
of nearly vertical Silurian limestone and then of underlying
Ordovician shales. The last give rise to a steep-sided but still
V-shaped valley; a gorge is cut in the lavas but it is outshone by
a slot canyon across the limestone, which has practically vertical
walls about 300 m high (PI. 15). In young folded mountain
ranges, limestone gorges of much greater dimensions occur such
as those of the Verdon in the southern French Alps or of the
Strickland in New Guinea. However, the same habit persists in
minimal local relief. In the Limestone Ranges of West Kimberley,
98
 Surface Land forms 99

#. C ■M ?

' ■■■w.*•« l;;WJb ■

%
-ft § r '■
'V \ - ; $.

f Jr
'
: i\
J* v ■

/5 Bungonia Gorge, N.S.W. Canyon eroded by Bungonia Creek across

strike of Silurian limestone dipping nearly vertically.

for example, there are Geikie Gorge on the Fitzroy River, Wind-
jana Gorge on the Lennard River, and Mount Pierre Gorge on
Mount Pierre Creek, amongst others; their walls are vertical but
the relative relief is everywhere less than 100 m.
The prevalence of gorges in karst is primarily due to the balance
100 Karst

between slope processes and river incision favouring the latter to

an exceptional degree. Longitudinal profiles in karst tend to be
flatter than in many neighbouring rocks, probably because solu­
bility permits more lowering of talwegs than does corrasion, even
in the case of massive and compacted limestones and dolomites.
Dissolved load is easier to transport than clastic load and a flatter
river profile results in limestone areas, and so gorges can be
remarkably persistent from entry into limestone to leaving it.
However, the prime factor in their formation is failure of slope
processes to flare back the valley sides to a V cross-section.
Marked infiltration and reduced runoff minimise slope wash
and many kinds of mass movement which would otherwise bring
about such change of form. Additionally, cave roof collapse has
been frequently offered as an explanation of gorges in karst
terrain. This is no doubt the correct explanation in some cases—
in the Rakova Kotlina northeast of Postojna, Slovenia, in which
there are two natural bridges; between the arch and the main
Maungawharawhara Cave in the King Country, New Zealand;
between the first and second sections of Stockyard Gully Cave,
southwestern Australia. However, many gorges previously regarded
as collapsed caves are no longer interpreted as such. This is the
case with Cheddar Gorge, Mendip, England, where surface
rejuvenation forms can be traced into the gorge in which the caves
are relict phreatic caves (Ford and Stanton 1968).

MEANDER CAVES

A contributing factor to the formation of gorges in karst is the

effectiveness of solution in lateral action by rivers. Corrasional
undercutting of valley sides encumbers a river with debris but
corrosion does not. Consequently meander caves are better
developed in karst than elsewhere, though it must not be thought
that they are very frequent or very important landforms. A good
example is provided by Verandah Cave, Borenore, New South
Wales, which is situated in the undercut cliff of the concave bank
of an ingrown meander of Boree Creek (Fig. 28). The Verandah
itself is a remnant of a higher abandoned meander cave corres­
ponding to a rock terrace remnant on the upstream side and is
much less impressive than the 30 m deep active meander cave.
 Surface Landforms 101

30 M e tre s

100 Fe e t

C ontour in te rv a l 5 fee t

-------- Cave w a l l C liff V e ra n d a h A llu v iu m

Me tres
R em nan t o f oLder m e a n d e r cave r 36
re la te d to te r ra c e A

Verandah Cave
— C a l c i te p i l l a r
S l ip o f f slope
- l BOR E E CREEK

28 M ea n d e r cave on Boree Creek, Borenore, N .S.W .

NATURAL BRIDGES

Natural bridges are more common in karst valleys than in others

but they vary much in form and genesis. Cleland (1910) made a
terminological distinction, which has not established itself in the
literature, between natural bridge through which a river runs or
has run, and natural arch where the span does not cross a valley
but perforates buttresses, spurs and ribs of rock as a consequence
102 Karst

of weathering, e.g. Porta di Prada, La Grigna, north Italy. The

distinction between bridges and caves must be an arbitrary one.
As satisfactory as any is probably the criterion that daylight
reaches through a bridge. The Arch Cave at Abercrombie, N.S.W.,
which is 180 m long, is at about the limit of bridge on this defini­
tion. The larger and straighter the way beneath a bridge the greater
length can still be lit from outside.
Some bridges are due to a very narrow band of limestone lying
across a stream course. A cave developing here as the river incises
is very likely to become nothing more than bridge. Steeply dip­
ping beds are most favourable to this and the Grand Arch at
Jenolan, New South Wales, is about 140 m wide in an outcrop of
180 m, dipping at 60°. The greatest span is 50 m and the maxi­
mum height is 20 m above Harry Creek which flows through
the Arch.
Other bridges are the surviving elements of former cave roofs,
which may have been of considerable length as in the instances of
the Rakova Kotlina and the Maungawharawhara given above
(Fig. 29a). Frequently natural bridges of this type seem to be
associated with underground river capture with the valley below
the bridge being much more deeply incised than that a little way
above it. This is the explanation which Woodward (1936: cf.
Wright 1936) gives of Natural Bridge on Cedar Creek near
Lexington and of Natural Tunnel on Stock Creek near Clinchport,
both in Virginia, U.S.A. (Fig. 29b).
Self-capture can also lead to natural bridges. The simplest case
is where meander caves on one or both sides of a meander spur
finally breach the wall of limestone between the two beds of the
river. But generally water escaping from the river will have pierced
the spur by solution along joints and bedding planes before the
edge plane action of the main body of stream water will have
destroyed it bodily. London Bridge, a very beautiful if small span
on Burra Creek near Queanbeyan, New South Wales, provides a
very straightforward example, where the creek and valley are
aligned along the strike in a general way but meanders necessarily
cross it. One meander spur includes a narrow band of nearly
vertical limestone. The long bend was cut off by cave development
across the neck of the spur where the limestone was so favourably
placed (PI. 16). An alluvial fan built by a tributary has partly
filled the old meander.
 Surface Landforms 103

29 (a) Natural bridges due to collapse of cave roof. After Maksimovich.

(b) Natural bridge due to river piracy, Natural Bridge, Cedar Creek,
Virginia. After Woodward 1936.
16 Natural bridge due to river self-capture across meander spur. London
Bridge, Burra, N.S.W.

Present alluvial plain

and river course

\ \

a l l
106 Karst

Bridges can also develop without meandering by self-capture at

rejuvenation heads. Waterfalls can cascade down the steep drop
in the longitudinal profile where joints may open up behind and
engulf more and more of the flow to leave a span of rock over a
retreating and degenerating fall. Cleland (1910) has proffered
this as an alternative explanation for Natural Bridge, Lexington,
Virginia. Other bridges are involved in karst river deposition under
which heading they may be more conveniently considered.

CONSTRUCTIONAL ACTION OF RIVERS

Because of the high concentration of carbonate they may carry,

karst rivers have a special capacity for constructional activity
through precipitation of a portion of their chemical load. This
may be due to diffusion of carbon dioxide to the atmosphere after
emergence from beneath the surface, to evaporation, or to the
intervention of plants. Many kinds of plants secrete calcareous
skeletons or carbonate is deposited around their organic tissues and
these can accumulate to form significant deposits. At Plitvice on
the River Korana in Yugoslavia, Pevalek (1935) distinguishes
four kinds of tufa formed from and around plant remains; two are
dominantly formed by mosses (Cratoneuron and Bryum), one by
the blue green alga Schizothrix, and another depends on a grass
(Agrostis) together with the alga. Each is associated with particu­
lar microrelief and flow conditions. In other parts of the world
many other plants take on similar roles, including different orders
such as the Characeae. Organic debris washed in— leaves, twigs—
is also incorporated.
Tufa is the best term to use for the porous primary deposits
formed in this way and travertine may be restricted to the more
solid and crystalline carbonate deposits from flowing water, which
may be secondary in nature through the infilling of the voids in
tufa, or primary deposits from the water without plant participa­
tion. Another term is calc-sinter which some would restrict for the
last category mentioned.
Plant growth, evaporation, and C 0 2 diffusion are all promoted
by aeration accompanying vigorous turbulence. Therefore deposi­
tion will preferentially take place where the water flows over any
initial irregularity and a barrier or dam gradually builds up there
(PI. 17). This in turn favours further accumulation on the front
17 Tufa dam formation in McKinstry's Canyon, Guadalupe Range, New
Mexico

where the steep slope involves frothing and bubbling, and the
thinner layer passing over the actual top of the barrier is also
conducive to growth. So barriers and waterfalls develop across
karst rivers through their own action. These constructive water­
falls (Gregory 1911) may advance down the valley leaving a
broader fill of travertine behind. The barriers can become over­
hanging with pendulous curtains of tufa and caves enveloped
behind.
Alluviation may take place above the growing barriers. This has
happened at the 20 m high Topolje Falls on the River Krka near
Knin in Yugoslavia, above which there is an alluvial plain. The
same is true lower down the Krka near Sibenik where there is a
series of tufa dams totalling 40 m. However, at Plitvice the barriers
have grown up too rapidly for deposition above to keep pace with
them so that a number of small lakes have formed along 5 km of
the valley of the River Korana (Fig. 30). The largest is nearly
one km- in area and the deepest is 50 m deep. In one of the lakes
108 Karst

30 Plitvice Lakes, River Korana, Yugoslavia, due to travertine and tufa

dams. After Gavazzi.
 Surface Land forms 109

there are two submerged barriers due to a lower barrier growing

higher than ones farther up valley and so incorporating two
formerly higher lakes.
Phases of deposition may alternate with phases of erosion. Emig
(1917) distinguishes three phases of travertine formation in the
Arbuckle Mountains of Oklahoma. The first period of deposition
gave rise to higher and wider falls than the present Turner Falls
on Honey Creek and Prices Falls on Falls Creek.
Natural bridges occur in travertine (Cleland 1910). Some are
due to actual construction of travertine above the stream through
splash and spray till it meets, e.g. on Pine Creek, Arizona, but
others are left in the air by the removal of underlying gravel on
which the travertine was built up.

SEMIBLIND VALLEYS

Persistent sinking at a point on a river’s course leads to a lowering

of the bed there through bedrock solution and engulfment of
sediment load. Below the sinking point the river has less power
to erode its bed by solution or mechanical action. Gradually an
upward step or threshold develops in the longitudinal profile and
the underground course enlarges its capacity to accept more of
the flow. Eventually it can take the whole flow at normal stages.
Flow is continuous down the valley only after heavy rains or
snow melt when the streamsink cannot accept the flood discharge.
Then water banks up behind the counterslope to the level of the
threshold and overflows. The bed below the threshold becomes
vegetated to differing degrees and the gravels or bedrock become
weathered subacrially in varying amount according to the
frequency of this intermittent channel use. This is the condition of
the semiblind valley.
Figure 31 represents an instance from Cooleman Plain, New
South Wales, where a small creek flowing from a granodiorite hill
sinks in normal conditions after flowing for 150 m over the out­
crop of Silurian limestone into an earth hole in a small, circular
alluvial flat. A low, grassy threshold, about 3 m high, separates
this depression from a gravel stream bed which is only rarely
followed by overflow.
1 10 Karst

*£££> Rock outcrop

4080

0P0P

, i j — ___
100 Feet (Contours in f e e t ) - 0 ^
Normal streamsink
-— Perennial watercourse
4090
flood water 4027 5 4017
Stream in swamp
100 Metres
300 Feet

31 Semiblind valley on Cooletnan Plain, N.S. W.

BLIND VALLEYS

Eventually a sinking stream cuts down its bed so far and enlarges
its underground course to such an extent that the stream is always
completely engulfed and never flows beyond. Thus a blind valley
is produced, closed off at its downstream end. The height of the
closure may be only a few metres in the case of a small stream
but may reach into the hundreds of metres, particularly with large
streams. The larger the stream the more likely it is to disappear
into a penetrable cave. After floods, a temporary pond or small
lake may form in a blind valley. The example of Fig. 32 is one of
many at Yarrangobilly, New South Wales, where a strike belt of
Silurian limestone forms a plateau between the Yarrangobilly
River gorge and the steep slopes of a porphyry range. This stream
sinks into the Bathhouse Cave below crags in a steep counterslope
 Surface Landfonns 111

THE BATH HOUSE

/ CAVE / " ' " - i

100 M e t r e s
3Ö0 Feet \ (Contours in f e e t ) \

P la te au surfa ce

DOLINE DOLINE BLIND VALLEY

Stream* inx
xswamp v
- 3400' THE BATH HOUSE
CAVE Y / A A /
A A /A A K

-3300'

^X a a W a o a a ^ M etres
'AAAAAAAAAAAAAA* 3 b 0 F e e t

32 Blind valley at Yarrangobilly, N.S.W.

I 5 m high. The blind valley closes off only about 45 m from the
limestone and porphyry contact. It is continued by a shallow
valley some 8 m deep in the limestone plateau and interrupted by
small closed depressions. This is the much modified former
onward course of the stream. Sometimes a blind valley has a series
of closed depressions into which the stream spills successively
after banking up in flood, each one providing additional entry into
the underground conduit. There are several cases of this on
Cooleman Plain, N.S.W.
Blind valleys sometimes reach far into the karst area and are
cut deeply below its surface. Many factors such as the characteris­
tics of the karst rocks, the discharge and the chemistry of the
sinking stream control the dimensions of blind valleys. Gams
(1962, 1965) has made a comparative study of a number of
112 Karst

blind valleys in the Slovenian karst, some of which are formed by

streams flowing from headwaters on the impervious Eocene
Flysch. Some of these reach depths of 200 m. He concludes that
there is a relationship between the total carbonate content of the
stream where it crosses on to the limestone and the nature of its
blind valley. Those with low carbonate contents have longer and
wider blind valleys; the ratio of valley lengths before and after
crossing on to the limestone is 37:1. In a second group with
much higher carbonate concentrations, the blind valleys are
narrower and shorter with an equivalent ratio of 11:1. The width
of blind valleys may also vary according to their alluvium. Where
a blind valley acquires an impervious alluvial fill, this will seal
off the bottom from corrosion which then becomes directed against
the sides of the valley, thus widening it. With pervious alluvium
such as much of the Pleistocene cold period gravels of the Yugo­
slavian karst, solution can still go on beneath the openwork fill,
continuing to deepen the valley without much widening.

STEEPHEADS

Some minor geomorphic aspects of springs have already been

touched upon in Chapter V. Perhaps their most important mor­
phological trait is the way springs, both of exsurgence and resurg­
ence type, occur at the heads of valleys which begin very abruptly
(PI. 18). These are usually short valleys in the margins of plateaux
or on the flanks of mountain ranges. The Fountain of Vaucluse in
Provence arises beneath a cliff of 200 m height at the head of
such a valley as is implied in the derivation of the name— val
clos. Another popular name for such cul-de-sac valleys in several
languages is ‘World’s End’ but the American term steephead is
becoming the customary word in the English language. Often these
valleys are gorge-like for some distance downstream and frequently
impermeable underlying rocks crop out along the bottom beneath
cliffs of limestone or dolomite. A distinction is sometimes made
between steepheads incised to an impervious basement and pocket
valleys of the same general nature but within the karst rock
outcrop.
Steepheads can form in more than one way. The form may be
due solely to spring sapping. A spring undermines the slope or
cliff above it, rock and soil gravitate into the spring and are
 Surface Land for ms 113

removed in solution or as clastic load. In this way there is head-

ward recession of the valley. Alternatively the gorge may be the
result of collapse of substantial lengths of cave roof. In this case
it may not have developed headward but in an irregular manner,

18 Spring at Sogöksu, Turkey, rising at head of short, blunt valley or

steephead
1 14 Karst

19 Dry valley, Cooletnan Plain, N.S.W., sunk below corrosion plain of

Silurian limestone. Valley contains Pleistocene periglacial-fluvial deposits
and small dolines occur along valleysides.

separate sections of collapse being eventually united. If collapse is

recent, evidence in the form of natural bridges may survive but if
it is ancient, all such evidence may have disappeared and criteria
to distinguish between the two modes of origin may be lacking.

DRY VALLEYS

Dry valleys are similar in many respects to ordinary river valleys

but there are no stream channels in their floors (PI. 19). They are
by no means restricted to karst; on many rocks there are short dry
sections at the heads of valleys subject to throughflow and minor
tributaries may be entirely of this nature, channels only appearing
when overland flow gathers sufficiently to cut and maintain them.
Dry valleys can be longer and may form branching systems on
other permeable rocks such as sandstone and pumice, but these
characteristics are accentuated on limestone and dolomite.
One kind of dry valley usually presents little difficulty of inter-
 Surface Landforms 115

Y / / \ Limestone
j 1 Impermeable
I____I rock
20 Kilometres
20 Miles

33 Dry valley of Pazinski Potok, Istria, Yugoslavia. After Roglic 1964a.

pretation but can be the most spectacular of this category of

landform. After sinking on entry into karst terrain, allogenic river
valleys are often continued by dry valleys, which may pass right
through the pervious rock outcrop. In the Istrian plateau (Roglic
1964a), three river systems, which have eroded wide valleys in
the Eocene Flysch, drain southwestwards into the limestone (Fig.
33). The northern, the Mirna, and the southern, the Rasa, cross
the karst in canyons to the sea. The middle system sinks at the
limestone but Pazinski Potok, a dry valley of canyon form, carries
on to the coast.
Very many karst regions can provide similar examples. In
Craven, the Watlowes is a craggy dry valley whereby the stream
from Malham Tarn on Silurian slates formerly flowed across the
limestone to cascade over the 75 m cliff of Malham Cove (Fig.
23). In south central Java, the Sadeng dry valley winds through
the Gunung Sewu conekarst to the south coast (Fig. 34). It is the
former course of the Solo River now flowing out of the Batoeretno
basin north of the karst (Lehmann 1936). Beginning 20-25 m
116 Karst

34 Reversal of River Solo in Batoeretno basin, Southern Java, causing

Sadeng dry valley through the Gunung Sewn cone karst (a) before
reversal, (b) after reversal. After Lehmann 1936.

above this basin, the dry valley is 250-300 m wide and 150 m
deep; its long profile’s southward fall is interrupted by elongated
basins with shallow lakes or telagas separated by gentle swells
as much as 30 m high. Figure 32 shows the beginning of a dry
valley system which continues the line of the Bathhouse Cave
blind valley to the Yarrangobilly River gorge. This is much inter­
rupted by transverse rock barriers and shallow closed depressions,
some of which represent former sinking points of the creek when
it flowed further across the limestone plateau.
Semiblind valleys may be continued by dry valleys instead of
by intermittently active stream beds when overflow is infrequent
enough to maintain a channel. Water may rise up through the
bottoms of depressions along such dry valleys when extreme flood
 Surface Landforms 117

35 Dry valley systems, Peak District, England. After Warwick 1964.

conditions re-establish an onward flow. This is the situation with

the Owenterbolea streamsink-St Brendans Well rising system
mentioned earlier (p. 84). Upper St Brendans is the chief inter­
mittent rising along the dry valley.
More problematic are the branching systems of dry valleys
within karsts where there is an absence of stream channels
throughout. Most of the Peak District is covered by such systems
and there are few river channels in the area other than the
118 Karst

allogenic Wye, Dove, Manifold, and Derwent (Fig. 35). These

dry valleys often begin in shallow, bowl-like basins, which
develop into rock-walled valleys and gorges. Many smaller dry
valleys hang above major dry valleys or the allogenic river
valleys. The general pattern of the dry valleys resembles that of
surface streams on surrounding impervious rocks, and nickpoints
on dry valley tributaries of the Manifold can be related to phases
of rejuvenation witnessed by the landforms of the major valley.
For these reasons Warwick (1964) ascribes the dry valleys to
inheritance from a former cover of impervious shales. After this
cover was stripped, the rivers incised into the limestone beneath
until solution opened up planes of weakness in it to permit drainage
to go underground. The hanging condition is a product of main
valleys continuing to cut down after tributaries ceased to have a
surface flow. Hanging valleys are frequent landforms in karst;
Sweeting (1950) describes them from Craven. On Cooleman
Plain several dry valleys hang above the North and South Branches
of Cave Creek. Discordance of tributary junction occurs along the
dry valley as well as the perennial and intermittent flow sections
of the South Branch.
The dry valleys of the Cretaceous Chalk of eastern England
and northern France have occasioned more attention than any
others. Valley systems dominate the relief of the cuestas and low
plateaux in this weak and porous limestone, closed depressions
are scarce and small. Allogenic rivers pass through, often in con­
sequent water gaps, but otherwise only the lower parts of more
important valleys have stream channels, usually over alluvial flood
plains. Very rounded cross-sections are characteristic of the dry
valleys, though thousands of years of cultivation have tended to
flatten the floors and cause breaks of slope along the valley sides
(Sparks and Lewis 1957). They drain both dipslopes and scarps
and sometimes exhibit rectangular patterns suggestive of joint
control. Exceptionally wet seasons can cause surface flow over
their grassy floors.
It is difficult to attribute the Chalk dry valleys as a whole to
inheritance of surface drainage from overlying impervious rocks
though some have this nature near the margin of Tertiary covers.
Some of the English Chalk country was planed off by an Early
Pleistocene marine transgression across which dry valleys are as
well developed as elsewhere. Some writers have attributed them
to glacial meltwater streams, mistaking the anthropogenic flat floors
 Surface Landfonns 119

for the characteristic trough cross-section of meltwater overflow

channels. Though there are a few dry valleys of this origin in the
Chalk of England, it is impossible as a general explanation because
much of the Chalk was unaffected by glaciers and elsewhere the
dendritic patterns of dry valleys do not fit any conceivable pattern
of ice retreat.
Periglacial conditions have been called on, with permafrost
inducing surface flow of snow meltwater and summer rain to cut
the valleys (Reid 1887). Dry valleys commonly have deposits of
coombe rock— angular chalk rubble in mud matrix— which are
at least in part periglacial solifluction phenomena and the rounded
chalk forms are attributed to periglacial masswasting as a whole
by many investigators. There seems little doubt therefore but that
Pleistocene phases of frozen ground and surface flow have affected
very many Chalk dry valleys (Brown 1969). However, it seems
unrealistic to assume plane surfaces over the whole Chalk terrain
prior to the Pleistocene cold periods.
Other authorities have sought their origins in changes in the
watertable. Sparks and Lewis (1957) made a strong case for
interpreting some at least of the scarp dry valleys, which are
steeper sided and blunter headed than the dipslope valleys, by
spring sapping, followed by lowering of the watertable and down
valley migration of springheads, now countersunk in the dry
valley floors well below their heads. The Pegsdon dry valleys in
Hertfordshire have valley fill yielding molluscs indicative of a
Postglacial warm period during which springs emerged higher up
the valley. However, they inclined to think that much of the spring
action creating the valleys belonged to earlier periods than that,
perhaps of more effective precipitation than at present. Lack of
statistical correlation between joint and dry valley directions has
been used by Brown (1969) to argue against spring sapping as
the mechanism of dry valley formation in the scarps.
Chandler (1909) and Lagg (1923, 1954) explained the dip-
slope dry valleys of the Downs of southeast England by scarp
retreat which lowered the springline at the scarp foot at the
contact of the chalk with underlying clay and thus also the water-
table, causing surface drainage to go underground on the dipslope.
But this hypothesis implies that the scarp dry valleys must be of
a different later generation; for this there is no evidence. More­
over, erosion surface remnants below the North Downs scarp show
120 Karst

it has not retreated much for a long time. However, Sparks (1961)
points out that suspended erosion surfaces are evidence of rejuve­
nation in the clay vales which must have lowered watertables in
the cuesta of chalk without scarp retreat and so could have dried
out both scarp and dipslope valleys.
The long profiles of dry valleys in the Chilterns show alternat­
ing graded reaches and nickpoints due to successive rejuvenations.
C. D. Ollier and A. J. Thomasson (pers. comm.) consider that
these facts deny genesis by meltwater runoff over frozen ground
which should result in a simpler profile, but support the idea that
headward sapping consequent on a fall of base level dried out
higher reaches of the valleys. Similarly, Pinchemel (1954) relies
on valley deepening over a long period of time to dry out
tributary valleys. Springs would shift down subsidiary valleys as
each incision in major valleys lowered the watertable. He terms
this general process one of auto-desiccation. It can be measured by
his index of desiccation, which is the ratio of valley density to
stream channel density (cf. Williams (1966b) dry valley/stream
valley ratio). Chalk country near Amiens gives a high ratio of 7 4
whereas Limousin granite country had an index of L3 and Jurassic
clays, marls, limestones and sandstones of the Bas Boulonnais
one of 1-85.

DOLINES AND COCKPITS

The simpler forms of karst closed depression are now commonly

considered under the name ‘doline’ of Serbo-Croat origin; English
words such as sinkhole, swallet, and swallow hole having very
loose connotations. Dolines are generally circular or oval in plan,
with depths varying very much in relation to diameters. There is
thus a range of forms — dish and bowl shaped, conical and
cylindrical. When in bedrock they approach the form of a shaft,
there is transition to the pothole, which is best considered along
with caves. When they become elongated in plan, there is again
transition to the karst corridor (PI. 37). If a stream runs into
them, the gradation is to the blind valley.
In size they vary from a few metres in their dimensions to more
than a hundred metres in depth and to several hundred metres in
horizontal dimensions. With increasing size, there is usually com­
plexity of form, which takes them into other categories of closed
depression.
 Surface Landforms 121

Bedrock crops out on the floors and sides of some dolines,

which are often then more angular in form, but others are
surfaced largely or entirely by soil or superficial deposits. Still
others are formed in bedrock formations overlying karst rock. If
there is a flat floor, this consists of detrital materials, often
impervious, including insoluble clay residues. Swamps and inter­
mittent or permanent ponds may occur on these impervious seals.
In other dolines there are cave entrances or open shafts.
Several processes are responsible for doline formation: surface
solution, cave collapse, piping, subsidence and stream removal of
superficial covers. These often occur in combination. Nevertheless,
it is useful in first analysis to consider dolines in a five-fold
categorisation on the basis of certain dominant mechanisms
(Cramer 1941).

36 Block diagrams of collapse and solution dolines

(a) Solution dolines (Fig. 36). These are due primarily to

pronounced surface solution of the karst bedrock around some
favourable point such as a joint intersection. The solutes and some
insoluble residues are removed down solution-widened planes of
weakness, though once the latter are enlarged to shaft dimensions
there will be sliding and falling of residues and rock fragments
brought to their apertures. As soon as a focus of downward perco­
lation is established by solution, it will gather drainage to itself
and the embryonic doline will further its own development. In
fairly uniform rock the interaction of solution, mechanical slope
wash and mass movements of material with the angle of the doline
sides can result in conditions of dynamic equilibrium on uniform
slopes. A conical shape is therefore characteristic of dolines of
122 Karst

this kind. However, residues may accumulate at the bottom too

rapidly for removal down widened joint planes, so that they form
flat and perhaps swampy floors or even cause pools. The slopes
around the floor may maintain a constant and characteristic angle,
usually in the 30-40° range. But if water is deflected laterally and
attacks the side slopes as the bottom rises through the accumula­
tion of impervious clays, the bedrock walls will lose their uniform
angle (Aubert 1969). The character of a doline will thus depend
on the two ratios:
vertical solution evacuation of solutes and clasts
------------------ X _________ ________________
lateral erosion clastic fill
The origin of a solution doline will be apparent if bedrock is
exposed over much or all of its surface (PI. 20), as commonly
happens in high mountains or dry country karst or where forest
clearance has been followed by loss of soil and litter from the
doline slopes. If soil or waste mantle covers the sides and bottom,

20 Solution doline with bedrock exposed over floor and sides of closed
depression. Velebit Mountains, Yugoslavia.
 Surface Landforms 123

Dolme-
-Trench

^ Bedrock
Local limestone fragments
A A Glacial erratic
I?/ d Yellow clay
Black humic earth
c = o Bone
nöjfl« Wood 0________ ) 3 Metres
Void o 10 Feet

37 Section through a solution doline at Grandes-Chaumilles, Swiss Jura.

After Aubert 1966. Numbers explained in text.

origin by surface solution may still be witnessed by a shaft or

solution pipe in undisturbed bedrock in the bottom of the
depression.
Aubert (1966) has sectioned a solution doline at Grandes-
Chaumilles in the Jura, revealing a bedrock solution pipe down
which the doline contents descend (Fig. 37). Frost shattered rock
(1) and residual clay soil with local rock fragments (2) underlie
clay soils containing Würm glacial erratics as well (3) and stone-
free clays above (4). This fill was partly evacuated in Postglacial
times and the inner doline thus formed has been partly refilled by
rcndzinas containing talus from the doline wall, bones and wood
(5). Thus the form and contents of the doline register its history
from the pre-Würm interglacial or a Würm interstadial.
Symmetrical form results from uniform slope processes on
uniform rock but asymmetry can arise in a number of ways (Fig.
39d). If the bedrock has a substantial dip, the dipslope side of
the doline is likely to be less steep than the antidip side, bedding
planes having a stronger influence on the one and joint planes on
the other. Again if the dolines develop in a steep slope, the upslope
sides are likely to gather more water than the downslope sides and
124 Karst

to be reduced in angle by greater solution. In conditions of

prevalent snow drifting, the leeward sides of dolines will gather
more snow which will persist longer. These may thus get degraded
more than the snow-free flanks. However, the opposite may be the

/F lat \
\fLoor

\ \ / /

Formline interval 5 feet

--------Doline margin
e/ r* \ Cave
Overhang
Overhanging cliff
30 Metres
l o o Feet

38 The Punchbowl, Wee Jasper, N.S.W. Collapse doline on right; closed

depression on left may be a solution doline.
 Surface Landfonns 125

case if the snow-free sides have much more humus on them than
the snowbanl; side, solution beneath the organic cover being
greater than beneath the snow.
(b) Collapse dolines (Fig. 36). The prime cause of a doline
may be collapse of a roof of a cave formed by underground
solution. To begin with, these will be mainly vertically walled and
often angular in plan through the influence of joints (Fig. 38).
Moreover, the depth-width ratio may frequently be greater than
with solution dolines where it is not likely to exceed 1:3-5.
However, inless there is further collapse, these dolines will
change progressively to a conical or bowl shape through the wear­
ing down of tie sides and filling of the bottom (by solution, rock-
fall, frost wedging and salt wedging according to climate), soil
creep once soil forms, and other processes. In time overt signs of
collapse are lost and a superficial similarity to other dolines
prevails. Onh examination of the cave below or excavation in the
doline bottom can reveal the origin.
In a dry climate, the original form may persist a long time. In
the Nullarbor Plain, for example, all the large dolines seem to be
of collapse origin (Jennings 1967c; PI. 31). Where collapse has
been into wa.er-filled caves or there has been subsequent rise of
water level, the collapse doline may have a lake, which can be
deep, occupying its floor (Fig 40a). Such are the cenotes of
Yucatan, the ‘obruk’ lakes of the Turkish plateau (PI. 21), and
similar features in the southeast of South Australia, including
Piccaninny Blue Lake and Goulden’s Hole. Divers have
descended 55 m into the latter.
(c) Subjacent karst collapse doline (Fig. 39b). Cave collapse
occurs in karst rocks beneath overlying bedrock formations.
Initially steep-walled, deep dolines may result but weathering will
turn them into conical features which may be degraded into still
gentler forms in due course. Thomas (1954) has shown that in
South Wales there are more and larger dolines on the Carboni­
ferous Millstone Grit, a conglomeratic sandstone, than on the
outcrops of the Carboniferous Limestone that extends beneath it.
He attributes them to collapse into caverns in the subjacent lime­
stone; however, there is the difficulty that bigger caves than any
known in Britain as yet must be inferred from the size of the Grit­
stone depressons.
126 Karst

21 Collapse doline with ‘obruk’ lake near Konya, Turkey. A cenote.

No other explanation seems possible for the Big Hole near

Braidwood in New South Wales where Silurian limestone is
inferred to underlie a Devonian quartz, sandstone in which this
115 m hole with a high depth/width ratio is found (Fig. 40b).
Here a chamber large enough to account for a cavity the size of
the Big Hole is known in Wyanbene Cave nearby beneath the
Devonian-Silurian unconformity, with shafts stoping to the over-
lying sandstone (Jennings 1967b).
(d) Subsidence doline (Fig. 39a). Where superficial deposits
or thick residual soils overlie karst rocks, dolines can develop
through spasmodic subsidence and more continuous piping of
these materials into widened joints and solution pipes in the bed­
rock beneath. They vary very much in size and shape. A quick
movement of subsidence may temporarily produce a cylindrical
hole (PI. 40) which rapidly weathers into a gentler conical or
bowl-shaped depression. More or less continuous small-scale
movements may initiate and maintain forms of this second kind.
The ‘shakeholes’ of Craven are conical subsidence dolines in
the glacial moraine left on the limestone by the final Pleistocene
 Surface Landforms 127

_SjJ[^£ce_of^aLLuviaL plain
^ — r-—Strearn f

Subsidence doline Alluvial streamsink doline

Subjacent karst
collapse doline

Prevalent wind
Prevalent snow drifting wind

Snowbank Snow Soil and humus

39 (a) Subsidence doline.

(b) Subjacent karst collapse doline.
(c) Alluvial streamsink doline.
(d) Some causes of doline asymmetry.

glaciation (Sweeting 1950). In the Mole Creek area of Tasmania,

many dolines occur in Pleistocene gravel fans, of glacifluvial and
periglacial-fluvial nature, resting on Ordovician limestone
(Jennings 1967a). Lowry (1967a) has shown that the conical
doline feeding into open solution pipes above Easter Cave,
Augusta, southwestern Australia, is a subsidence doline in thick
residual soils above aeolian calcarenite (Fig. 7b). Dolines can also
be formed by solution at the top of subjacent karst rocks and by
gradual subsidence or recurrent spasms of subsidence of overlying
weak bedrock through progressive removal of their support.
Cramer (1941) and Hundt (1950) cite many areas in central
Germany where this happens above evaporites.
(e) Alluvial streamsink dolines (Fig. 39c). Dolines form in
alluvium where streams sink into underlying karst rock. The
128 Karst

30 Metres
100 Feet

Contour interval 10 f e e t

o o o o o o o o o

Rubble I
pile r—
60 Metre1
200 Feet

40 (a) Hells Hole, Mt Gambier, South Australia, a cenote. After Cave

Expl. Grp S. Australia survey.
(b) Braidwood Big Hole, N.S. W., a subjacent karst collapse doline.

processes which create subsidence dolines operate here but addi­

tionally the stream provides a good channel for mechanical
removal of the insoluble alluvium. Stream-cut trenches lead into
the side of otherwise conical dolines produced in this way. This
kind of doline is frequently blocked by detritus and much of the
time there may be no karst bedrock visible at all.
Because dolines often occur in large numbers close together
(Fig. 41, PI. 22), geomorphologists were early led to rudimentary
quantitative analysis in their study (e.g. Lozinski 1907). Drawing
on these earlier efforts and adding to them, Cramer (1941) was
 Surface Landforms 129

1 Kilometre
1 Mile
Contour interval 20 f e e t
* Lake

41 Duline karst near Mam moth Cave, Kentucky. Drawn from U.S. Geol.
Surv. map.
22 Doline field at Craig more, South Canterbury, New Zealand. Dolines
along joints and along dry valleys, partly due to solution in Oligocene
limestone and partly due to subsidence in loess cover. Photo by
E. Thorn ley by courtesy of New Zealand Geological Survey.

the first to venture more consciously in this direction in a carto­

graphic analysis of doline fields in twenty areas. His measures
were: mean size in the doline population, the doline density
(numbers per km-), and total doline area per km-. Williams
(1969) proposes an index of pitting which is the reciprocal of this
last measure.
Cramer determined mean sizes ranging widely between 17 m-
and 159,200 m-, densities of 0 57/km- to 2460/km-, and total
doline areas of 006 m-/km- to 299,000 m-/km2. The sample
areas, selected on the basis of available detailed maps, varied
lithologically (limestone, dolomite, gypsum), climatically (alpine,
cool temperate continental, humid subtropical, tropical semiarid,
and tropical humid), and in other ways. The dolines themselves
varied genetically, though solution and subsidence dolines were
regarded as the dominant types except in the gypsum karst where
 Surface Landfortns 131

subjacent karst collapse as well as subsidence dolines were

prevalent. With data complex enough to invite factorial analysis
today, Cramer did not argue very much from his results. He
recognised that there was climatic as well as geological influence
on the mean size and density of both subsidence and solution
dolines without being able to disentangle them. The doline fields
consisted of (a ) few, small dolines; (b) many, small dolines; and
(c) few, large dolines. However, these types do not represent an
evolutionary sequence, for which relief energy would be a better
indicator. Size and density are instead measures of corrosion
intensity and total doline area per km-, the product of these two
measures, is a single index for this. His results show this is highest
in humid and semiarid tropical and subtropical instances and next
greatest in alpine karsts. The lowest values came from cool and
warm temperate continental interior examples. Many more such
analyses would be necessary to substantiate these relationships.
Relief energy can be measured by depth/diam eter ratio and
this was early used by Cvijic (1893) descriptively. Considering
small depressions of part of the Mendip plateau, Coleman and
Balchin (1959) argued that if they were of solutional origin, there
should be a tendency to dynamic equilibrium in their slopes and
therefore a depth/diam eter graph should cluster them along a
straight line. Collapse dolines on the other hand would be initially
variable in this ratio and weathering back of their sides would
increase this variability; a depth/diam eter graph would have a
wide scatter as a result. A plot of 140 measured Mendip depres­
sions gave such a scatter. It was argued that this pointed to a
collapse origin but the initial data included many old mining holes.
Subsequently Ford (1964) examined a larger sample of 566
definitely natural dolines. Eighty per cent of them lay in dry
valleys where their longitudinal profiles had less than a certain
critical angle. Thus the dolines formed by surface solution where
most surface water was concentrated but where it did not run off
too fast to percolate underground. Mapping of the caves in the
area showed that they do not underlie the dry valleys and
collapses in the caves do not correspond with surface depressions.
Measurement of a field of ninety-four dolines on the Craigmore
plateau, South Canterbury, New Zealand (PI. 22; Fig. 42 ), showed
that there is a strong modality in their depth (m ean 5 9 m, standard
deviation 2 3 m ). Also there is a strong correlation between depth
132 Karst

Doline
measurements
Ft M

100 200 0 25 100

1/2 ( Length + Width) Depth Width
( Feet) ( Feet) ( Feet)

42 Dolines at Craigmore, South Canterbury, New Zealand

(a) Dimension measurements.
(b) Augered profile across one doline.
(c) Frequency distribution of mean diameter.
(d) Frequency distribution of depth.
(e) Dispersion diagram of depth against width.

and mean diameter ( = 0 84, significant at 01 per cent confidence

level). These shape characteristics argue against a collapse origin
and point to parallel slope retreat in them, which is compatible
with solutional or subsidence origin. Boring across one of these
dolines indicated a composite origin, with a younger subsidence
doline in loess covering a solution doline beneath.
These New Zealand dolines tend strongly to circularity (mean
length/width ratio or elongation ratio (Williams 1969; La Valle
1967) L32, standard deviation 0 32) and there is little tendency
to a preferred orientation in the 25 per cent sufficiently elongated
for the direction of elongation to be measurable. In south central
Kentucky over several limestones, La Valle (1967) found there
is significant elongation. He calculated mean elongation ratio for
 Surface Landforms 133

dolines in sample areas and also the percentage of dolines elon­

gated along joints or faults. Multiple regression analysis was used
to relate these doline characteristics to other karst parameters. The
two traits were found to be strongly associated with one another;
the more elongated they are the more frequently aligned along
structural lineaments. Both elongation ratio and structural align­
ment increase with purity of the limestone. They are also directly
related to the gradient of the underground drainage as calculated
greatest height above lowest rising
by the karst relief ra tio ------------------------------------------------- .
distance from underground watershed
This is thought to be due to more rapid and more efficient removal
underground of solutes and residues from the limestone where
gradients are greater. Also the two characteristics are more pro­
nounced the nearer the sample area is to a rising; underground
drainage will be better developed near the outflow points and so
lead to more pronounced adjustment to structure in the closed
depressions above. In addition percentage of structurally aligned
depressions, and probably also mean elongation ratio, is correlated
with the percentage of each sample area in closed depressions
(cf. total doline area/km 2); the greater the development of dolines
the closer their adjustment to structure in length and direction.
In tropical humid karst there are many simple closed depres­
sions which differ from those already described in important ways
(Fig. 43). They are star-shaped, with their sides lobed convexly
inwards and with gullies between which carry streams after heavy
rain, as Lehmann (1936) described from south-central Java.
Moreover, they do not perforate a fairly simple surface but are set
amongst steep residual hills. The Jamaican name ‘cockpit’ has
been associated with this kind of closed depression. The slopes of
these residual hills drain directly into the cockpits and the whole
area belongs to the closed depressions. Aub (in press b) has
shown that 60 per cent of the cockpits of a Jamaican karst area
have bedrock shafts at or near their lowest points. This is probably
true also of most of the remainder so that here there is a field of
closed depressions due to surface solution.
For their morphometry, Williams (1969) draws their boun­
daries over the tops of the enclosing hills and through the cols
between these so that all the slopes draining into the cockpits are
included. From air photographs he analysed samples of various
kinds of karst in eastern New Guinea using such limits. Such
134 Karst

43 Characteristics of tropical cockpits, ordering, dimensional analysis and

some resulting relationships from karst on M t Kaijende, Central High­
lands, Australian New Guinea. After P. W. Williams.
 Surface Landforms 135

depressions were ordered on the basis of their gully systems as if

they were normal drainage basins and various dimensions were
measured, including distances from the lowest point in each
depression to the periphery along the major axis and along one at
right angles. Analysis of these dimensions revealed system in a
kind of relief previously described as chaotic. For example, in
karst on Mt Kaijende in the Central Highlands, the frequency of
depressions was greater in intermediate orders than in the simplest
and the most complex orders. Their mean area, however, increased
steadily with order and the number of residual hills on their boun-
length
daries also. There was not much change in elongation -------------—
max. width
with order nor of distance from nearest neighbour in the same
order. On the other hand the product of symmetry, a measure of
departure of the lowest point from a central position, increased
substantially with order. Analyses of these kinds impose con­
straints on hypothesising about their origin and evolution.
Matschinski (1962) has devised a tensor analysis to determine
any alignments in a pattern of dolines but it cannot be applied to
close-set fields of dolines for which more complex tests must be
devised.

UVALAS

This Serbo-Croat name has also passed into international usage to

refer to complex closed depressions with more than one hollow in
their make-up. Size is not a criterion but perforce they are larger
than small dolines because they are formed by the merging of the
simpler type of closed depression. When the rocks are dipping
substantially, uvalas are generally elongated along the strike, with
a chain of dolines in this trend; Cvijic (1960) cites Ceteniste
uvala in Triassic limestone in southwest Serbia. Faultlines can also
occasion uvalas; the depression in which Haggas Hole is situated
in the upper Waitomo valley, New Zealand, is of this nature. In
horizontal beds, uvalas are likely to be more lobate in plan than
those structurally aligned ones.
Cvijic (1960) attributes uvalas to surface solution, but little
work has been done on their formation and other modes of origin
cannot be excluded. Northeast of Mole Creek, Tasmania, there is
a uvala 250-350 m in diameter and 25 m in depth, made up of
136 Karst

14 hollows of varying size (Jennings 1967c). It forms an inlier

of Ordovician limestone in a Tertiary basalt cap of the divide
between the Mersey River and the catchments of two tributaries,
Mole Creek and Lobster Rivulet. There has been relief inversion.
A small stream drains into the deepest hollow and gullies its flank.
Here it is possible that cavernous development beneath the basalts
played a part in forming the depression.
Complex depressions of uvala type occur over the course of
underground rivers as in the case of River Mangapu in the King
Country, New Zealand, in the neighbourhood of The Lost World,
a large collapse entrance into the river cave below (PI. 23).

23 Uvala in Oligocene limestone at The Lost World, King Country, New

Zealand

P O L J ES

In the Yugoslavian karst there are many large closed depressions

with flat floors across which streams flow (Fig. 44). They are
called poljes and this term is now used generically for such
features, though there are many local names for them, e.g. ‘plans’
in Provence, ‘wangs’ in Malaya, and ‘hojos’ in Cuba (Lehmann
and others 1956). They are usually elongated along the strike and
 Surface Landforms 137

tectonic axes (PI. 25), but can also be compact or of irregular

plan. In the Dinaric karst the small Blato polje is practically as
wide as it is long but the Slamoc polje is 26 times as long as it is
wide. The largest is the Livanjsko in Croatia which is about 40 km
long and 6-8 km wide. Gams (1969) requires the flat floor to be

44 Kupres polje, Bosnia. After Roglic 1939.

138 Karst

24 Planina Polje, Slovenia, under winter floods which recur for several
weeks each year. Photo by I. Gams.

25 Small polje in Napier Range, West Kimberley, Australia. Alluviated

area beneath woodland.
 Surface Landforms 139

several square kilometres in area before a closed depression

qualifies as a polje.
There is generally a sharp break of slope between the floor and
the sides, which are normally fairly steep. Where these sides are
in impervious rocks, streams flow down on to the floors; elsewhere
they rise in springs at their margins. Then they flow over alluvial
fans or plains to low points where they sink. These streamsinks
are called ponors, which vary from cave entrances in the lime­
stone walls to alluvial streamsink dolines. There may be several
ponors along a stream or else it may branch to several ponors.
There may be flood ponors and old abandoned ponors at levels
higher than the normal place of engulfment. Small poljes may
have only one alluvial floor and one stream; others have several,
with low, doline-riddled bedrock areas between.
In many poljes, the ponors cannot carry away the runoff fast
enough after rainy weather, even when higher-lying ones have
come into action. Then shallow lakes form over part or even the
whole of the polje floor. Thus Popovo Polje, inland from Dubrov­
nik in Dalmatia, was inundated over three-quarters of its floor in
each of seven years between 1891 and 1900, with depths of more
than 30 m at its lower end. In some poljes certain ponors change
function for a period in the wet season and spew out water instead
of imbibing it. Indeed this is the commonest context for estavelles.
During the dry season temporary lakes recede as runoff fails
and evaporation returns water to the atmosphere. There are, how­
ever, permanent lakes over parts of some polje floors, e.g. in
Cerknica polje, Slovenia. Near coasts, lowlying poljes are some­
times entirely occupied by permanent lakes, e.g. Jesero Lake near
Zara in Dalmatia.
In the Dinaric karst and elsewhere there are depressions fully
comparable with poljes except that they have external surface
drainage along a narrow defile which seems almost adventitious
to the major feature. Cvijie (I960) terms them open poljes; semi-
closed may be a better term.
Residual hills of limestone or hums in some cases protrude
through the alluvial plains or rise from slightly higher bedrock
floors of limestone, particularly towards the sides of the poljes.
In Yugoslavia they vary from 15 to 90 m in height and have a
sharp break of slope at their foot. They are often pyramidal in
shape with rather uniform slopes but those in the alluvium tend to
140 Karst

more convex profiles with a basal steepening of slope (Klaer

1957).
Early debate on the formation of the Yugoslavian poljes has
been confused by the facts that many are closely related to the
mid-Tertiary folding of the Dinaric mountains and that late
Tertiary lacustrine beds occur in some of them. This led to the
idea that they are of tectonic origin, occupying fault-angle
depressions, fault troughs and synclines (Fig. 45a). There can be
little doubt that some poljes constitute primary tectonic relief.

Fault trough Synclinal

polje polje

Anticlinal
polje

45 Structural relations of poljes

(a) Primary tectonic poljes in syncline and fault trough.
(b) Enlargement of tectonic poljes and development of secondary poljes
by corrosion.

Thus in New Guinea, the depression in which Lake Tibera lies is

a tectonic depression due to high angle reverse faulting in a Plio-
Pleistocene orogeny, subterraneously drained through limestone.
However, the Dinaric karst is older and there are thought to be
high-lying erosion surfaces which argue against survival of tec­
tonic relief. The Tertiary lacustrine beds have been tectonically
deformed themselves and are older than the time of polje forma­
tion. Some poljes are formed in tectonic highs, namely anticlines
and horsts. Moreover poljes in synclines and fault troughs often
have irregular margins which transgress the faultlines and synclinal
limbs broadly defining the polje floor. Additionally these floors,
whether exposed or alluviated, are found sometimes to truncate
structures in the limestone (Fig. 45b). Therefore in the classical
Dinaric karst the close relationship of poljes to tectonic lines
 Surface Land forms 141

seems to be a reflection of structural guidance of erosion. Even

the simplest tectonic poljes usually involve some erosion, e.g.
Mende polje, Portugal (Birot 1949; Fig. 46c).
Erosional poljes in Yugoslavia have been interpreted by Roglic
(1939, 1940) as due to lateral solution undercutting, a mechanism
applied to various areas by others (e.g. in Turkey by Louis 1956;.

Corrosional plain
remnant Kupres Polje
Cretaceous "SE
limestone r Alluvium
Triassic
dolomite
F Oligo-miocene Werfen Shales
Cretaceous freshwater marl
limestone Triassic limestone

Corrosional plain
Duvno Polje
Cretaceous Alluvium
limestone
Cretaceous
F Tertio, limestone

Mende Polje,
Portugal

Hum Flood ponor

Ponor

Limestone

Impermeable beds Buried ponors

46 (a) Section across Kupres polje, Bosnia. After Roglic 1939.

(b) Sections across Duvno polje, Bosnia. After Roglic 1940.
(c) Block diagram of Mende polje, Portugal. After Birot 1949.
(d) Theoretical section to show polje formation by lateral corrosion
from alluvial floor. Based on Roglic and Louis.
142 Karst

in Italy by Lehmann 1959). Parts of the walls and floors of the

poljes Roglic studied are developed on impervious Mesozoic and
Tertiary beds and this is where they initiated (Fig. 46a, b).
Surface drainage on them carried detritus to streamsinks in the
surrounding limestone where alluvium accumulated (Fig. 46d).
Because of its lower solubility, dolomite has acted similarly so
that some poljes have developed along the contact of limestone
and dolomite (Gams 1969). Periodically the surface of the
alluvial flat formed in this way was flooded and aggressive waters
attacked the limestone margins. Rotting organic matter washed to
these contacts provided biogenic C 0 2. By this biochemical solu­
tion at the edge of the plain, a flat floor was extended into the
limestone and an alluvial seal simultaneously spread over it. This
seal normally protected the karst rock beneath, permitting this
extension by lateral solution. Poljes created in this way can
obviously be regarded as developments of blind valleys.
At certain junctures, ponors may open up substantially, enabling
polje streams to incise the alluvium; subsequent partial blocking
of the ponors would result in fresh lateral planation at a lower
level, producing a countersunk floor. Rock terraces along polje
floor margins occur which could correspond with such a history.
Polje floors represent interior planation surfaces. H. Lehmann
and others have stressed control of their altitudes by the level of
the outflows (Vorfluters) of the underground drains; this can result
in common levels for many poljes where the outlets are controlled
by sea level or that of a coastal plain. The other school of thought
led by Roglic emphasises control by ponor level within the poljes
whereby each polje, even parts of poljes, are independent in their
planation level. The cave systems between ponor and outflow vary
so much that there is good reason to think that both circumstances
occur in nature.
From Cuban tropical humid karst Lehmann and others (1956)
described karst margin poljes where the flat floor is largely
surrounded by impervious rocks with a limestone wall on the out­
flow flank only. Panos and Stelcl (1968) have denied that much
of these floors is developed on former limestone floors subse­
quently stripped, as Lehmann inferred. There is little doubt but
that this has happened in the case of a karst margin polje behind
The Tunnel in the Napier Range, West Kimberley, where the
backreef facies of a Devonian reef structure has largely been lost
 Surface Landforms 143

in this way. Karst margin poljes have been recognised in many

karsts. However, the term polje may not strictly apply to closed
depressions without any limestone floor in their evolution. Some
of the ‘inland valleys’ surrounded by limestone in Jamaica but
entirely floored by impervious bedrock are of this type.
Though the Yugoslavian terms— doline, uvala, and polje—can
be used to classify very many closed depressions in karst, there is
greater variety than this and no attempt should be made to force
all occurrences into this framework.

KARST MARGIN PLAINS

Attention has concentrated on the flat surfaces which have

developed enclosed within karst because they are the most
distinctive. However, many of the factors which operate there can
also be effective in the condition of external surface drainage.
Morawetz (1967) has redrawn attention to the lower Neretva
valley in Yugoslavia, a plain only 4-6 m above sea level as much
as 20-30 km inland. It is a bedrock plain with an alluvial veneer,
hums project through it and the margins are embayed by steep-
heads. This active karst margin plain is being extended today by
spring sapping and also by general flooding in winter and spring
from the Neretva River and the springs. Corrosion plains of this
type tend to greater perfection than most other kinds of degrada-
tional plain. Jennings and Sweeting (1963) have argued that
tropical semiarid pediments likewise tend to a higher degree of
planation as a result of special processes obtaining when they are
developed on karst rocks. But it is in the humid tropical karsts
that the best examples of karst margin plains are to be found, for
example in the Celebes (Sunartadirdja and Lehmann 1960) and
in Tabasco, Mexico (Gerstenhauer 1960) because the climate and
vegetation are there very favourable to extension of plains by
lateral solution (see Chapter IX ). The extensive opencast work­
ings for tin in the plains of the Kinta valley in Malaya reveal
vividly how intricately corroded subsurface relief with an ampli­
tude up to 10 m is nevertheless horizontally truncated most
sharply.
 VII

KARST CAVES

Penetrable natural cavities are formed in diverse ways— by lava

flow, wave attack, weathering, landslides, and movement and melt­
ing of glaciers, but caves formed by karst processes are the most
numerous, largest and most complex. Different karst rocks vary
in their propensity to form caves, chiefly according to their
chemical purity and mechanical strength. Insoluble residues tend
to block incipient cave development and inadequate shear strength
results in collapse of developing cavities. The last factor explains
why caves are poorly developed in evaporites (Krejci-Graf 1935;
Wagner 1935; Würm 1953). Rock strength increases with com­
paction, itself the result of compression, cementation, and
recrystallisation, but decreases with frequency of planes of weak­
ness. Some planes of weakness are necessary for substantial per­
meability, which is another prior condition for cave formation.
Yet no karst rocks seem too massive for caves, though they may
be few and large in very massive rocks. Coarse, intergranular
porosity may maintain dispersed percolation and be inimical to
speleogenesis, the formation of caves; this may explain the com­
parative poverty in caves of the Chalk of Britain and France.
Water is also necessary for caves to form. The Nullarbor Plain is
possibly the largest continuous limestone area in the world, a free
karst of approximately 200,000 km-, yet cave development is
modest because it is entirely subject to semiarid and arid climates
(Jennings 1967a). Where rock temperatures are below freezing
point, water is present as ice only and caves cannot develop, though
they may survive (Corbel 1954).
Though predominantly the effect of meteoric waters, solution
of karst rocks is not quite exclusively so. Thus hydrothermal water
produces caves and cave decorations and these must be considered
144
 Karst Caves 145

true karst features. One of the best known occurrences of hydro-

thermal karst is at Hranice on the River Becva in Moravia
(Kunsky 1958).

GENERAL CHARACTERISTICS

Karst caves are extremely varied in all respects. They range in

length from a few metres to well over 100 km in the cases of
Flint Ridge Cave, Kentucky, and Hölloch in Switzerland and in
depth to 1311 m in Gouffre St Pierre Martin in the French
Pyrenees. Calculations of area and volume are few so the range
cannot be specified, but very large caves in these two ways are the
Hölloch with 7 km2 of area and Gouffre Berger in the French Alps
with 2-3 km3 of volume (Geze 1965). In complexity they vary
from single rooms, short passages, or open shafts (potholes) to most
intricate systems of passages linking rooms, shafts, chambers, and
halls of all sorts of shapes and sizes. Single chambers reach the
size of Grotta Gigante near Trieste, which is 200 m long, 130 m
wide, and 136 m high, and of the Big Room in Carlsbad Cavern,
New Mexico, with dimensions of 400 m, 230 m, and 100 m.
Associated with penetrable caves, there are usually innumerable
enlarged joints and small tubes, embryonic caves, which together
may amount to a much greater volume than the caves proper. The
density of caves can vary from a few m /km 2 in several small caves
to hundreds of metres of cave per square kilometre. Some residual
hills of limestone seem to be mere shells riddled with large and
small cavities, e.g. Mt Etna near Rockhampton, Queensland,
though more sober examination proves that much more rock
remains than voids.
Some caves are completely dry, others are filled with water.
The Masocha pothole in Moravia has a lake more than 100 m
deep and there are more than 10 km of lakes (more like canals
than true lakes) in Padirac in the Causse de Gramat, central
France. Some caves are completely surfaced by bedrock but most
have deposits of various kinds in them, which can accumulate to
total fill when the cave becomes fossil.
Caves occur in all kinds of topographic situation from beneath
coastal plains and valley bottoms to the tops of mountains. High,
steep ridges may be poor in caves because gradients have favoured
runoff at the expense of infiltration (Grund 1910b). High plateaux
146 Karst

tend to be rich in deep potholes and caves with a great deal of

vertical development whereas low plateaux will chiefly have
horizontally developed caves. Cave exploration beneath lowlying
plains and valley bottoms usually involves diving, so not much is
known about them in this situation.
Cave entrances may be vertical shafts and fissures or lateral
openings in slopes and cliffs. When small, the aperture may have
smooth forms and be entirely erosional, but when large, rockfall
has usually contributed to their enlargement. Even then they need
not be angular because subaerial weathering can round off large
arches produced by collapse. Tiny entrances can lead into large
systems and many caves have only artificial entrances, usually
made by removal of rockfall or digging out of finer detritus. Some
caves have only become known when quarrying or mining in bed­
rock intersected them (Warwick 1968). Other caves have water-
filled entrances.
Very many caves have rivers running through them and are
called active river caves. These caves can usefully be considered
on the following basis (Grund 1910b):
(a) Inflow cave where a river is followed downstream from
its point of engulfment to a sump which is the limit of
exploration.
(b) Outflow cave where a river is followed upstream to
impassable obstacles or to the beginnings of concen­
tration of seepage water.
(c) Through-cave where a river is passable from its sinking
point to its rising, possibly through water-filled sectors.
(d) Between-cave where a river passage is entered from
above or laterally but cannot be followed to the surface
either upstream or downstream.
The term ‘dead cave’ is usually applied to caves which are not
being enlarged by water action today; many of these have forms
and deposits which reveal they were active river caves in the past.
The simple classification above can be applied to these also.
Drainage through karst rocks tends to develop lower routes to the
risings. This happens both through self-capture and through
piracy of neighbouring cave rivers. River caves can become com­
pletely abandoned in these ways. Often cave systems consist of
abandoned levels at successively lower altitudes, with the lowest
one remaining active (Fig. 47).
 Karst Caves 147

In caves, there is an almost infinite variety of bedrock forms

both large and small, which result from the interaction of active
processes with the passive factors of rock types and arrangement
(Bögli 1956; Renault 1958). These are called speleogens and
some of them will be mentioned where appropriate in the follow­
ing discussion.

Far Chamber Punchbowl Cave Pitch Chamber

Metres h \

100 Metres Signature Cave

300 Feet

47 Long profile of Punchbowl-Signature Caves, Wee Jasper, N.S.W., with

four levels of development and subhorizontal solution roofs [hatched).
Bedding is vertical.

CAVE FORMATION

It is impossible in this small book chiefly concerned with surface

geomorphology to do justice to the great controversies of the past
about cave formation or even to the manifold views held today by
speleologists about the caves with which they are individually
familiar. Only the main threads can be followed.

Tectonic caves
Tectonic origins for karst caves lost their early appeal during
the gradual overthrow of catastrophism by uniformitarianism in
the explanation of landforms. The great numbers of caves in karst
rocks invalidate such origins as a general cause of these caves
because tectonism would affect all rocks of comparable strength
to much the same degree. Associations of caves and tectonic
features such as the slickensided fault-plane roof of Terrace
Chamber in Marakoopa Cave, Mole Creek, Tasmania, generally
imply no more than structural guidance of the processes forming
caves. Nevertheless some caves may be due directly to earth
movements as Geze (1953) maintains for some potholes of fissure
type. The 90 m deep Igue des Landes in Tarn, France, is due to
the opening up of joints in massive limestone in an anticline,
148 Karst

whereas the Avens du Pic St-Loup near Montpellier are in

vertically dipping beds, which have gaped apart through spread
under gravity (Fig. 48a). It is in Cainozoic orogenic belts that
there is the greatest likelihood of some caves being due to earth
movement; thus some of the large caverns in the Dachstein plateau
in Austria have been attributed by Arnberger to differential move­
ments of rock masses under thrust faulting (cf. Groom and
Coieman 1958; Pig. 48b). It is postulated that these were later
joined to one another and to the surface by passages of erosional
origin.

'k
100m

48 (a) Averts du Pic St-Loup, Montpellier, France, potholes due to gravity

spread of vertical beds. After Geze 1953.
(b) A rnberger’s concept of tectonic cave formation by differential move­
ment of limestone bodies along bedding planes.
(c) Shelves in walls of caves, Clare, Ireland, through differential solu­
tion of limestone beds. After Ollier and Trat man 1969.

Renault (1967) has argued that residual tectonic stresses

released by both surface and underground erosion play a signi­
ficant role in the localisation and growth of caves.
 Karst Caves 149

Corrosion versus corrosion

By the late nineteenth century the central controversy in speleo-

genesis had shifted to the relative roles of corrasion and corrosion
in the erosional work of underground rivers.’ Cave explorers have
understandably been impressed by the noise of waterfalls and the
grinding of boulders together in underground confines where even
small rapids and whirlpools can present real obstacles to move­
ment. There resulted a tendency to attribute more or less all the
work in forming caves to the enlargement and fusion of fluvial
potholes or rock mills. It was necessary for the case to be argued
for the less obvious quiet chemical action of underground water.
In extensive karsts, underground rivers often lack boulders,
pebbles, and sand for corrasional attack, precisely because the
rock fragments are dissolved, though with sufficient velocity purely
hydraulic action remains in their power. But the speed of move­
ment required with or without rock tools presupposes a fairly large
passage, which must first be produced by solution. Where on cave
surfaces there is differential solution of different limestone beds
or of chert inclusions, the dominance over mechanical attack is
also proven (Ollier and Tratman 1969; Fig. 48c).
Indeed appreciation of such facts and concentration on the
distinctive karst process of solution have led in recent decades to
undervaluation of mechanical attack. But small impounded karsts
usually have their underground rivers fed with detritus from
surrounding impervious rocks. For example, on Cooleman Plain,
N.S.W., nearly all caves have trains of igneous pebbles through
them from surrounding ranges, even though they are practically
horizontal systems. The more extensive limestone areas of the
King Country, New Zealand, are interspersed with outcrops of
other sediments and there have been extensive ignimbritc and
volcanic ash covers. Signs of powerful mechanical action such as
plunge pools and rock mills are very evident in the caves of this
high rainfall area. High mountain karsts reaching to frost wedging
levels have good cave supplies of rock fragments fed down pot­
holes and streamsinks into their caves.
Martel (1921) rightly stressed that the two aspects of erosion

1 It should be noted that in much French and German literature

corrasion and erosion are equated.
150 Karst

26 Cave river waterfall and plunge pool, The Chute, Coolagh River Cave,
Clare, Ireland. Water about 70 cm below normal level. Photo by
D. M. M. Thomson.
 Karst Caves 151

generally accompany one another but that their relative roles vary
with lithology and hydrodynamic circumstances.

Vadose zone action

The conditions governing cave forming processes vary between
different parts of a karst hydrologic system. At one time or another
speleogenetic theory has stressed each of the hydrodynamic and
chemical contexts which occur in these systems. Stress was first of
all laid on the action of seepage water and free surface streams in
the vadose zone because this was the most readily examined. In
some ways matching grikes and potholes on the surface, there are
blind fissures and shafts rising from cave passages and chambers
towards the surface but not reaching it. Some blind shafts- and
some potholes are attributable to seepage water, which
undoubtedly seams the walls of many of them with vertical
grooves separated by thin, sharp ribs. Cylindrical vertical shafts
of the central Kentucky caves in Pennsylvanian limestone are
significantly located beneath the edges of sandstone-capped
plateaux at the heads of receSses. According to Pohl (1955) they
are fortuitously related to the horizontal cave systems, which
themselves disregard the surface topography. He argues therefore
that these blind shafts are due to solution down joint intersections
by aggressive seepage water percolating through the sandstone
(Fig. 49a). Burke and Bird (1966) have similarly interpreted
closely comparable forms in the Carboniferous Limestone beneath
the Millstone Grit margin in South Wales; the Gunbarrel and
other blind shafts in Wyanbene Cave, N.S.W., stoped to the
unconformity between Devonian conglomerate and Silurian lime­
stone, provide another match.
Very much larger potholes, often of inverted conical form, such
as GoufTre Henne Morte (466 m) in the French Pyrenees, are
attributable to seepage water since they are fed only by small
doline catchments in mountain tops (Fig. 49b, PI. 33). Something

2 Terminology for blind shafts in English is not very satisfactory. The

British have long misused the French aven for them but in France avert
means a shaft open to the sky, a pothole, in fact. The Americans use the
Italian foiba (pl.-<?) but a similar twisting of meaning on transfer seems to
apply here also. The American ‘dome pit’ is also not satisfactory for general
use because many blind shafts are not domed at the top but taper off. Blind
shaft seems short and specific enough to act as a generic term and is
therefore used here.
152 K a rs t

Shale
Sandstone

- 1215m

Metres
Dry -----45
entrance

passage

W e t^ ^
entrance
100 150 Active river passage
Wet exit 'Cave
Creek

49 [a) V e rtica l b lin d shafts o r fo ib e in lim estone caves beneath sandstone

and behind scarp recesses, K entucky. A fte r P a id 1955.
(b) U pper p a rt o f G o u fjre Henne M o rte , Pyrenees, France, po tho le due
to seepage fro m sm all surface h o llo w . A fte r Geze 1953.
(c) Lo n g p ro file o f Barber Cave, Coolem an P lain, N.S. W. D ry passage
shows riv e r concavity o f upper vadose section and su b h orizo n ta l
so lu tio n ro o f o f lo w e r shallow p hreatic passage.

of the role of a sandstone capping not present here may be taken

on by snow drifting into the dolines in great banks which release
water slowly. Geze (1953) calls them absorption potholes. He and
others such as Maucci (1960) maintain they grow upwards though
the water is descending. Water condensing on cold rock surfaces
in the vadose zone may in some circumstances significantly contri­
bute to the formation of these potholes.
Similar potholes are caused by streams sinking vertically after
passing on to karst rock from a catchment on impervious rocks,
which may overlie it or be faulted against it. There must be deep
valleys in or around the karst for deep potholes. One of the deepest
potholes in Britain, Gaping G ill Hole in Craven, was formed in
Carboniferous Limestone by a stream from the overlying and
mainly impervious Yoredale Beds which include much shale.
Only flood waters of Fell Beck nowadays fall down the pot; its
 Karst Caves 153

normal discharge sinks a little upstream and enters it below the

surface. As the boundary of the Yoredale Beds retreats under
erosion, fresh streamsinks are created and potholes of this origin
are abandoned. Bar Pot appears to be an inactive predecessor of
Gaping Gill Hole with its entrance blocked by rockfall, recently
opened artificially.
Another kind of pothole is the consequence of the ancillary
process of collapse also characteristic of the vadose zone.
Successive breaking away or stoping of a cave roof eventually
penetrates to the surface. A block pile or debris cone accumulates
below (subject to removal by river action). The Daylight Hole
in the Dip Cave, Wee Jasper, N.S.W., has a block pile and is
situated on the crest of a spur so that it is unlikely to have been
a streamsink. However, rockfall can contribute to the growth of
seepage or streamsink potholes so there are transitional forms just
as enlargement of collapse potholes produces collapse dolines.
In the vadose zone, free-surface streams have downhill gradients
under gravity, sometimes developing an exponentially concave
longitudinal profile as do surface streams (Fig. 49c). Plunge pools
occur below waterfalls and rock mills are found in the bottoms of
canyons, the wails of which sometimes show remnants of the
cylindrical pits drilled by them (Ford 1965a). Where gradients are
less, the walls may still give evidence of vertical downcutting in
sequences of nearly horizontal stream grooves (PI. 28). Lateral
erosion accompanying the incision may result in ingrown meanders
with the roof matching the slip-off slope. Meander spurs may be
undercut and left suspended (Jennings 1964). Alternatively
meanders may be cut off and oxbow passages left abandoned. In
plan the confluences of tributaries produce branchwork patterns
(Fig. 50a), but the confined nature of caves, with no equivalent
of the flood plain on the surface, results in ready rise of level in
floods and there is a strong tendency for oxbows and higher level
passages normally left dry through capture to be maintained in
periodic action as flood overflow routes. In some caves such as
Honeycomb Cave, Mole Creek, Tasmania, the river in flood flows
through channels which divide and rejoin complexly, not in one
plane but several; it is in some degree like a three dimensional
braiding system. Structural lineaments—joints, bedding planes,
faults— often strongly influence the plans of vadose caves but
gravity stream action will frequently impose independent curving
elements.
154 K a rs t

Active river passage

Inactive former river passage

Fred Cave - New Zealand

W aterfall
High inactive
passage

Potholes Entrance pothole

300 Metres
1000 Feet

New Entronce Cameron Cave

Missouri USA

90 Metres
300 Feet

Surface contour intervol 5 metres

lid Entrance’

50 (a) Plan o f F re d Cave, A u ckla n d , N Z . Be tween-cave o f m a in ly vadose

developm ent w ith b ra n ch w o rk pattern influenced by jo in ts. A fte r
N .Z . Spel. Soc. survey.
(b) Plan o f C am eron Cave, M isso u ri, U.S.A . T w o-d im e nsion a l netw ork
o r maze developed p h re a tica lly w ith jo in t c o n tro l. A fte r M isso u ri
Spel. Soc. survey.

Microforms consequent on turbulent How in the vadose zone are

current markings or scallops (Fig. 51), asymmetric hollows with a
steep semi-circular step on the upstream side and a gentle rise
downstream ending in a point between the steps of the next down­
stream hollows (Coleman 1945; O llier and Tratman 1969). Fric­
tion with the surface causes small eddies to develop which alternate
with laminar flow; solution is greater in the eddies and so hollows
the rock. The slower the flow the larger the scallops (Glennie
 Karst Caves 155

1963). Flutes (Curl 1966a) are similarly asymmetrical solution

hollows but with long parallel crests transverse to the current.
Curl has developed a mathematical function relating their dimen­
sions to the hydrodynamics of the flow.
Scallops and related forms are caused by solution, but in the
vadose zone it is possible that corrasion may match or even sur­
pass corrosion in enlarging caves in some circumstances.

Deep phreatic solution

A free-surface stream presupposes a cave passage through which
it can move freely. In a paper stimulating much rethinking about
cave formation, Davis (1930) argued that most of these initial
passages were inherited from a previous phase of solution by slow
moving water in the phreatic zone below the watertable. Such
water infiltrated along the planes of weakness, enlarging them by

• • tv :

51 Current markings or scallops developed from small initial irregularities

by eddy currents
156 Karst

solution into three-dimensional networks of small galleries and

shafts, irregularly connected, with outgoing as well as incoming
branches and with rising as well as failing gradients. These
phreatic caves would develop best beneath erosional planes trun­
cating the limestones. Uplift and dissection then resulted in partial
drainage and modification of the caves into branchwork systems,
with corrasion and breakdown added to solution as sculpturing
processes. The very close adaptation to structure of the phreatic
caves would be partially destroyed at this time.
Bretz (1942) elaborated and gave precision to these ideas. The
elaboration included the insertion of a red clay infilling between
the phreatic and the vadose phases. Though such a history may
be true of particular caves, many cave studies over the world show
that it is not usual to have a stage of red clay fill. On the other
hand much of Bretz’s discussion of criteria to distinguish between
vadose and phreatic origins for caves has proved valid and useful.
Thus substantial development of the following features is
interpreted as due to solutional work in the saturated zone without
definite currents:
(a) Spongework; intricate cavitation like the pores of a
sponge on various scales (PI. 27).

27 Phreatic spongework in Tunnel Cave, Borenore, N.S.W.

 Karst Caves 157

(b) Bedding plane and joint anastomoses-, repeatedly branch­

ing and joining patterns of small sinuous tubes in such
planes. With bedding planes, the holes of the anasto­
moses often have flat bottoms in them, residues sealing
off the lower bed.
(c) Wall and ceiling pockets-, hemispheroidal hollows in
these surfaces of passages and chambers. Bögli (1964a)
attributes ceiling pockets or ‘hellholes’ along joints to
mixing corrosion.
(d) Joint wall and ceiling cavities-, similar but elongated
along joints.
(e) Ceiling half-tubes-, larger features than (b) in roofs of
large chambers.
(f) Continuous rock spans across cave chambers; bridges if
more or less horizontal, partitions or blades if more or
less vertical.
(g) Two- and three-dimensional networks or mazes of
passages (Fig. 50b); once created, these can be main­
tained and enlarged in other hydrodynamic states.
Rock pendants are bedrock projections hanging from ceilings,
smoothed erosionally and often in groups reaching down to a
common level. Bretz attributes them to vadose modification of the
rock left between anastomosing half-tubes, but others consider
them completely due to some kind of phreatic solution.
Most caves exhibit some of these characteristics and true
phreatic solution must initiate most cave development. However,
substantial enlargement in deep phreatic conditions beyond this
phase is thought to apply only to particular cases. Dip Cave at
Wee Jasper. N.S.W., appears to be a little modified phreatic
system with quite large chambers; breakdown since emptying is
the chief modifier. But clay residues will be susceptible to floccula­
tion in the alkaline water and will settle out as circulation is slow.
Diving into water-filled caves has revealed that clay coatings are
common on walls, floors and ceiling, which must hinder further
solution of these rock surfaces.

Watertable stream, shallow phreatic and pressure passage action

If one tube in a bedding-plane anastomosis or one element in a
spongework gets larger than its neighbours, water will pass through
158 Karst

it more rapidly and enlarge it still further. This autocatalysis

changes hydrodynamic conditions substantially and is most likely
to occur at the top of the watertable or rest level where accessions
of fresh water may renew aggressiveness.
Considerations such as these and the other restraints on deep
phreatic solution cited above led Swinnerton (1932) to maintain
that most solution went on near the top of the permanent phreatic
zone and in the intermediate zone through which the watertable
oscillates. Fast turbulent watertable streams fashion big subhori­
zontal tunnels disregarding structure. Roofs as well as walls and
floors are subjected to solution so that asymmetrical current
markings can be found on ceilings also, though smooth surfaces
can occur in these tunnels as well. Flat ceilings cutting across
structures (Fig. 47, PI. 28), have been regarded as diagnostic of
action at this level by several investigators (Halliday 1957;
Jennings 1964). Lange (1962) discusses the related question as
to how ceilings are planed off by solution at regulatory water
levels of cave lakes.

28 Dry passage in Barber Cave, Cooleman Plain, N.S.W., with shallow

phreatic fiat roof, vadosc current markings, and successive channel
grooves in walls
 Karst Caves 159

Most cave investigation in the last two decades has confirmed

Swinnerton’s views rather than Bretz’s, though the term shallow
phreatic has been preferred for this important zone of action
(Davies 1960; Deike 1960; White 1960; Thrailkill 1960, 1968).
The master caves of Craven seem closely comparable to the
shallow phreatic type (Glennie 1954b) and Glennie’s term
epiphreatic is also commonly used for this hydrodynamic zone
(Glennie 1958). Where caves are developed in gently dipping
beds which vary in solubility, abrupt terminations of passages are
regarded by White (1960) as pointing to watertablc control.
Successive horizontal levels boring through steeply dipping beds
and related to surface river terraces (Davies 1960; Ek 1961; Fig.
52) and the association of major cave development in karst areas

Grotte de
Briamont
R Ourthe

G r o tte Saint A nne Upper Level

Middle le v el / /

Lower le vel

----- Active river cave -----Dead river cave ----- River terrace

52 Relationship of cave levels in Grotte de Briamant and Grotte Saint

Anne at Tiliff, Belgium, to terraces of River Ourtlic. After Ek 1961.

with the major valleys are also interpreted in the same way. Pitty
(1968) quotes various groundwater engineers to the effect that
there is a rapid decrease downwards in limestone of cavities and of
rate of water circulation. The chemical characteristics of the top
of the phreas with several modes of renewal of aggressiveness
may be as important as the kinetic considerations already set out
here.
On the continent of Europe where there arc frequently great
depths of cave and large inputs of water and where the notion of
watertables in most karst is often unaccepted, speleologists have
chiefly contrasted pressure flow and free surface flow (Chevalier
1944). In the parts of caves permanently or temporarily filled
with water, hydrostatic heads can build up through inadequate
capacity to cope with increases in discharge; solution and corra-
160 Karst

sion can then affect all surfaces of passages in rising as well as

falling passages. Both true and inverted siphons come into violent
action.

53 Types of passage cross-section. After Renault 1958, Bögli 1956, Ollier

and Tratman 1969.
(a) Phreatic tube in massive rock.
(b) Elliptical phreatic passage in horizontal bedding plane.
(c) Phreatic passage in vertical joint plane.
(d) Phreatic passage in group of more soluble beds.
(e) Phreatic passage with ceiling half-tube due to air entrainment along
roofline.
(f) Phreatic passage aggraded to roof with development of ceiling half­
tube.
(g) Elliptical phreatic passage modified by breakdown.
(h) Vadose canyon with horizontal channel grooves incised in floor of
phreatic passage.
(j) Vadose canyon with inward meandering. Some aggradation of
primary phreatic passage before incision.

Pressure passages of this type can assume various forms

(Renault 1958) according to structural constraints, the distinctive
criterion being liability of all surfaces to erosion (Fig. 53a-d).
Cylindrical tubes develop where there is little structural influence
(PI. 29); vertical or horizontal elliptical passages where planes
of weakness retain control; rectangular passages where beds differ
 Karst Caves 161

markedly in their response to erosion. A ceiling half-tube

( Wirbelkanal: Bögli 1956; Fig. 53e) may result from air entrain­
ment accentuating solution at the highest roofline (PI. 30). These
pressure passages may be modified subsequently by the introduc­
tion of fill, vadose flow or collapse. Some apparent tubes are not
due to pressure flow at all but to negative exfoliation due to rock

29 Keyhole passage in Metro Cave, Charleston, New Zealand. Upper part

phreatic with vadose incision in floor. Photo by D. L. Homer by
courtesy of New Zealand Geological Survey.
162 Karst

30 Speleothems in Metro Cave, Charleston, New Zealand. Straw and

conical stalactites, stalagmites, and a column. Part of a shawl at top
left. Photo by D. L. Homer by courtesy of New Zealand Geological
Survey.

pressure (Fig. 6a). Signs of rock spalling will distinguish these

from fresh pressure tubes but the solution surfaces of long aban­
doned pressure tubes will come to approximate to the condition
of those due to mechanical breakdown alone.
 K a r s t Caves 163

In te rm itte n t resurgence
p o th o le o f La Lu ire

In t e r m itt e n t s u b a e ria l
course o f R.Vernaison

R esurgences o f
B o u rn illo n and
A rbois
/A / La Bourne
P ere nnial underg rou nd
course o f R. Vernaison

P otho le o f ab sorptio n
W in te r w a te rta b le and flo o d resurgence
A b s o rp tio n
S um m er w atertable p o th o le

Resurgence
p o th o le

54 (a) S ym m etrica l so lu tio n h o llo w in g o f Picznice Cave, H u n g a ry. A fte r

C ra m e r 1933.
(b) G o u ffre de la L u ire , Vercors, France, resurgence pothole. A fte r
G'eze 1953.
(c) Potholes ro u n d m argins o f po lje , G ra n d Plan de Canjuers, Provence,
France. A fte r Geze 1953.
164 Karst

Forms intermediate in nature between those of deep and shallow

phreatic or pressure flow have been recognised. Injections of
powerful currents into large water-filled spaces seem capable of
producing large symmetrical surface hollows. Cramer ( 1933)
describes this from the Piznicehöhle in Hungary (Fig. 54a), and
similar sculpture is characteristic of Cathedral Cave, Wellington,
N.S.W. Small symmetrical hollows transitional to scallops are also
known, e.g. in Dip Cave, Wee Jasper, N.S.W. (Jennings 1963).
Pressure flow is involved in two kinds of pothole (Geze 1953).
Resurgence potholes are essentially due to such flow. The Gouffre
de la Luire in the Vercors plateau, French Alps, is perhaps the
most striking example; hydrostatic pressure forces water more or
less vertically up 200 m from a deep river cave when it is in flood
(Fig. 54b). High ground above is necessary for the pressure
required and usually a fault or a syncline provides favourable
structural conditions. Other potholes are due to alternations of
rising pressure flow and descending gravity flow according to
changing water levels. These occur round the margins of poljes,
e.g. the estavelles around the Grand Plan de Canjuers in Provence
(Fig. 54c).
As a karst develops, potholes may change their function per­
manently, so each one needs careful analysis. Indeed a catholic
attitude towards hypotheses of origin for all caves is necessary
because most caves are composite in their nature. Ascertaining the
relative importance of each kind of action which has produced the
present form and pattern of each cave depends on reconstructing
its history. In many parts of the world the conclusion has been
reached that the most diverse histories may be expected from
caves quite close together, e.g. in the Sierra Nevada (Halliday
1957, 1960), the Mendips (Ford 1965b), and Wee Jasper,
N.S.W. (Jennings 1967c).

Large chambers
Large chambers in cave systems are often difficult to under­
stand because much of the evidence is destroyed in their forma­
tion (Renault 1967). Such chambers may be located excentrically
with respect to the main passages but often they lie at the junc­
tion of several passages. Sometimes favourable structural factors
may be discerned, e.g. faults, close-set joint fields, lithological
 Karst Caves 165

weaknesses. The Grotte de la Cigalere in the French Pyrenees

has chambers wherever the cave encounters schists encompassing
the limestone in which it is found.
Low, wide chambers may be due to solutional removal of
blocks between a labyrinth of small passages in a single bed. This
seems to apply to Flat Roof Lake Chamber, Jewel Cave, Augusta,
southwestern Australia, which is 160 m long and up to 45 m
wide, and developed in a shallow phreatic zone. When a roof gets
too wide to support itself, collapse results in a domed chamber.
The deep caves of the Nullarbor Plain have many such which
stope through weak chalky limestone of the Eocene Wilson Bluff
Formation to the stronger overlying Miocene Nullarbor Lime­
stone. Other chambers seem to be formed from labyrinths at
several levels above one another when collapse will contribute as
well as solution. This is the explanation given for the large halls
of Mullamullang Cave in the Nullarbor Plain by Hunt (1970).
Rockpiles result from such developments.
Most of the chambers of Punchbowl Cave, Wee Jasper, N.S.W.,
occur where the stream forming the cave stayed in approximately
the same position during the time more than one of its four levels
formed (Jennings 1964); only in the Far Chamber is there
collapse material and this has entered laterally. The ceilings of the
chambers are mainly of solutional nature, e.g. Pitch Chamber.
Renault (1958) attributes the Grotta Gigante, which is as deep
as it is wide and not a great deal longer, to collapse associated
with pothole development; there are three pothole entrances and
the rockfall floor is hollowed. It is an ‘underground doline’ and
there must have been removal from below after collapse.

MORPHOMETRIC ANALYSIS OF CAVES

Caves offer abundant scope for morphometric analysis of both

large and small forms, yet not much has so far been essayed in
this field, partly because underground conditions are often not
conducive to tedious repetitive measurement. The commonest
exercise has been to analyse the directions of approximately
straight passage segments and relate these to structural controls
such as strike joints, dip joints, and (in inclined rocks) bedding
planes (e.g. Gams 1963). Glennie (1948, 1950) found that in
Ogof Ffynnon Ddu in South Wales the frequency maxima of
166 Karst

joint-aligned passages coincided with the strike and the dip in the
eastern part of that cave but in the western part they occur 20°
away from both these directions. He surmised there may have
been earth movement since the latter part of the cave formed.
Less frequent has been analysis of the attitude of planes of
weakness which have governed cave cross-sections. However,
Maucci (1960) has done this for a sample of 200 caves in sub­
horizontal limestone near Trieste. Not unexpectedly more than
half were vertically disposed or close to it, nearly half lay
between the horizontal and 30 from it, leaving only a very smali
number between 30 and 70°. However, with such figures he rein­
forced his arguments that his ‘orthovacuums’ (i.e. blind shafts and
potholes) have quite separate genesis from his ‘paravacuums’ (i.e.
caves proper), the first being dominantly due to seepage and the
latter to streamflow. He maintains that because of gravity, seepage
solution is most effective in vertical planes of weakness, producing
the dominant frequency about the vertical. Comparison with
similar data from karst with about a 45° dip would be interesting.
These examples used simple analysis for descriptive purposes
only. Employing established statistical theory for immigration-
emigration processes based on random walks, Curl (1958)
derived a stochastic function for the relationship between lengths
of caves and the number of entrances they have, on the assump­
tions that the longer a cave the more likely it is to gain additional
entrances by natural processes such as solution and the more
entrances a cave has the more it is likely to lose some by other
natural processes such as collapse. Applying this to limestone
caves from West Virginia and Pennsylvania, he got good agree­
ment between observed and predicted length/entrance data.
For the West Virginian caves, Curl (1960) selected a mathe­
matical model for changing length of caves with single entrances
based on a regional history of a phreatic phase in which there was
increase in length of a fixed number of caves, of a vadose phase
after rejuvenation of cave fragmentation and length reduction, and
of a final decay phase of length reduction without loss of cave
numbers. This was applied to the observed length data and past
and future length frequencies predicted.
Later Curl (1966b) applied the same approach to a greater
range of limestone karsts for which appropriate data were avail­
able. The assumption that the number of cave entrances is related
 Karst Caves 167

in a simple direct manner to cave length was tested against two

other assumptions and was found the most acceptable statistically.
The preferred model predicted a considerable number of entrance­
less caves, on the average shorter than those with entrances. Any
comparisons between cave areas should rest on these as well as
the known caves. Warwick (1968) has discussed from direct
observations the importance of entranceless caves in a number of
British karsts. For each karst Curl calculated a karst constant
which secures the best fit between the observed and the predicted
length/entrance data. No relation was found between karst
constants and the accepted theories of origin for the caves in the
different areas. But there was a good correlation between the
constant and the mean length of all caves in each area, with the
exception of one karst, that of Clare, Ireland, where the length
frequency distribution was also anomalous.
Some of Curl’s work has been criticised because so many
assumptions lie behind the mathematical manipulation (Ollier
1963). There are indeed some quite unrealistic aspects to this
work. For instance, if there is a water-filled passage in a cave, it
is treated as two caves for the calculations. But the presence of
the water is really a pointer to continued development of the cave
as a whole. This objection does not apply to the same extent to
treating barriers of rockfall or fluvial fill as the boundaries between
separate caves since they may well evolve differently after such
isolation. However, though these efforts may have been too broad
and over-ambitious in a pioneer field, the basic method of applying
stochastic models to assemblages of cave data, testing them and
using them to direct further field inquiry is a valuable supplement
to classic methods of karst study.
Lesser aspects of cave morphology may be more susceptible
of mathematical analysis. No doubt stimulated by recent work on
the meanders of surface streams, Ongley (1968) has analysed the
directions of passage segments in Serpentine Cave, Jenolan,
N.S.W. and shown that they are random. Preferred orientations
which joint influence must have induced in the phreatic origins of
this cave have thus been destroyed by subsequent solutional action.
The tendency to oscillation is analysed by logging whether each
segment of passage angles to the right or to the left of the previous
segment. A series of right-bending segments followed by a series
 168 Karst

of left-bending segments constitutes an oscillation. Using 4-feet

and 2-feet segments, the average wavelength was found to be
4-6 m, ranging from 18 to 7 6 m, and the wavelength/width ratio
averaged 5-5, ranging between 2-9 and 8-5; the sinuosity was 14.
He concludes there are no true meanders but that there is a definite
meandering tendency.

UNDERGROUND SOLUTION

Whereas in the past chemical analysis of cave waters was chiefly

used to help in water tracing (e.g. Oertli 1953), there has been
in recent years increasing use of it to determine the course of solu­
tion underground. Water tracing has shown there is a tremendous
range in speed of underground water movement. Geze (1965)
gives an average of 30 m /hour from French measurements. But
the range is from 1000 m /h in steeply falling caves during storms
to 4-5 m /h through low gradient caves with large storage in lakes
and water-filled sections. Very variable speeds can be registered
from the same system— 6 to 500 m /h in the Sourciettes cave of
the first type and 5 to 76 m /h in Padirac of the second type.
So much more repetitive measurement than has been achieved so
far will be necessary for a general understanding of cave solution.
However, several studies have shown that a great deal of
solution by seepage water takes place close to the surface. Smith
and Mead (1962) demonstrate this for a Mendip cave and Gams
(1962) maintains most solution in Slovenia takes place in the
top 10 m of limestone. Considering all available data from Clare,
Ireland, Williams (1968) comes to the conclusion that as much
as 80 per cent of solution by percolating water may occur within
8 m of the surface there with high concentrations of 200-300
mg/1. Pitty (1968) takes the same view for the Peak District
with concentrations of the same order. In both of these British
Isles karsts very shallow caves sometimes have much active
deposition of dripstone confirming the measurements.
Whereas Smith and Mead (1962) found increasing amounts of
limestone in solution with increasing depth of seepage water in
G.B. Cave in the Mendips, Pitty (1966) found no correlation
between depth and hardness in drips in Peak Cavern in the Peak
District. A partial correlation analysis of the time series of obser­
vations from that cave showed greatest correlation between the
 K a r s t C aves 16 9

amount in solution and surface temperatures 3-4 months before.

Soil temperatures largely govern soil biochemical processes and
the influence of the root zone, the ‘rhizosphere effect’, on C 0 2
availability resulted in highest concentrations in the warmer
summertime, together with a ‘spring burst’ of microbial activity
and of solution. Although percolation waters in Postojna Cave in
Slovenia had a different annual rhythm of carbonate concentration
in the different climatic regime there, Gams (1966) stressed the
important effect of surface vegetation in the matter. Seepage water
was less rich in dissolved limestone under grassland than under
forest but its concentration fluctuated much less. Two-thirds of
the solution is attributed to biotic factors.

Devonian
Sandstone'
25-100,

'0-180

V/50-250
16-200
l- 3 k m

3Q00 Feet 1500 Feet

90-18'
10b0 M etres 5(3o Metres
^ 30-50 Stream sample, CaCo3 in m g / l
Rising Streamsink «^120 Drip sampLe, CaCo3 in m g / l

55 D ia g ram s o f solution in cave systems in

(a) C lare, Irela n d .
(b) M e n d ip , England. Both a fte r S m ith, H ig h , and N ich o lso n 1969.

If so much solution takes place close to the surface, the question

of how large caves develop at depth comes to the fore. Where
there is inflow from impermeable rocks, vadose streamflow can
account for this. Water analyses through Clare caves such as
Poulnagallum (Smith and others 1969) suggest that the bulk of
the water passing through these caves has come from higher shale
terrain (Fig. 55a). On entry low carbonate values of 15 to 50
m g/l are measured. Along the main cave there are increments of
dripwater and some tributaries are entirely of percolating water
with concentrations averaging 170 m g/l. The result is that when
little runoff from the shales enters the caves, the risings have high
170 Karst

concentrations though still well below saturation, and when flood-

waters enter the sinks the risings have concentrations only half as
great. In Mendip, the surface catchments on sandstone and shale
inliers are much smaller and contribute much less (Fig. 55b).
Along G.B. Cave there is an increase of 20-100 mg/1 along the
main stream, the increase varying inversely with discharge; Smith
and Mead (1962) think this is mainly due to drips and trickles
and little to solution along the stream channel. As a result the
risings here have a much more uniform concentration than the
Clare ones. Smith and Mead (1962) postulated a steady increase
of solute content along Mendip caves. However, Ford (1966) has
shown that this is not so for Swildon’s Hole in which a tributary
contributes most of the increase. There is also considerable
variation in carbonate content of the flows into the entrance of
G.B. Cave, St Cuthbert’s Swallet, and Swildon’s Hole. Observations
thus confirm theoretical expectations of much variation in the
contributions of seepage water and streamflow of surface origin
from cave to cave and karst to karst.
Because of practical difficulties of observation, the problem of
solution at greater depth along water-filled passages (Bögli 1964a)
remains one largely of theoretical discussion of mixing corrosion
and possibilities for other renewals of aggressiveness high in the
groundwater body.
 VIII

CAVE DEPOSITS

Cave deposits of various kinds often occupy much bedrock cavity,

affect cave form in other ways, and are also important because
cave history is frequently better recorded by depositional evidence
than by erosion, which is liable to destroy its own traces.
Genetic classification of cave deposits is difficult because there
is much inheritance of characters from other environments and
much mingling. Kyrie (1923) made a basic division between
materials formed in place and transported sediments but this does
not correspond simply with autochthonous and allochthonous
origins (Kukla and Lozek 1958) since river deposits, for example,
may be from either source. Others have gone straight to a division
on the basis of mechanism of deposition, e.g. mass movement,
water-laid, chemical and biological. Other mechanisms not so
common include glacial, aeolian, and the freezing of water. Here
space permits only an empiric and highly selective outline.

SPELE OTHEM S

Depositional forms growing from chemical precipitates are

ambiguously referred to as ‘cave formations’. ‘Concretion’ is too
restrictive in its English connotation. ‘Cave decoration’ could be
satisfactory but the American term, speleothem, is gaining general
currency. The various processes of deposition are discussed in
Chapter III; precipitation of calcitc brought about by diffusion of
C 0 2 from water to cave air is the dominant one. However, the
chief controls of growth of speleothcms in Postojna Cave, Slovenia,
are the amount and hardness of the seepage water, not variations
in COo content of the cave air (nor of evaporation, of course;
Gams 1968).
171
172 Karst

Calcite deposited from drips before they are detached from

ceilings and walls forms downward growing stalactites and upward
growing stalagmites form where drips splash (PI. 30). With straw
stalactites (Fig. 56a) the drip deposits a circular ring of calcite

C4 \

56 (a) Development of a straw stalactite, blockage of capillary and growth

to conical form.
(b) Crystalline structure of angular helictite, with abrupt changes in
growth direction. After Prinz 1908.
(c) Curving helictite produced by gradual change in direction of c-axis.
After Prinz 1908.

to form a vertical tube about 5 mm in diameter. The greatest

recorded straw length is 6 m in Easter Cave, Augusta, south­
western Australia. Usually straws get blocked internally while
much shorter and then films of water flow down the outside to
build a conical stalactite with radially oriented crystals around
the central tube. Such stalactites may become very much longer
 Cave Deposits 173

than straws; in Aven Armand, Lozere, France, one is over 30 m.

The rate of growth varies in place and time in accordance with
the drip rate, the state of the CaCO.rCCF-hFO system, and
the shape of the stalactite. A common rate of growth for straw
stalactites is about 0 2 mm/year (Moore 1962). If the drip point
shifts, a fresh stalactite usually forms but with stalagmites which
are usually much broader in relation to height because of splash
and flow, a shift in the drip point simply leads to a complex
doming. Very slow drips may inhibit stalagmite growth, whereas
too fast a rate may prevent stalactite formation.
A distinction is sometimes made between dripstone and flow-
stone. However, all but the simplest stalactites involve flowing
films as well as drips. When a stalactite and a stalagmite meet to
form a column (PI. 30), dripping is eliminated but not growth.
Curtains and shawls are due chiefly to trickles down walls. Rim-
stone dams or gours grow up on floors to impound pools of
various sizes and depths. Aeration over a growing rim promotes
loss of CO.» and calcite precipitation. Siffre (1959) has invoked
laminar flow as an essential condition for gour formation. They
form in rivers as well as from thin flows over floors as is well
exemplified over a kilometre of stream in Croesus Cave, Mole
Creek, Tasmania, but deposition may only occur when flow is
small. Calcite deposits in rimstone pools often have an external
crystalline form reflecting their internal structure such as dog tooth
spar. Floe calcite forms as a thin film on pool surfaces and can
accumulate on the bottom.
Helictites are eccentric speleothems which defy the law of
gravity, growing in any direction from solutions fed along tiny
capillaries. Angular helictites involve abrupt changes in crystal
lattice orientation (Fig. 56b), but gradual changes in axis of
curving forms are as common (Fig. 56c; Prinz 1908). Some
helictites consist of or are covered in calcite crystals with external
crystalline form without ever having been immersed in standing
water. Many forces and processes have been cited to explain
helictites (Moore 1964) but it cannot be said that any explana­
tion has yet acquired general support. Curving crystal forms are
very common in gypsum and halite speleothems (Lowry 1967b).
Many other kinds of speleothem occur and are discussed in the
literature (Warwick 1962).
Speleothems occur in actively developing river caves but their
174 Karst

formation can become dominant in passages and chambers aban­

doned by streams. It has been claimed that speleothems form rarely
when water temperatures approach freezing (Corbel 1960), that
stalactites are more frequent relative to stalagmites in cool
temperate climates (Corbel 1952), and that stalagmites and bulkier
forms generally assume dominance with increasing warmth in
tropical caves (Corbel 1959b).

CAVE ICE

In snow climates, snow drifts into potholes. Thick banks may

block shafts and stop warm air circulation in summer. Parts of
the drift may survive summer melting and turn into ice masses of
semi-permanent nature, often conical in form.
More diversified ice bodies accumulate in the lower parts of
caves where cold air enters in winter and freezes descending
seepage water, e.g. in Eisriesenwelt and Rieseneishöhle near Salz­
burg, Austria. Floor ice may be as flat as an ice rink or steeply
but smoothly sloping. It is misleading to label such ice masses
underground glaciers because they are immobile; the inter­
nationally used French term glaciere designates caves with ice in
them (Balch 1900). Icicles, stalagmites, and columns of ice are
broadly equivalent in form to their speleothem counterparts and
may become very big. They can form and degenerate over com­
paratively few years (Kyrie 1923).

BIOGENIC DEPOSITS

Animals living completely in caves are small in size and even as

large populations cannot build significant accumulations, so
biogenic deposits are ultimately derived from outside caves. Farge
masses of excrement are piled up by colonies of larger animals
which feed outside caves. Bat and bird guano in great conical
masses beneath roosts in chambers are the chief materials of this
kind.
Animal bones also accumulate. They may be from animals
which sheltered in caves (e.g. in Mixnitz Cave, Austria, 500,000
individual cave bear skeletons have been counted) or from
animals that fall accidentally into potholes and vertical shafts.
 Cave Deposits 175

The prey of carnivorous animals also contributes to cave bone

deposits, including the pellets of owls and food refuse of human
kitchens.
With time these organic materials are mineralised. Phosphates,
nitrates (saltpetre), and other minerals are formed.
Primary biogenic accumulations are subject to water transport
at any stage and may be incorporated in clastic sediments. In most
well populated countries, biogenic deposits have usually been
mined for fertiliser and, in the case of saltpetre, for gunpowder,
so that the floors of caves from the United States through Europe
and Asia to Australasia which had such deposits are now in a
very modified state through removal and translocation.

CLASTIC SEDIMENTS

Clastic sediments vary a great deal in caves because of diverse

origins, modes of transport, and environments of deposition.
Breakdown deposits. Breakdown of roofs and walls usually
results in massive crudely conical heaps of angular blocks of rock
on cave floors (PI. 3; PI. 38), though some blocks may be partly
rounded by solution before detachment. Finer grained aggregates
result from salt weathering of the surfaces of dry caves. Textural
sorting is generally poor, and textural parameters are frequently
bimodal because inherited textures are superimposed on break­
down textures.
Gash-breccias and bone-breccias are exceedingly characteristic
of caves and have long occasioned interest (e.g. Pengelly 1864).
These are very mixed deposits with large bedrock fragments in
clay matrix and sometimes with sand. Bones and also speleothem
materials, which may be in place or transported, may also contri­
bute. These breccias come from karst bedrock, insoluble residues,
surface soil materials, secondary precipitates, and animal remains
sliding under gravity down fissures and shafts. They are found
mainly as fissure fills but also in cave chambers, in part because
of cave enlargement since their emplacement, as in Dip Cave,
Wee Jasper, N.S.W. (Jennings 1963).
Stream deposits are commonly of external origin but insoluble
residues, rockfall, and speleothem fragments may be shifted around
within caves by streams. Well sorted and rounded sediments of
greatly varying modal size from boulders to clay result. The size
176 Karst

frequency is normally skewed to the fines because of clays

inherited from the insoluble fraction of the bedrock. Mineral
composition is very varied where the cave streams are bringing in
materials from outside the karst. Calcite sand is rare but dolomite
commonly yields sand-sized sediment, e.g. in Grotte de Moulis,
Ariege, France. Quartzose sands may be from karst bedrock or
from surrounding impervious rocks.
Many sedimentary structures are encountered, such as cross­
bedding and cut-and-fill. Graded beds, the products of big floods
into the cave, are common, not only in areas which suffered
glaciation, with attendant glacial meltwaters, but in a wide variety
of karst environments. Frank (1971) describes graded beds of
19,000 B.P. from Koonalda Cave in the Nullarbor Plain;
mineralogically these do not differ from similar sediments being
deposited now in the present semiarid climate so he finds no
evidence in this for a former wetter climate. Lamination may be
especially common in cave stream sediments because of fine frac­
tions supplied by insoluble residues and of great fluctuations in
flow (R. M. Frank, pers. comm.). Siffre and Siffre (1961) have
presented shape data of limestone pebbles which have passed
through inverted siphons and argue that pressure flow results in
highly rounded but flattened forms. Cementation of cave stream
deposits is common because of frequent occurrence of super­
saturated carbonate waters.
As well as active channel deposits, abandoned stream deposits
are frequently found in caves, e.g. as terraces. Some caves may
be practically filled with gravel like Baldocks Cave, Mole Creek,
Tasmania, where the fill is probably of Pleistocene glacifluvial
origin. However, too frequently inference of thick or complete fill
has been made from small patches of deposit in high positions in
cave walls. Often these are no more than slip-off slope deposits
and meander niche fills which have no such implications, but
register only former levels in cave incision and stream lowering
(Fig. 53j). Floods can deposit fine sediments on ledges at many
levels at the same time. Alluvial passage blockages sometimes
deflect cave streams along different courses or cause a fresh
channel to be cut in bedrock above the fill (Fig. 53f).
Red clay. Red or ochreous clays, sometimes unctuous, are
found widely in limestone caves in many climates; within cave
systems they are probably most frequent in phreatic sections. They
 Cave Deposits 111

may include silt and fine sand but clay minerals predominate,
kaolinite being the most common. They are unbedded but liable
to desiccation cracks, into which other materials may be intro­
duced subsequently. The reddish colour is usually attributed to
ferric iron.
In the past these clays were regarded as a product of internal
weathering of cave bedrock surfaces. But Bretz (1942) thought
that in many caves the volume of limestone removed could not
have provided the necessary bulk and he maintained they were
derived from surface soils and were deposited in stagnant phreatic
conditions. Detailed work by Reams (1968) on the Ozark Moun­
tains caves shows that the caves are shallow phreatic in origin but
that the sediments were deposited by streams bringing in surface
materials after the surface rivers had incised their valleys and a
vadose phase had supervened.
In Austria and Switzerland, cave loam {Höhlenlehm) was
similarly regarded as cave weathering residue. The term has been
applied to a range of sediments from sandy loams to heavy clay.
Bögli (1961b) has shown that cave loam in the Hölloch system,
Muototal, Switzerland, was brought in from the surface by streams
and only calcium carbonate cement is spelean in origin.

ENTRANCE FACIES

Deposits in and near cave entrances have been studied more than
all other cave deposits, partly because of their accessibility but
more because of their valuable complex stratigraphic sequences
and archaeological associations (Schmid 1958).
Here surface materials are moved in by various mechanisms
and interbedded with materials originating in the cave. Surface
soil, weathering mantle and bedrock may fall, slide and creep,
especially in periglacial conditions when regolith is exceedingly
mobile and frost wedging provides much angular rock. Pleisto­
cene glaciers have forced moraine into cave entrances and pot­
holes; these often form a substratum for later accumulations.
Varves may be deposited by glacial meltwaters farther in.
Entrance facies may receive aeolian accessions in the form of
loess near Trade Wind deserts and near glacial outwash plains in
higher latitudes in Pleistocene cold periods. Bones and excrement
register phases of occupation of cave entrances. Fossil assemblages
178 Karst

may reflect external environmental changes in some detail;

mollusca have been particularly useful in this way (Lozek and
others 1956). Men have similarly left a physical record of their
occupation behind, with tools and hearth ashes as well.
Roof fall may be intercalated between these layers of allochtho­
nous material or be incorporated in them. Periglacial phases may
accentuate rockfall because frost wedging can reach tens of metres
into caves in rigorous climates. In addition, calcite floors may be
precipitated over entrance areas and authigenic minerals are
found in the voids of earlier sediments, chiefly calcite, aragonite,
gypsum, salt, and phosphates.
Bedding can be complex in the entrance facies, frequently with
high depositional dips. Nevertheless the variety of deposits laid
down in sequence has provided much evidence for Quaternary
history in many countries.
Riedl (1961) has stressed the importance of extending the
manifold field and laboratory analyses applied to the entrance
facies to all cave deposits and of relating these as closely as
possible to speleogens on all scales. In this he stresses the import­
ance of a climato-morphogenic approach. Isotopic studies of
speleothems are being intensified with promising results. Hendy
(1969) has combined C-14 age and oxygen palaeotemperature
determinations of stalagmite calcite to produce a palaeotempera­
ture curve over the last 100,000 years for the Waitomo caves area
of New Zealand, whilst Duplessy and others (1970) have con­
structed a similar record for the Riss-WUrm interglacial (130,000-
90,000 B.P.) from a stalagmite 3 km inside Aven d’Orgnac,
France.
 IX

INFLUENCE OF CLIMATE AND

RATES OF DENUDATION

Classical ideas about karst arose chiefly from studies of the Dinaric
karst and that of central Europe. Differences between them were
appreciated. For instance, central Europe was noted for possessing
a greater proportion of covered karst than the Dinaric region. In
order to explain these variations, climatic factors such as frost
incidence and rainfall intensity were invoked (Grund 1910b),
but the possible relations were obscured by complications due to
forest clearance and geological structure. The dominant mode of
thought as exemplified by Sawicki (1909) and Grund (1914)
was to range all karst phenomena into a single evolutionary
sequence or cycle. Even strikingly different karst forms described
from tropical humid Jamaica and Java (Danes 1908, 1910) were
regarded as falling into place in such schemes, and landform terms
from the Dinaric karst were applied without much discrimination
the world over, almost with the effect of mental blinkers (Lehmann
1960). With the emergence of general concepts of climatic
morphology in the 1930s came a clearer realisation that karsts in
contrasting climates might have independent modes of develop­
ment. This was most evident in a study of Javanese karst by
H. Lehmann (1936) but had been foreshadowed in the work of
O. Lehmann (1927) on changes of karst features with altitudinal
climatic zonation in the Austrian Alps.

RATES OF KARST DENUDATION

After World War II geographers interested in karst became pre­

occupied with climatically controlled morphogenic systems; at the
1953 meeting of the International Geographical Union’s Karst
179
180 Karst

Commission (Lehmann and others 1954) no fewer than eight

climato-morphogenic regimes were discussed. Corbel in particular
(e.g. 1957) pursued this question and he gave it a quantitative
slant by advocating measurements of rates of karst denudation
(Corbel 1959a). His formula

E= runoff in deci-
Limestone denudation 4 E T n metres,
(m3/km 2/year, or = ----------- where T = mean CaCO.j
mm/1000 years) content in mg/1
1
— = fraction of
n catchment in
limestone and
limestone
alluvium
has subsequently been improved by Williams (1963) and Douglas
(1964) by allowing for departures from assumed density of 2-5
for the karst rock and by substituting total carbonate hardness for
calcium carbonate alone. Greater difficulties arise in obtaining
adequate field data. Corbel frequently employed far too few figures
of solute content for the mean to be representative and, since
discharge data were frequently unavailable, annual surplus of the
water balance (precipitation minus evaporation) was substituted.
Climatic figures for evaporation are notoriously problematic and
abnormal infiltration in karst affects this calculation particularly.
Table 6 presents a selection of karst denudation rates from
different climates, about half of which are based on discharge
values and sufficient chemical determinations.
In recent years there has been some reaction to what is claimed
as overemphasis
v
on karst differences attributed to climate (Panos
and Stelcl 1968), moreover detailed studies reveal sharp variation
in spatial and temporal distribution of solution within given karst
situations (Gams 1966; Douglas 1968; Pitty 1968; Smith and
others 1969). Undoubtedly there has been a tendency to overlook
greater differences due to other causes such as lithology in the
search for the operation of the climatic factor and to claim greater
interpretative certainty in the face of the complex interactions with
many other factors. Nevertheless the broad reality of climatic
control in karst cannot be denied, even if the nuances have not
been evaluated.
 Influence of Climate 181

T able 6 Rates of karst denudation

Net rate of
Koppen Mean denudation
Karst area climatic annual m 3/ km*/y Source
type precipita­ or m m /
tion (mm) 1000 y

Som erset Island, ET 130 2 Smith 1969

N . C anada
T an an a R., Central Dfc 450 40 Corbel 1959a
A laska
Svartisen, N. Norway Dfc 740-4000 275-5000 Corbel 1957
Vercors, France Dfb 1500-2500 240 ,, ,,

Punkva R., M oravia, Dfb 620 25 Stelcl and

Czechoslovakia others 1969

Fergus R. and Cfb 1000-1250 51-53 Williams 1963,

Shannon R., Ireland 1970
C raven, England Cfb 1250-1500 40 Sweeting 1965
Peak District, Cfb 800-1200 75-83 Pitty 1968
England
Mellte R., Wales Cfb 1600 16 G room and
Williams 1965
M endip, England Cfb 900-1100 40 Corbel 1959a
Slovenia, Yugoslavia Cfb 1250-2000 10-100 G am s 1966

S. Algeria Bwh 60 6 Corbel 1957

Los Alamos, New Bwh 25-40 < 1
Mexico, U.S.A.
G rand Canyon, Bwh 25-50 7 9» 99

Colorado, U.S.A.

Kissimee R., Cfa 1200 5 Corbel 1959a

Florida, U.S.A.
Yucatan, M exico Aw 1000-1500 12-44 99 99

Indonesia Afi 200-3000 83 Baläzs 1968

THE ARID EXTREME

Water is necessary for karst development; vegetation and soil

microbial activity promote it, chiefly by providing much more
C 0 2 than the atmosphere alone. Regions of low precipitation,
particularly those of high temperature where evaporation renders
31 Kestrel Cavern No. 1, Nullarbor Plain, Western Australia. Extensive,
almost featureless surface of retarded arid karst, with rare collapse
dolines which retain freshness of form for long periods.

32 Hemispheroid conekarst in Miocene limestone at lehi, Lower River

Kikori, New Guinea. Partly cleared of rainforest.
184 Karst

rainfall less effective, have poorly developed karsts in consequence.

The Nullarbor Plain in Australia, one of the world’s largest
karsts, with not far short of 200,000 km2, is entirely semiarid and
arid in climate and much of it is treeless as its name implies.
Uplifted in Miocene times with only gentle tilting, it remains an
unbroken low plateau (PI. 31). There are virtually no valleys but
a very gentle shallow undulatory relief, generally no more than
3-6 m in amplitude, is a product of differential surface solution,
guided by joints. There are a few collapse dolines which long
retain their sharpness. Minor surface sculpture is minimal and
best developed near the coast where sea spray drifts on to the
rocks. Less than a score of large deep caves are known which
reach down to a flat watertable with brackish slowly moving water
in a lower highly porous chalky limestone. Shallower caves are
more numerous but the underground as well as the surface karst
is greatly retarded despite ample time since emergence (Jennings
1967c). Gypsum and halite speleothems assume a greater than
normal importance as does salt weathering in caves (Lowry
1967b). There is a great deal of calcrete induration of surface
limestone.
These characteristics are common for karst in hot, dry climate
(Wissmann 1957; Conrad and others 1968). Corbel (1959) has
reported low rates of present denudation in accordance with these
traits (Table 6). Where there are regional relief contrasts of
tectonic origin or inherited from previous geomorphic phases,
limestone or dolomite karst acts as resistant rock because it is
almost immune from surface water erosion. Evaporite karst
persists at the surface more than in humid climates and even salt
domes provide projecting relief for some time after stripping, then
leading to craterlike forms and salt swamps (Harrison 1930).

THE COLD EXTREME

In cold climates where water on the surface and in soil and rock
is frozen most or all of the year in association with geophysically
polar glaciers and with permafrost, karst development is also
inhibited. Beneath glaciers, claims Corbel (1954), limestone acts
as resistant rock but resultant landforms comprise glacial, not
karst relief. In the permafrost zone, underground karst develop­
ment stops and even though there is surface water in a short
 Influence of Climate 185

summer, total amounts are low. Frost wedging destroys incipient

surface solution features and shillow clutters outcrops. Because
structural forms are prevalent, it is thought that the surface is
lowered parallel to itself, though solution is more active beneath
snow banks than along streams. According to Corbel, surface
lowering proceeds slowly because of low available amounts of
water and, according to Smith (1969) on the basis of observations
from Somerset Island in northern Canada, because of poverty in
biogenic CO 2 through low vegetative productivity. However,
Corbel (1959) reports extremely high carbonate contents from
deep wells below permafrost levels in northern Alaska where
karst processes may be slow but have not stopped.
Conditions are different in maritime and high mountain cold
climate and in cool temperate snow climate. Here glaciers are
accompanied by abundant meltwater and bedrock temperatures
permit infiltration of water. Beyond the glacier margins, frozen
ground is only seasonal and snow melt provides seasonal abund­
ance of liquid H 2 O. In these conditions there can be vigorous
development of karst. Close-set fields of small solution dolines
and potholes occur beneath forest and in alpine meadow and

33 Close set potholes and dolines in Palaeozoic marble, Mt Arthur, Nelson,

New Zealand. Shillow on outcrops due to frost shattering.
186 Karst

tundra conditions (PI. 33) though they may be subordinate to

larger glacial forms. In the Queyras district of the French Alps,
glacial rock basins in dolomitic limestone lack lakes through
underground drainage and their glacial moraine cover is pitted
with subsidence dolines. Minor surface solution sculpture is some­
times well developed, but the effects of recent ice action may not have
been removed or frost shattering may still be too persistent for the
development of minor forms. Some kinds of Karren, e.g. solution
funnel steps, appear to be genetically linked to snowmelt (but they
are also found on the Irish coast (P. W. Williams 1966a)).
Cave development can be extensive and in high mountain karsts
of this type, very deep systems have developed, especially in the
Alps and the Pyrenees. The caves are usually poor in speleothems
and glacieres occur. Corbel (1957) claims that in these climates
limestone acts as a weakly resistant rock, though young tectonism
may prevent it from finding present expression. In northern
Norway and western Tasmania, limestone outcrops are disposed
along valley floors and karst has largely been destroyed in these
older tectonic milieus.
There is argument about the cause of advanced karst in these
kinds of climate. Corbel (1954, 1957) has stressed higher C 0 2
saturation equilibria of cold waters and the presumed high C 0 2
content of the voids in snow and ice. However, a more common
view is that saturation values rarely control development and the
kinetics of solution are more important. Observations by Ek
(1964) on ice melt and meltwater rivers in the French mountains
show higher pH and lower carbonate concentration than Corbel
claims; they do not testify to special aggressiveness. High precipi­
tation and low evaporation are admitted by all as explaining high
denudation rates in karsts such as those of Svartisen in Norway
and Vercors in the French Alps. Even in the inland taiga-tundra
basin of the River Tanana in Alaska, with a modest annual
precipitation of 450 mm, the bulk of solution takes place during
the short summer season of great floods when carbonate concen­
trations are lowest (Corbel 1959a). Sheer volume of water and
its turbulence dominate solution.

THE ‘BOTANIC HOTHOUSE’ EXTREME: TROPICAL HUMID KARST

The most important fact about karst in tropical humid climates

with or without a dry season is its greater variety than elsewhere
 Influence of Climate 187

(Verstappen 1960a, 1964; Jennings and Bik 1962). The many

type names applied— cockpit karst, tower karst, labyrinth karst,
crevice karst, polygonal karst, arete and pinnacle karst are
examples—testify to this. That climatic factors are not the only
ones involved is illustrated by the range of explanations given for
doline karst in hot, humid climates. These consist of dolines
perforating planation, structural or tectonic surfaces in a manner
and to a degree similar to that found in classical mid-latitude
karsts. Such doline karsts have been attributed in the tropical
humid context to marly limestone or to interbedding with
impervious rocks as in the Aguada Limestone of Puerto Rico
(Monroe 1968); to mechanically weak limestones as in Yucatan;
to longer dry season as in northern Jamaica; or to closeness to
sea level as in subtropical Florida.
Great variety is a pointer to extremely active karst development
and many investigators have interpreted the most striking kinds
of tropical karst as the product of rapid, vigorous solution. These
are the kinds now commonly subsumed under cone karst (Ger.
Kegelkarst) in which the relief is dominated, not by closed
depressions, i.e. reduction forms, but by projecting residual
forms, rarely, however, of conical form (Fig. 57, PI. 32). Two
varieties have been contrasted in the literature, though in reality
they grade into one another in some contexts and there are many
areas which conform to neither (Aub, in press a).
Cockpit karst is typified by the Gunung Sewu karst of south
central Java (Lehmann 1936; Flathe and Pfeffer 1965), though
the name comes from similar country in Jamaica (Lehmann
1953; Sweeting 1958). Here the residual limestone hills are
characteristically hemispheroidal in shape (Kugelkarst). The
Gunung Sewu are about 30 to the square kilometre and 30-70 m
high. The closed depressions between them, the cockpits of
Jamaica, are often of star shape, with sides bulging inwards (see
p. 134; Fig. 43). Often the hills are aligned and the cockpits
strung out in chains along ‘glades’ between. The relief is clothed
naturally in dense rainforest; red earth and rendzina soil (Pfeffer
1969) cover variable proportions, sometimes confined to the
depression floors, whilst forest litter patchily hides craggy slopes.
Despite the forest, water may course down the slopes during
intense rainfall of tropical storms, especially where adjacent slopes
converge.
188 K a rs t

XT^

0_____________________________ 5p0 M e tre s

0 500 1000 1500 Feet
C o n t o u r i n t e r v a l 10 m e tr e s

57 Cone karst in n o rth e rn P u e rto Rico. F ro m U.S. G eol. Surv. map.

Lehmann (1936) interpreted the Gunung Sewu cockpit karst

as due to rejuvenation of surface drainage by tectonic updoming
of a planation surface on limestone (Fig. 58). This produced
ridges and valleys aligned down the domes. Then with the
development of vertical infiltration of the water and underground
drainage, the valley systems were broken up into chains of closed
depressions of surface sollutional origin. Detailed work by Aub
(in press b) in a small area of Jamaican karst confirms a surface
 Influence of Climate 189

solutional origin of its cockpits. In other cockpit karsts, alignment

of hills and cockpits is controlled by jointing, not necessarily
related to the latest tectonic uplift, and in others, e.g. along the
lower Kikori River, Papua, the former presence of a planation
surface is in doubt. These differences, however, do not preclude
a basically similar evolution.
Tower karst (Ger. Turmkarst) differs from cockpit karst in
much steeper lower slopes to its residual hills, which can be
vertical or overhanging, and in the presence of swampy alluvial
plains around the towers (Fig. 59, PI. 34). The towers themselves
may surround flat-floored depressions, basically of polje type.
Karst margin plains frequently accompany tower karst and deep
karst corridors, vertically walled linear depressions, often cut up

58 Evolution of cockpit karst of Gunung Sewu type according to Lehmann

1936
190 Karst

the hill masses. Surface drainage over the flat plains may pre­
dominate areally. Horizontal tunnel caves through the towers link
the poljes with plains which have surface drainage to the sea and
the levels of their outflows (Vorfluters) regulate the polje floor

Active High Level Abandoned river caves

alluvium with gravels and guano Cliff-Foot
I V
River

River

59 Section through tower karst based on Kinta valley, West Malaysia

34 Tower karst in Permo-Carboniferous limestone in Perlis, West Malaysia.

Towers rise 360 m above alluvial plain.

heights. Inactive higher level caves are the sites of guano accumu­
lations and phosphates (Wilford 1964).
Most Malayan karst, e.g. in the tin and iron mining Kinta
valley with towers up to 500 m high, is of this type and so also
 Influence of Climate 191

that of Sarawak (Wilford and Wall 1965). Tower karst is

described from Cuba where the towers are called ‘mogotes’
(Lehmann and others 1956) and southwest China and Vietnam
(Wissman 1954; Geliert 1962; Silar 1965).

In Gunong Temperong In Gunong Rapat

25 0 6Metres
A Swamp slot B Solution notches C Stalactite D Alluvium
E Bat guano

WSW Trade Winds

Thin calcrete and Thick calcrete on

cavernous weathering windward side with
on leeward side with much evaporation
less evaporation

60 (a) Cliff-foot cave sections from Kinta valley. After Walker.

(b) Mogote asymmetry in Puerto Rico resulting from differential calcrete
formation according to Monroe 1966.

Tower karst is generally attributed to lateral solutional under­

cutting by river floods and swamp waters around the margins of
alluvial plains, which tin, iron and gold mining in Malaya and
Sarawak, for example, have shown to overlie limestone surfaces,
intricately cut up in detail by subsurface solution but planed off
flat overall. Cliff-foot caves (Fig. 60a) are frequent round the
bases of towers though by no means present everywhere; these
are low-roofed, narrow caves elongated along the tower margins
with occasional passages leading more deeply inwards. Swamp
slots and solution notches (Chapter IV) are also common.
192 Karst

frequently in association with depressions marginal to the alluvial

plains. Many tower karsts give strong expression to the tendency
for karst solution to operate dominantly in vertical and horizontal
planes.
Whether it is cockpit or tower karst which develops in any area
has been variously explained. In Tabasco, Mexico, tower karst is
thought to follow cockpit karst as a later stage of development
(Gerstenhauer 1960) whereas in southwest Celebes nearness to
impervious basement is the control preferred, shallow karst
favouring cockpit karst (Sunartadirdja and Lehmann 1960).
Renault (1959) thinks that less compacted, more jointed and
faulted rocks support cockpit karst only in Kouilou, Gabon, but
for West Irian, Verstappen (1960a) relies on variation in porosity
as the determinant, the more porous giving rise to cockpit karst.
In Puerto Rico, presence or absence of a former cover of super­
ficial deposits is regarded as critical, cover resulting in tower karst
(Monroe 1968).
In Puerto Rico there is asymmetry in the residual hills that is
not related to attitude of the limestone but to the prevailing winds;
Thorp (1934) attributed it to greater rainfall and consequent
solution on the windward eastern side, producing gentler slopes
there. According to Monroe (1966), however, the mechanism is
one of greater superficial induration of weak limestone on the
windward side (Fig. 60b). More wetting and drying takes place
there with greater solution and reprecipitation. The thinner cap-
rock of the western sides allows readier attack and undercutting
of the weak interior to steepen these flanks. His evidence makes
plain the importance of case-hardening of this type in the resist­
ance of the residual hills to denudation here but many other tower
karsts, e.g. most Malayan karsts in dense Palaeozoic limestones,
owe little or nothing to such processes.
Tropical limestone towers are not always set in alluvial plains.
In Malaya many stand with customary abruptness on steep granite
slopes (Fig. 62) and in New Guinea (Jennings and Bik 1962)
they often rise from supporting conical or pyramidal limestone
hills. How the clifT-foot angle is initiated and maintained is easier
to understand in the first case (because water must issue at the
limestone-granite contact) than in the second circumstance. Nor
are the plains associated with towers always developed on karst
rocks. For example, tower karst surrounds some of the ‘interior
 Influence of Climate 193

valleys’ of Jamaica, large closed depressions of which the floors

are entirely inliers of an impervious basement.
These variations do not invalidate the general view that enor­
mous amounts of limestone have been removed rapidly to create
these elaborate karst styles. Most authorities have explained this
in terms of the climate and the vegetation (Lehmann 1953; Birot
1949). Chemical reactions proceed faster because of higher tem­
peratures; high rainfall and high intensities of rainfall make for
prolonged and rapid solution; rapid plant growth and decay and
intense microbial activity with high soil Pco^ make tropical water
very aggressive.
Corbel (1959a) opposed this ‘botanic hothouse’ interpretation
on the grounds of lower saturation equilibria for CCL and car­
bonate solution with high temperatures and of low observed
carbonate concentrations in tropical karst waters. He explains the
undoubted elaboration of these karsts by the great length of
geological time which has been available for uninterrupted
development, whereas Pleistocene glaciations have intervened in
many high and mid-latitude karsts, some of which have developed
entirely in the Holocene. Such histories may be true in certain
cases; in Clare, Ireland, the caves and underground drainage are
of Postglacial age (Ollier and Tratman 1969); Malayan karst
probably reaches back into the Tertiary in continuous develop­
ment. On the other hand some mid-latitude karsts, e.g. the
Dinaric, have inherited some features from the Tertiary and
there are conekarsts in New Guinea which were not uplifted till
the Pleistocene and are very young (Jennings and Bik 1962).
There is much reprecipitation of carbonate in humid tropical
karsts so that net removal rates may conceal greater absolute
solution. But it is probable that saturation equilibria rarely control
action here. More detailed studies of discharge and water chemistry
in high precipitation tropical karsts are needed.
Nevertheless the complexities of tropical humid karst will not be
solved in climatic terms alone as is evident within one fairly
uniform tropical humid climate with a short dry season in Perlis
and Kedah states of Malaya. Along strike belts of Permo-Carboni­
ferous limestone there are lines of great towers rising up to 360 m
high from coastal alluvial plains (PI. 34). Immediately west of this
typical tower karst there is an anticlinal belt of Ordovician-Lower
194 Karst

Silurian limestone along the Thailand frontier; much of it is of

high purity though some is dolomitic and some impure. In the
main it forms a massive plateau area of 300-600 m, which is
predominantly a doline and polje karst. The poljes are deep and
steeper-sided than typical Yugoslavian ones but otherwise much
resemble them. Small dolines pepper the plateau surface. However,
a lithological explanation is not immediately apparent because the
plateau karst is continued along identical strike belts into the
Langkawi Islands across a very shallow sea (PI. 39). Some of
this karst is comparable but there is tower karst also, some rising
from mangrove swamps and some from the sea. Though marine
notches indicate present-day recession by seawater solution of the
latter, it is unlikely that the presence of tower karst can be the
result of the same conditions as are now found. The presence of
large, round dolines in the island of Pulau Dayang Bunting which
are occupied by freshwater lakes, suggests that the karst may have
formed in part during lower sea level stands. Only combinations
of factors explain this complex of karst types in northwestern
Malaya.

35 Vertical walls and aretes round dolines in Miocene limestone on

Mt Kaijende, Central Highlands, New Guinea
 Influence of Climate 195

FURTHER KARST CLIMATO-MORPHOGENIC SYSTEMS

Corbel (1957) has sought to distinguish a karst style in cool

temperate maritime climates as in Ireland where Williams (1970)
has pointed to the importance of peaty waters rich in organic
acids and where there are persistent light rains and low frost
incidence. However, Corbel cites few distinctive forms other than
badland sculpture of scarps exposed to wind-driven rains but not
subject to frost wedging. Tropical subhumid and semiarid karsts in
Brazil (Tricart and Silva 1960) and in northern Australia
(Jennings and Sweeting 1963) present some traits of tropical
humid karsts, e.g. evolution to tower karst, but others of drier
climate affiliations, including pedimentation, are also present.
Whether this can be explained entirely on the basis of present
climate is not certain (Jennings 1969). Indeed all efforts to
distinguish climato-morphogenic systems other than those asso­
ciated with climatic extremes must be regarded as tentative in
terms of present knowledge (PI. 35).
 X

INFLUENCE OF GEOLOGICAL
STRUCTURE

Geological structure is so powerful and pervasive an influence on

karst that none of the preceding chapters has escaped reference
to it. Nevertheless some separate treatment of it is necessary.
Here structure is taken to include both original depositional
structure of the rocks with their varying lithology and later
tectonic features such as folds, faults, and joints.
Primary depositional and tectonic forms are more evident in
karst than with many other rocks, and karst regions appear dis­
proportionately in landform atlases and geomorphological texts
to illustrate tectonic relief. This has often been explained as due
to a special immunity from surface attack conferred by the
development of underground drainage. Though such karst
immunity may be true in arid and semiarid climates, many
measures of surface solution of limestone in other climates are
of magnitudes which preclude long persistence of surfaces intact.
Thus Aubert (1969) has argued that if the present low rate of
surface ablation of 0 05 mm per year in the Jura Mountains has
been maintained since the Pontian orogenic movements, 500 m
of limestone have been removed. Lowering of surfaces along with
preservation of form by uniform reduction by solution may
resolve this paradox in some cases (P. W. Williams, pers. comm.).

DEPOSITIONAL RELIEF IN REEF AND DUNE LIMESTONE KARSTS

Biogenic reefs are particularly illustrative of this attribute.

Verstappen (1960b) describes various types of primary relief on
emerged ‘coral’ reefs, according to steepness of the basement on
which reefcaps rest, vigour of reef growth, and course of uplift,
196
 Influence of Geological Structure 197

and he also discusses their destruction by solution (Fig. 61).

Rapid emergence of fringing reefs on the north Huon coast of
New Guinea has left a sequence of marine constructional terraces
of well preserved form in a humid tropical climate rising to over
500 m and ranging back to 450,000 ± 100,000 B.P. (dated by
IJ238/234 (Veeh and Chappell 1970)). Narrow, close-set sequences
relate to times of rapid negative movement when tectonic uplift
and falling eustatic sea level worked in opposition; broad terraces
with lagoon facies as well as reef rim mark Stillstands when rising
land and rising sea levels were matched.

1 ; ; :J Reefcap V / / \ Basement

61 Types of emerged reef relief according to Verstappen 1960a

(a) Plateau reefcap on broad platform with good reef growth.
(b) As above but with poor reef growth.
(c) Terraced reefcap on steep basement with intermittent uplift and
good reef growth.
(d) As above but with poor reef growth.
(e) Undulating reefcap with thin reef and no reef edges due to rapid
uplift.
(f) As above but with occasional reef edges.

Emerged atolls such as Rennell Island in the Solomons and

the Trobriand Islands (Ollier and Holdsworth 1968, 1969, 1970)
have an annular outer ridge which was the former reef and a
swampy interior plain which was the former lagoon. The coastal
cliffs of the rim have certainly suffered sea attack, witnessed by
marine notches at various levels, but they also owe much to the
primary precipitous nature of the outer face of the reef.
Active reef growth tends to incorporate many primary voids,
which may be quite large, and there is the possibility that emerged
 198 Karst

reefs have inherited caves from marine conditions. However,

Ollier and Holdsworth found no evidence of this in the Trobriands
where caves commonly are of single level development by solution
close to the present watertable, followed by collapse and much
speleothem development, though there is surprising variety. On
some emerged reef islands caves develop as tunnels leading
radially to the sea such as Tumwalau Cave on Kiriwina and
Bwabwatu on Kaileuna. Most are much modified by collapse.
On Kitava several large caves have large symmetrical scallops
and hellholes indicative of phreatic development; a freshwater
lens may have reached below the sea level of the time of their
formation.
In Bermuda in a limestone formation part reef, part aeolian,
there are caves reaching below sea level with submerged speleo-
thems: according to Swinnerton (1929) they formed during a
period of lower sea level in the Pleistocene. Bretz (1960) main­
tained that when the island was larger with lower sea level, it was
big enough to maintain a body of fresh groundwater that
developed the caves phreatically. Further lowering allowed air to
enter which permitted speleothem growth. But on the small
Trobriand Island of Vakuta the caves seem to have developed
in brackish water so that the establishment of a fresh groundwater
body does not seem to be essential.
The intricate pattern of sounds and bays on Bermuda is entirely
attributed to cave solution and collapse in Pleistocene glacial low
sea levels by Bretz, but Swinnerton thinks only some like Castle
Harbour are of this origin, others being simply due to the flooding
of primary hollows in dunes deposited during the low Stillstands.
It is evident from much of the western and southern coasts of
Australia that primary dune forms can persist through the consoli­
dation of calcareous dune sands to aeolian calcarenite (Jennings
1968). There are multiple sequences of these dune systems and
those on Rottnest Island off Perth, Western Australia, dated at
100,000 B.P. and retaining strong primary relief, are not the
oldest. For Bermuda, both Swinnerton and Bretz postulated that
diagenesis of aeolian sands preceded karst development. However,
in the Australian dune limestones development of some karst
features has accompanied consolidation (Jennings 1968). Some
caves occur in only partially consolidated sands even today. This
partly syngenetic karst has some distinctive features: calcite
 Influence of Geological Structure 199

deposition contributes to solution pipe development (p. 50);

some gorges are depositional in origin (PI. 36); collapse plays a
precocious role in cave development; the form of the crystalline
basement beneath the dunes affects cave formation also as does
differential lithification.

STRUCTURAL RELIEF IN REEF KARST

Even in ancient reefs, structure may remain evident through its

guidance of solution as is well illustrated by the Devonian reefs
of West Kimberley (Fig. 3b). Only modestly disturbed by tec­
tonics, these were buried beneath Permian sandstones and then
were truncated by a Mesozoic-Tertiary planation surface. Subse­
quent uplift and rejuvenation have led to removal of the weaker
impervious basin and inter-reef facies so that the barrier and
patch reefs stand out now from shale plains as the low plateaux
and ridges of the Limestone Ranges. Though pedimentation has
cut into reef structures to margin the ranges with erosional cliffs
and scarps, these parallel the strikes of the forereef beds faith­
fully (Jennings and Sweeting 1963).
The narrow band of the reef proper is solutionally etched as
a shallow trench below the forereef and backreef facies. Wherever
either of these facies (though more commonly the lagoonal) tends
to impurity, rounded relief with a normal valley system occurs

36 Deepdene, Augusta, Western Australia. Gorge due to building of Pleisto­

cene dune ridge athwart river course. Consolidated to aeolian calcarenite.

äst-?
200 Karst

with only minor solution sculpture, some caves and springs indi­
cative of karst tendencies. Where pure, both facies give rise to
distinctive, rugged karst relief (PI. 25) evolving from ‘giant grike-
land’, intricate karst corridor and fissure cave terrain, through an
angular box-valley stage, to a pedimented tower karst. Differ­
ences between forereef and backreef karsts are obvious but minor,
largely arising from the depositional dip of the former and the
horizontal attitude of the latter, and from other lithological details
such as the megabreccias of the forereef (Playford and Lowry
1966).
The Chillagoe karst of north Queensland, though subject to
basically the same processes in similar climate of modest rainfall
and long dry season, differs geomorphically through a different
structural context. The rocks here dip steeply, with biohermal
lenses of limestone interbedded with impervious rocks. Pedimen-
tation has produced a tower karst only, each lens encroached on
all sides to yield a tower or group of towers, riddled with caves
and wildly etched by surface solution (Danes 1916; Jennings
1966).

HOLOKARST AND MEROKARST

Some of these differences due to structure are illustrative of

Cvijic’s classic distinction between holokarst and merokarst in
Yugoslavia. The holokarst is found in the extensive coastal belt
of the Dinaric karst which has great thicknesses of pure limestones
reaching deep below sea level and high above it (Gams 1969).
Here the full gamut of mid-latitude karst forms and drainage,
including erosional poljes, occur; karst processes dominate
morphogeny without significant vertical or horizontal boundary
effects.
Inland the karst rocks are thinner and interrupted in outcrop
by impervious rocks which greatly influence the karst. Karst here
is less complete, particularly with regard to poljes, only that of
Pester being typical, and normal valleys are more frequent. This
inland area was that originally designated as merokarst by Cvijic
(1893).
Later, however, Cvijic (1918, 1925) transferred this term to
karsts still less elaborate in character where the rocks are litho-
 Influence of Geological Structure 201

logically even more unfavourable to the evolution of some karst

attributes, especially through the accumulation of residual covers
on the karst rocks. Surface solutional sculpture is weak, but not
subsoil sculpture; dolines form chiefly in the covers rather than in
bedrock; poljes are lacking; caves and potholes may be poorly
developed also, whereas integrated valley systems are frequent.
These characteristics in the main apply to the Cretaceous Chalk
country of England and France, the Jurassic dolomites and lime­
stones of the Franconian Jura, and to other central European
karsts according to Cvijic (1960). The inland Dinaric karst he
then placed in a transitional group with such karsts as those of the
Causses of central France and of the Alps. It may be that there
is every gradation from highly developed karst to its most modest
manifestation over a range of structural frameworks with no
critical thresholds for geomorphic response.

STRUCTURE IN TROPICAL HUMID KARST

Variety of structural relationships is well illustrated by limestone

towers in Sarawak (Wiiford and Wall 1965; Fig. 62). Their
margins may be defined by the top and bottom of vertically dip­
ping limestone interbedded with weaker rocks(a). Where there
are sharp changes in facies, the cliff may be not far removed from
where the limestone passes laterally into other rocks(b). The
facies change may be a less marked one within the karst rocks;
some of the limestone flats are developed in bedded, less pure
calcarenites and calcirudites whereas the hills are made of very
pure, massive, recrystallised coral-algal biolithite(c). Some sides
run along faults juxtaposing karst and impervious rocks(d), and
igneous dykes cut up other limestone bodies into towers separated
by corridors(e) (PI. 37). Nevertheless many margins are not
structurally controlled and lie within the one limestone facies;
these are the product of lateral solutional undercutting from
alluvially veneered limestone plains(f) or of sapping at basal
contact with sloping igneous surfaces(g).
In northern Puerto Rico, five calcareous formations strike west
and the karst has a similar latitudinal pattern of different associa­
tions of karst forms (Monroe 1968). Cockpit karst is associated
with pure, massive limestone, whereas mogote karst requires
similar limestone which has had a young surficial cover of sands
202 Karst

62 Types of limestone tower in Borneo. After Wilford and Wall 1965.

and sandy clays, now dropped by solution to veneer the plains
between the towers. Less pure, well bedded limestones give rise
to zanjones country, shallowly joint-trenched karst. Alternation of
mechanically strong and weak limestone gives rise to deep doline
karst with which are found the greatest cave developments and
the natural bridges. Although the southern part of the karst has
been uplifted more than the northern (Gerstenhauer 1964), the
climate is uniform and it is lithology which chiefly differentiates
karst in this region.
STRUCTURE WITHIN CAVES

Where limestones or dolomites are nearly horizontal, cave plans

often show much joint control (Fig. 50), whereas in cross and
 Influence of Geological Structure 203

longitudinal sections, bedding becomes important in delineating

roof and floor, with joint planes influencing walls, e.g. in many
caves of Craven, England and Clare, Ireland.
When dip becomes substantial, strike directions frequently
become important in plan whilst cross-sections tend to be tri­
angular with a sloping bedding plane roof. Along the strike, level
passages occur in the bedding but down the dip bedding and joint­
ing alternate in control to yield ‘up and down’ profiles developed
phreatically until destroyed or evaded in vadose development, e.g.
in Mendip, England (Ford 1965b; Fig. 63). Deep potholes such
as Gouffre Berger have oblique passages down the dip alternating
with gravity-controlled vertical pitches.
When beds dip more or less vertically, the cave plan may be
dominated by the strike of the beds and the walls are frequently
bedding planes, e.g. in Mair’s Cave, Flinders Ranges, South
Australia (Fig. 64). Joints governing breakdown make ceilings
irregular where these are not planed by water, e.g. in the Dip,
Wee Jasper, N.S.W.
Recently Bögli (1969) has argued that the role of joints in
cave formation has been exaggerated. Because of their greater
continuity, bedding planes are more important in the initiation of
phreatic caves but the elliptical passages developing in them mis­
leadingly lose that form by secondary breakdown (incasion) along
joints (Fig. 53g). Vadose systems, however, will normally
originate in open joint conditions which then can prevail in struc­
tural guidance. The world’s longest caves do not owe much to
joints in their passage directions as Bögli himself shows for the
Hölloch Cave in Switzerland and Deike (1967) for the great
Kentucky caves.
The different attitudes of cavernous rocks are part and parcel
of larger tectonic structures, but closer links can exist between
the latter and caves through their guidance of solution. Narren-
gullen Cave near Yass, N.S.W., lies along the axial trough of a
syncline at the base of limestone beds surviving only as an inlier
in this tectonically low position. Some caves have developed along
faults; in Craven, Rift Pot. Long Kin East Pot, Hull Pot and
Meregill seem to be of this nature (Myers 1948), though more
general claims of fault control in this area are invalid. Faults do
not always favour cave development, possibly because of
mineralisation or too great a tendency to collapse; Glennie (1950)
204 Karst

points out how Ogof Ffynnon Ddu in South Wales has two main
systems separated by a fault zone where passages are few.
The details of caves generally reflect structure with a faithful­
ness as great as that of desert weathering. In northwest Clare,
chert bands form false floors and narrow shelves on passage walls,
which also result from modest differences in limestone lithology

37 Joint controlled karst corridor in tower karst in Kinta valley, Perak,

West Malaysia
 Influence of Geological Structure 205

(Ollier and Tratman 1969). Calcite veins sometimes are resistant

to solution and may project as boxwork (Bretz 1942); alterna­
tively voids along the centre of the veins promote solution and

38 Rectangular cross-section and blockpile in horizontal Ordovician lime­

stone in Exit Cave, Tasmania. Photo by R. Curtis.
206 Karst

then they are recessive (Ollier and Tratman 1969). In Craven,

bedding-plane passages are frequently associated with thin shale
bands in generally pure limestones (Sweeting 1950). Tn Black

Fir st base
Level

Second
base Level

of f i r s t w a terta b le
Vadose p a s s a g e P hrea tic passag e
of second w a t e r t a b le

63 Influence of steep dip on Mendip cave development. After Ford 1965b.

64 Influence of vertical beds in Mair’s Cave, Flinders Ranges, South

Australia. After Cave Expl. Grp S. Australia survey.
 Influence of Geological Structure 207

Range Cave, Cooleman Plain, N.S.W., felsite dykes cause doglegs

when the stream passage meets them. These few instances must
suffice to point to an aspect of cave geomorphology of infinite
variety.

HYDROGEOLOGIC SYSTEMS

Not only are the details of individual caves responsive to structure

but the whole nature of cave development and the associated
underground drainage. This is illustrated by the seven types of
hydrogeologic systems that White (1969) has conceived for karst
of low to moderate relief (Table 7, Fig 65). Karst rock lithology,
depth of karst in relation to river talwegs, relationship to
impervious formations and attitude, are the genetic factors incor­
porated, and all these are essentially structural in the sense adopted
here (see overleaf).
208 Karst

T able 7 Hydrogeologie systems in low to moderate karst relief

(White 1969)

I. DIFFUSE FLOW
In impure limestones and coarse dolomites. Little surface karst.
Random small solution cavities along joints and bedding planes
but well connected. Darcy’s Law applies. Marked watertable.
Large number of small springs. Deep flow.
II. FREE FLOW a. DEEP 1. OPEN
Thick, massive Karst reaches Karst reaches surface.
soluble rocks. below river Intakes through many
Well integrated valleys. dolines and stream-
branchwork sinks. Much clastic
caves. Large load. Main waterfilled
turbulent flows caves just below river
under gravity at level. Abandoned
low gradients. A tubular cave fragments
few big springs. above, sediment-
choked.
2 CAPPED
.

Intake from edge of

impervious cap down
shafts. Large nearly
horizontal caves.
PERCHED 1 OPEN
Impervious base­ Karst reaches surface.
ment above river Intake through many
valleys. Shallow dolines. Much clastic
flow paths. Cave load. Short cave
streams often segments.
free-surface, small CAPPED
water storage. Vertical shafts round
edge of cap and lateral
inflow. Long integrated
caves.
III. CONFINED a. ARTESIAN
FLOW Under inclined cap. hydrostatic pressure
Flow restricted forces water deep. Slow flow. 3-D network
by beds. caves in joints, inclined down beds.
b. SANDWICH
Thin karst rocks between perching and
capping impervious rocks. Horizontal 2-D
network caves in joints. Small diffuse
recharge from above. Backflooding from
rivers.
 In flu e n c e o f G e o lo g ic a l S tru c tu re 209

Recharge
R iver

D iffu s e re cha rg e
River

S h a ft S haft
G ravity S tre a m -
sp rin g / sink
River

R ec h a r g e

I D if fu s e recharge J

R iv e r

R e c h a r g e a t h ig h r i v e r s t a g e

I—I 1I C a rb o n a t e
1‘ 11 aqu ife r Sandstone S h a le

Ig n e o u s r o c k -------------W a t e r t a b l e

Types o f hydrogeologic systems in lo w to m oderate karst re lie f accord­

in g to W hite 1969
 XI

HISTORICAL GEOMORPHOLOGY OF
KARST

THE KARST CYCLE

Soon after W. M. Davis’s ‘cycle of erosion’ made impact in

Europe, theoretical conceptions of a karst cycle of erosion
appeared from those of E. Richter in 1907 onwards.
The most universally based scheme is that of Grund (1914;
Fig. 66), which may be summarised as follows. As soon as the
land is raised above the watertable, dolines develop at points
favouring solution. In youth these are irregularly scattered over
the initial surface and most European karst is at this stage. Youth
ends when dolines increase in numbers and size so much as to
destroy all vestiges of the initial surface and are separated by
ridges only. Then the more favoured dolines expand at the
expense of neighbours, eliminating intervening ridges. Uvalas
develop in this way. Flat floors appear in the depressions and

66 Karst cycle according to Grund 1914

210
 Historical Geomorphology 211

cone-shaped hills result from ridge destruction. This is maturity

and Grund mistakenly thought that this concept corresponded to
the cockpit karst of Jamaica and Java. It is in fact much more
like some high mountain karst in New Guinea (Jennings and Bik
1962).
The whole surface is progressively lowered until the depression
floors reach the watertable when gentle streams run across them.
At this old age stage there is also much cave collapse; gorges and
collapse dolines result. Corrosion plains with scattered residual
hills are the end point. Poljes are regarded as extraneous tectonic
features which can introduce old age forms at younger stages.
Though later work has sustained Grund’s setting apart of uvalas
and poljes, his scheme is vitiated by the amalgamation of tem­
perate and tropical forms into one framework with no regard to
climatic factors.
Cvijic (1918) restricted himself to features found in the
Dinaric karst in his scheme and he also assumed the simplifying
conditions of a thick limestone mass sandwiched between
impervious formations but lying entirely above sea level (Fig.
67). As the impervious cover is stripped off, the limestone
inherits a normal surface drainage and a valley system develops.
Then follows progressive loss of surface streams as solution opens
up ways for infiltration and engulfment. Dolines begin to form and
tectonic poljes may be present. Eventually all surface drainage
has been lost at maximal karst development, dolines occurring on
interfluves as well as along the valleys. Elaborate cave systems
feed risings around the karst margins and around the poljes,
which may, however, be punctured by uvalas now.
Thereafter the karst is destroyed. Normal valleys reappear
through steepheads recessing the margins, through cave collapse
and through floods planing the polje floors afresh at lower levels,
leaving hums here and there. Surface erosion breaches dolines and
uvalas, though poljes continue to develop. Finally the impervious
basement is widely exposed and only scattered hums remain of
karst relief.
Cvijic’s scheme can also be criticised. For instance, fluvial
drainage may never attack the top surface of the limestone since
subsidence and collapse dolines may develop in the overlying
rocks before it is exposed. However, it does not claim univer­
sality, though Cvijic, like Davis, suffered from his disciples’ mis-
212 Karst

67 Karst cycle in thick limestone between impervious formations and

above sea level according to Cvijic 1918

representing his ideas (e.g. Sanders 1921). It retains value as a

theoretical construct against which to match regional instances,
and furnishes the basis for a restatement by Geze (1965) with
the speleological aspects elaborated.
 Historical Geomorphology 213

CLIMATIC CHANGE

The recognition that karst development may have a number of

separate rationales in different climates led investigators to hypo­
thesise evolutionary schemes for particular climato-morphogenic
systems, e.g. Lehmann’s interpretation of the south-central
Javanese karst (p. 187).
However, as Jennings and Sweeting (1963) found in the
tropical semiarid karst of the Limestone Ranges of West
Kimberley, the problem is complicated by climatic changes. Is the
sequence of relief forms evident in these ranges an active product
of the present alternation of a short but intensely rainy wet season
and a long dry season? Or is it the result of historic events, a
tropical humid climate followed by a tropical semiarid or even
arid one (Jennings 1969)?
Similar problems arise in better investigated areas such as the
Dinaric karst itself. In certain parts, e.g. in the Beljanica Moun­
tains (Gavrilovic 1969), hums are set so close as to resemble
tropical conekarst. Lateral solutional undercutting has been pro­
posed as the cause of mid-latitude poljes as it has for corrosional
plains of tropical humid karst, though minor related features such
as tropical cliff-foot caves are lacking in Yugoslavia. It has there­
fore been suggested that these features are relict from a Pliocene
subtropical humid climate (Roglic 1954; Rathjens 1954). Further,
Melik (1955) considers that later cold periods in the Pleistocene
had the effect of halting or slowing down polje development so
that the Tertiary forms survived. In these periods slopes became
more active; frost wedging increased, solifluction and landslips
moved masses of regolith downslope, and rivers alluviated polje
floors more abundantly. Thick layers of gravel, sands, and clays
accumulated and some of the underground passages were
obstructed. Some highlying poljes are filled to the level of the
lowest col and there has been no extension of their floors since
this event (Rathjens 1960). According to Roglic (1964b),
aggradation of glacifluvial, as well as periglacial fluvial deposits
has led to the enlargement, even the creation of polje floors in
many Yugoslav instances and in a few the associated blocking of
ponors led to Pleistocene lake clay accumulation. Gams (1969),
however, maintains that the effects of the cold periods were not
solely constructional; increased amounts of runoff from melting
214 Karst

snow and ice, accentuated by frozen ground, were directed against

marginal limestones and poljes enlarged in this way also.
Cold climates intervene in other ways in karst development. The
limestone pavements of Craven, Western Ireland and the Alps
belong here (Williams 1966a). These horizontal or sloping plat­
forms of bare karst are due to exposure of well bedded, hard
limestone to strong Pleistocene glacial erosion. That erosion has
removed weathering mantles and much of the preglacial solutional
relief. Most of the minor solution sculpture found on them has
developed since the ice retreated. Some variation in lithology
seems necessary; more resistant, pure and compact beds form the
pavements whereas weaker, impure or poorly cemented beds or
shale interbeds are found at the foot of the risers between the
platforms. Apart from large, level platforms where both beds and
ground are horizontal, pavements occur on moderate slopes
where they may be inclined pavements if dip and slope coincide
or stepped pavements (Schichttreppenkarst of Bögli 1964) if the
dip of the beds is with the slope but less than it. Past periglacial
phases also leave other effects in the present landscape of tem­
perate karst than their influence on polje development as Tricart
(1955) showed in respect of the Causses of France.
Much more extensive survival of well developed tower karst
from Tertiary times, in this case lower Tertiary, is claimed for
South China by Silar (1965). Geliert (1962) provides strati­
graphic evidence for the age in the form of Mio-Pliocene lavas
and sediments overlying some of the associated plains. In the
higher and more northerly tower karsts of the Yunnan and
Kweichow plateaux, inheritance from the warmer Tertiary climate
gives an apparent conflict between the nature of the karst and the
present climate.
Another contrast leading to inference of inheritance is the
presence of inactive cave systems in semiarid and arid karsts as
in the Nullarbor Plain in Australia (Jennings 1967c). They are
thought to have formed during phases of greater effective, if not
absolute precipitation in Pleistocene cold periods. The time of
formation of Carlsbad Caverns and other caves in the semiarid
Guadalupe Range of New Mexico has been cast farther back still
into the wetter Pliocene by Bretz (1949) and Horbcrg (1949).
They argue that the caves are unrelated to present topography and
formed by deep phreatic action beneath a summit planation
 Historical Geomorphology 215

surface, subsequently veneered by Pliocene gravels. Contrarily,

Moore (1960) recognises three levels of shallow phreatic develop­
ment in Carlsbad Caverns which would indicate that at least most
of the development took place after uplift and dissection of the
summit surface, probably in the Late Pliocene and Early Pleisto­
cene, when, however, conditions were still wetter than at present.
There are coarse fluvial sediments present from a late stage in
the cave evacuation which are quite incompatible with present
climate.
In caves in the Transvaal, alternation of deposition of red
sands and formation of speleothems has been related to drier and
wetter phases in the Pleistocene (King 1951). The red sands are
partly angular grains of local provenance only and partly well
sorted and rounded grains of quartz and also of minerals foreign
to the vicinity. The latter fraction is regarded as an aeolian
component associated with more arid climates than at present
which is one of moderate speleothem precipitation. In other
Transvaal caves, however, periods of speleothem formation, which
alternate with phases of filling with waterborne red earth, are
interpreted as the drier parts of arid-humid climatic oscillations
(Marker and Brook 1970).

TECTONIC MOVEMENTS AND CHANGES OF BASE LEVELS

Not only do karsts reflect episodes of climatic history, they often

register tectonic events and changes of base level more com­
pletely than do other terrains: the underground partakes of the
effects of these vicissitudes as well as the surface. This is illus­
trated by the 300 m fault scarp on the western side of Mt Hoyo
in the eastern Congo, which is penetrated by caves at various
levels (Ollier and Harrop 1963). Although opening from the
face of this high cliff, the caves are phreatic in nature, though
they have subsequently been modified vadosely, especially near
their entrances. The phreatic phase is incongruous with their
present position and they must antedate early Pleistocene fault­
ing. Cave survival through this tectonic phase is remarkable.
General uplift leads to rejuvenation which can be recorded both
below as well as above ground as Sweeting (1950) has shown
for Craven. Here the Carboniferous Limestone and part of the
overlying, mainly impervious Yoredale Beds are truncated by an
216 K a rs t

Yoredale Beds D ry v a lle y

P la te a u Stage

D ales S ta g e
— ~r"f C a rb o n ife ro u s r
—r L im e s to n e ,— L-
M a s te r
P re se n t le v e l o f saturation^

\ \ S ^ 1—
P re -C a rb o n ife ro u s im pe rm e ab le
' b a se m e n t

K ilo m e tre

G a te k irk
I Cave

Haws GilL W heel'

P h ilp in Hole

»^W e a th e rcote Cave

A 1 I frj
H u rtle Pot
I
I J& £)
£

P h ilp in H ole G a te k irk Cave

W e a th e rco te Cave Haws G ill
W heel
H u rtle Pot
Gods B ridg e

68 (a) Scheme o f karst e vo lu tio n in Craven, E ngland. A fte r Sweeting.

(b) C ave developm ent at rejuve n a tio n head on R iv e r G reta, C raven,
between Plateau and Dales Stages o f valley developm ent. A fte r
W a rw ick 1960.

I
 Historical Geomorphology 217

erosion surface at about 370 m (Fig. 68). Close beneath this

surface horizontal caves formed and valleys graded to it. In a
first rejuvenation the valleys in limestone became dry and under­
ground drainage descended to greater depths, deep potholes being
formed and vertical pitches added to earlier cave systems. Then
in a stable base level stage, the Dales Stage, straths were fashioned
along many valleys 70-100 m lower down and so also horizontal
cave passages at about 280 m, feeding surface streams at the
strath level. A second rejuvenation ensued with further incision
of the major rivers to grade to the Craven Lowlands, leaving more
tributary valleys hanging and dry. Further vertical development
took place in the caves and passages at the 280 m level were
enlarged into large chambers by collapse after emptying of water.
A final Stillstand has caused horizontal ‘master’ caves, largely
waterfilled, at present rest levels at about 200 m.
Not all caves in the area are old and complex in their history.
Residual hillmasses of the Yoredale Beds continue to recede
through surface erosion, and, as their margins retreat, fresh pot­
holes develop at streamsinks on the still extending 370 m
surface; some streamsinks are of Holocene age.
Warwick (1960, 1962) has stressed the importance of the
effects of successive phases of rejuvenation on British karsts
generally. In particular, rapid lowering of water levels under­
ground near to rejuvenation heads in limestone valleys leads to
sinking of rivers and cave formation (Fig. 68). This is particu­
larly liable to happen where nickpoints pass up incised meanders,
short cuts developing through meander cores in these circum­
stances (cf. p. 102).
Denudation chronology in karst is likely to have a more precise
record because of the perfection of corrosion plains (PI. 19)
compared with those formed in other ways and because of the
degree of immunity from valley erosion which they possess
through partial deflection of water underground.
The relict karst margin plains of the Dinaric karst have long
attracted attention (Krebs 1929; Cholley and Chabot 1930;
Kayser 1934) because of their high degree of planation evident
after dissection and the sharp breaks they make with the high
plateaux. These plains occur both along the inland margins of the
karst and along the Adriatic coast, e.g. the Karlovac-Sluin,
Kistanje, and Cetina plains. Rivers, such as the Krka and Cikola
218 Karst

in the case of the Kistanje plain, meander in trenches through the

surfaces. Hums rise from their surfaces and dolines occur
especially near to the river gorges. Morawetz (1967) compares
these plains at elevations of 200-300 m with the corrosion plain
of the lower Neretva valley (p. 143). He argues that higher
surfaces were fashioned in a similar way prior to a negative
movement of base level which raised them high above water rest
levels and began their protracted slow destruction.
Accordance of summit levels in tropical conekarsts has been
interpreted as evidence of planation surfaces from which the fields
of residuals have been carved. For example near Baisha southeast
of Kweilin in southern China, limestone towers betray a storeyed
relief with three surfaces represented by accordance at different
levels. Geliert (1962) considered there had been three corrosion
plain stages separated by uplift and rejuvenation (Fig. 69). High
up in individual towers there are cliff-foot caves and through-
caves also witnessing this storeying.

69 Storeying in tower karst, Baislia, Kweilin, China. After Geliert 1962.

Just as the interpretation of summit accordances and upland

plains in impermeable rocks as remnants of former erosion
surfaces partially destroyed by rejuvenation has been subject to
latter-day criticism in favour of other explanations, for example
in favour of dynamic equilibrium, so also have special objections
of similar purpose arisen within the realm of karst. Thus for the
Jura, Aubert (1969) maintains that doline deepening is eventually
stultified by the accumulation of insoluble fines within them and
erosion is then directed laterally. In this way relief is flattened,
especially because the more open jointing along anticlinal axes
favours the reduction of tectonic relief. It is his view that the
high plateaux of the Jura, previously interpreted as relics of a
Tertiary peneplain, are in fact relief conformal to tectonics but
attenuated, if not completely planed, by karst surface corrosion
 Historical Geomorphology 219

which has operated and still is operating laterally rather than

vertically.
Whether some of the rejuvenations discussed above are the
product of tectonic uplift or of large eustatic shifts of sea level
itself enters a debate not central to the theme of this book (see
Bird 1968 and Twidale 1971). Nevertheless there is no doubt
that glacioeustatic sea level changes have affected karsts, most
obviously through the alternating exposure and drowning of the
continental shelves. Where there are surface streams in coastal
karst, they deepen their valleys during glacial low sea levels and
these are drowned in the interglacials, finally by the Flandrian
transgression in the Holocene. In the Dinaric karst this is evident
in the Gulf of Sibenik which consists of the drowned lower valleys
of the Krka and Cikola. The Kistanje erosion surface extends over
some of the Adriatic islands and some of the sea channels between
them, the kanali, are probably drowned Pleistocene valleys. How­
ever, little has been done to see whether such drowned valleys
possess special karst characteristics. Certain of the ‘calanques’ of
the limestone coast of Provence are coastal inlets of a gorge-like
nature and Chardonnet (1948) explains them as due to the
formation by freshwater solution of caves reaching below sea level
followed by roof collapse and eventual destruction of the seaward
barrier to allow drowning without any change in sea level. How­
ever, no case for rejecting the alternative explanation of Flandrian
flooding of collapsed caves formed during low sea level times is
given.
Some submarine springs are a result of deepening of under­
ground circulation during glacial low sea levels and subsequent
submergence (p. 79). Other submarine features of which the
same may be said are the large closed depressions on the Florida
shelf (G. F. Jordan 1954). The ones surveyed vary between 40
and 160 m in depth and between 900 and 2200 m across; their
tops are at —300 to —500 m and lie on a fault block above a
fault scarp. Despite these depths and the unstable tectonic context,
it is likely that eustatic low stands of sea level have contributed
to their formation. R. H. Jordan (1950) previously argued that
vadose action during Pleistocene low sea levels had worked in
conjunction with deep phreatic circulation to fashion the deep,
water-filled dolines of the Florida peninsula itself.
Well known because of its spectacular landscape, the archi-
220 Karst

39 Drowned tower karst in Langkawi Islands, Kedah, West Malaysia

pelago of Vung Ha Long Bay in north Vietnam can be described

as a tower karst sitting in a shallow sea. The same is true of parts
of the Langkawi Islands of Malaya (PI. 39). Some of the Vung
Ha Long islands enclose lakes of sea water reached by boats
through caves. Though active marine notches undercut these
towers now, it is doubtful whether the relief as a whole has been
produced at present sea level. Marine processes can produce only
a few stacks in front of a shoreline retreating from marine attack
but cannot fashion such extensive archipelagoes on their own.
Some of the Vung Ha Long towers have bases at —20 m. There­
fore change of base level is involved. Silar (1965) postulates
tectonic downwarping (followed by uplift because of emerged
marine terraces on the nearby mainland shore and of caves
in the towers at +14-19 m ). However, adequate positive and
negative movements due to glacioeustatism are known to have
occurred which could explain all the phenomena without tectonic
deformation.
Reference has been made to the effects of glacioeustatic shifts
on cave history in Bermuda, ending in the drowning of some
caves and their speleothems. Caves above sea level sometimes
reveal effects of alternating sea level. In Auckland, New Zealand,
 Historical Geomorphology 221

Kairimu Cave lies along a shallow syncline in Lower Oligocene

limestone which was uplifted in Late Miocene times (Barrett
1963). Its oldest passages were then initiated phreatically and
subsequently much enlarged by vadose action. High sea level in
Late Pliocene or Early Pleistocene time flooded the cave which
opens at + 1 0 0 m, and deposited silt and clay with a dating micro­
flora. Some small passages have developed since. As this part of
New Zealand is relatively stable tectonically, the stage of marine
drowning of this cave could be of eustatic origin only.
A miniature relevant example is provided by Narrengullen Cave
well inland in southern New South Wales which has been affected
by water level oscillations in Burrinjuck Reservoir, an artificial
storage which laps to its exit in high stages. This has caused
alluviation of the lower end of the cave. Big flows and sediment
loads tend to correspond with high lake levels; reduced river
flows during low lake stages cannot readily remove the intervening
deposition.

COMPLEX EVOLUTION

The artificiality of separate discussion in this and previous chapters

of the operations of individual factors differentiating karst has
been made patent by inevitable cross-references. Most karst evolu­
tion is complex, both on the surface and underground.
In Clare, Ireland, Visean limestone underlies Namurian sand­
stone and shales, which mainly form gentle but higher land
draining on to the karst. The shales were breached in the late
Tertiary when large depressions were initiated (Williams 1970).
Though they developed further in warm Pleistocene intervals, their
catchments were gradually removed by glacial erosional stripping
of the shales and they ceased to develop, though one, the Carran
depression, had become a polje (Sweeting 1953). Bevelled spurs,
hill-top summits and river nickpoints aggregate into a series of
levels, which Sweeting (1955) regarded as late Tertiary erosion
surfaces, but Williams (1970) maintains are largely due to struc­
tural guidance of Pleistocene glacial erosion. Some closed depres­
sions are simply hollows in the surface of glacial till on top of
the limestone. Glacial erosion produced limestone pavements,
destroying previous minor solution sculpture so that many grikes
and other small forms now present are largely postglacial.
222 Karst

Despite this polygenetic history, Ollier and Tratman (1969)

find the evolution of the caves of a substantial part of the area
to be a simple one confined to the Postglacial. These young caves
are active branchwork caves at shallow depth, dominantly due
to vadose solution with sharp fluctuations of cave river levels.
Only a very few are conceded to have a longer history. The caves
follow the lines of dry valleys, which have been cut in the Post­
glacial with removal of moraine along them. The ice probably
retreated about 15,000 B.P. from the area so that the caves have
a short and simple history in contradistinction to the surface
landforms.
This is a valuable illustration of the important principle that
surface and underground forms can get quite out of step in their
development. The converse relationship to the one just described
may be even more common, namely that cave systems can develop
over long periods of time without having much effect on surface
topography as Aubert (1969) maintains with respect to much of
the Jura Mountains, where, for instance, collapse dolines are very
rare.
The Mendip Hills in southwest England are complex in their
whole karst evolution (Ford and Stanton 1968). Valleys cut in
the Carboniferous Limestone were buried by Triassic conglomerate
and deposits of the same age filled some caves. Subsequently the
valleys have been partially exhumed so that there are Triassic
elements in the landscape. The Mendip plateau surface is due to
Pliocene erosion with drainage oriented by Miocene doming,
though parts of a Jurassic unconformity survive through coincid­
ence with the Pliocene level. Valleys were incised into this surface
and erosional benches are recognised in them. These events are
also registered in the development of the caves which accompanied
stream sinking. The valleys are now dry except on impervious
inlier monadnocks and below the major risings. Cutting of the
valleys is largely attributed to permafrost causing resumptions of
surface drainage during the Pleistocene cold periods. Cheddar
Gorge is due entirely to surface incision and Ebbor Gorge is due
to the destruction of a cave by surface incision during a cold
period. In the cold periods the caves were chiefly subject to calcite
deposition and cave enlargements took place during the intervening
interglacials when there was no ground ice to minimise under­
ground circulation. However, these phases did not lead to stoping
 Historical Geomorphology 223

to the surface and dolines along the dry valleys are the expression
of surface solution where gradients are gentle. Large shallow basins
at the heads of former surface drainage were ponds during the
periglacial periods, when they collected fills of clay, partly wind-
borne. Thus from lower Pleistocene times onward some of the
caves have been affected by alternations of rejuvenation and Still­
stand, and by the impact of successive phases of cold and of
temperate climate in inextricable genetic association with the con­
temporaneously evolving surface of the karst.
In the light of such contrasts in evolution between two com­
paratively well studied karsts, both Carboniferous Limestone
plateaux with Hercynian tectonic histories and subject to similar
climate today, present knowledge may be inadequate for defining
systematic karst types on the total basis of structure, climato-
geomorphic processes and historical evolution, although Birot
(1954) has attempted this for the Alpine and Hercynian belts of
Europe.
 XII

PRESENT STATE OF KARST

GEOMORPHOLOGY AND ITS VALUE

Present knowledge of karst is microcosmic of the whole discipline

of geomorphology. Because it includes some of the most difficult
terrain for movement there is, even qualitative description is still
incomplete for many areas of karst, particularly within the tropics
(PI. 35). But these frontiers are being pushed back now with
modern aids, though the geomorphologist may be a ‘poor relation’
riding oil company helicopters. Techniques and equipment for the
exploration of practically any kind and size of cave have been
elaborated, and diving methods have been modified to permit a
degree of direct examination of waterfilled passages; yet many
karsts are speleologically almost uninvestigated. Photogrammetry
presents special difficulties in karst, requiring more ground checking
than in many other landscapes. Sharp variation in rainforest height
over conekarst points to the need for sensors other than black-
and-white and ordinary colour film for mapping here.
Complexity in much karst relief, with its apparently chaotic
nature, has perhaps daunted application of modern morphometric
methods in this field. However, recent endeavours along these
lines are bringing karst gcomorphology more into step with other
aspects of the discipline with promising results.
Karst is normally included in texts devoted to structural gcomor­
phology since it is fundamentally distinguished from other topo­
graphy by the effects of abnormal rock solubility. Nevertheless
there has been a tendency to take the nature of the rocks for
granted. Recent signs of pursuit of finer lithologic variations in
their karst effects keep this branch abreast of similar indications
elsewhere in gcomorphology.
The equal need to approach karst in terms of the role of different
224
 Present State of Karst Geomorphology 225

climates in conditioning its development in distinct morphogenic

systems was exemplified early in the impetus towards climatic
geomorphology between the two world wars. At least the con­
trasting morphologic regimes of periglacial, hot arid, and hot
humid zones with their attendant suites of karst landforms stand
forth from a proliferation of proposed climato-morphogenic realms
in karst. The readiness with which karst students turned to the
climatic approach went with an enduring interest in karst pro­
cesses; there never was on this side of the discipline quite as much
neglect of process as has been associated with the Davisian school
of geomorphology. In this connection it would be unjust not to
recollect the stimulating attention that Davis himself gave to the
analysis of different hydrodynamic circumstances in underground
geomorphology. In recent decades the chemical aspects of karst
processes, including rates of chemical denudation, have attracted
a great deal of effort and practitioners of karst geomorphology
can feel reasonably satisfied with their contribution to the surge
of activity in quantitative geomorphic process study. That this has
led in the last few years to increasing involvement of physicists
and chemists in what are essentially geomorphic problems is to
be welcomed (Ashton 1966; Roques 1969), even though some
matters thereby become insulated from criticism by more generally
trained geomorphologists. Glaciology has had this experience.
All these approaches must be brought together in order to
reconstruct the historical evolution of individual karsts. In this
task there are both special advantages such as the perfection of
corrosion plains and special obstacles such as the difficulty
frequently met in dating stages of cave development. Many more
results must come forth from the different approaches— morpho­
metry, process measurement, correlative sediment study— before it
can be resolved whether, in this final stage of synthesis, karst study
must remain a nomothetic rather than an ideographic science.
‘Hydrogeologie uniqueness’ is not infrequently invoked in respect
of individual limestone areas (Stringfield and LeGrand 1969a). At
the moment the essential point is that karst investigators are parti­
cipating in all the growing points of geomorphology, if to differing
degrees.

This book has been concerned with the purely scientific objectives
of karst investigation; it is proper to conclude with a brief indica­
tion of its practical advantages.
226 Karst

Questions of drinking water supply have long prompted much

activity. Martel’s name stands out in this connection, especially
with regard to danger of pollution arising from the large conduits
and rapid transit common in karst groundwater. In arid and semi-
arid zones, karst rocks assume very great importance as aquifers
and this has led to much work by FAO on karst hydrology in
the environs of the Mediterranean. Not only surface karst is
significant in this respect; buried karsts are important aquifers in
the United States (Stringfield and LeGrand 1969a). Many karst
springs have mineral contents of alleged curative value and give
rise to spas.
In times past, karst springs have been used to generate mechani­
cal power, but more recently knowledge of underground drainage
has helped in the assessment and manipulation of hydroelectric
potential as in the Gouffre Pierre St Martin in the Pyrenees.
Dubrovnik in the Dinaric karst has an electricity supply from
springs within the city limits. In the neighbourhood of Antalya
in southern Turkey the karst drainage system has required deflec­
tion for irrigation needs. Understanding of the functioning of poljes
in Yugoslavia has permitted engineers to diminish flooding of
their cultivable floors, and in Florida and Georgia wells are put
down into limestone to get rid of surface drainage. The handling
of water resources remains the prime economic spark for karst
research.
Mining of phosphate and nitrates from caves is more of historic
than present commercial interest. However, karst cavities of past
geological periods have been lodging points for other minerals
such as alluvial tin, sulphide and other ores of copper, and they
have also acted as oil reservoirs. Search for and exploitation of
these resources can be pursued more successfully in the light of
improved knowledge of karst in all its manifestations.
Engineering in karst country takes risks if its peculiarities are
disregarded. The May River dam near Konya in central Turkey
is a costly construction unlikely ever to perform its purpose
because its catchment, floored with limestone, behaves like a
sieve (PI. 40). There are also problems of dam stability to be
reckoned with on limestone foundations. Lowering of water levels
in dolomites by pumping from gold mines in the Far West Rand
in Transvaal, South Africa, has caused accelerated subsidence of
deep soils and regolith into cavities (Brink and Partridge 1965).
 Present State of Karst G eomorphology 227

40 Subsidence dotine in alluvium over limestone in May River dam near

Konya, Turkey. Formed since dam was constructed.

This had led to considerable loss with the extreme case of the
disaster at West Dreifontein in 1962 when twenty-nine workers
lost their lives in sudden subsidence which engulfed a crusher
plant. Nowadays in parts of the United States, geophysical sounding
is employed over proposed highway routes through cavernous
terrain so that thin roofs can be avoided. Better understanding
can lead to better management of land.
Caves have long been used as refuges, particularly by outcasts
and escapists from society. In Europe during World War II, people
fled to caves as a protection from air raids, and in New Guinea
evidences of recent cave occupation are ascribed by the New
Guineans to the ‘taim bilong big pait’. Since that war, consideration
has been given in some countries to the potential of caves as
refuges for large numbers in times of nuclear warfare. They can
also provide cheap storage— in peace as with Roquefort cheese
in the Causses of France and in war as with Nazi petrol and oil
reserves in Postojna Cave in Slovenia until partisans coming by
unguarded distant entrances set fire to the military dumps. Whether
cave air has any real therapeutic value for illnesses such as
228 Karst

asthma may be uncertain, yet in Hungary and Turkey there are

caves employed in this way.
The tourist attraction of well decorated or otherwise impressive
caves is a resource in process of wider realisation than ever before
as societies become more affluent and better educated. The same
is true of spectacular surface features such as natural bridges, tufa
dams and gorges which often accompany them. There are great
conservation problems arising, with tourism the largest single and
fastest growing industry of the world. For example, the placing
of hotels requires the greatest of care lest it involve partial ruin
of their raison d’etre. This has happened at various places such as
on the magnificent wall of rimstone dams at Pamukkale in western
Turkey. When a forest is destroyed, there is still the chance of
natural regrowth under protection in a century or two; when the
curiosities of the mineral kingdom which decorate caves in such
fantastic and beautiful fashion are broken, on the human timescale
they are lost forever. In this matter, however, tasteless commer­
cialisation is a less widespread danger than the enthusiastic but
often careless behaviour of cavers in unprotected caves.
Financial support from governments and private enterprises
for karst investigation and cave exploration is naturally most forth­
coming in countries with extensive karst areas such as France.
However, karst has such a fascination that even where this kind
of support is not available, geomorphologists, whether amateur or
professional, will surely find opportunity to contribute their part
to the understanding of it. Its economic and social values will be
not less for being a by-product of scientific curiosity and sporting
endeavour.
 BIBLIOGRAPHY

Adams, C. S. and Swinnerton, A. C. 1937. The solubility of calcium

carbonate. Trans. Am. geophys. Un., 11:504-8.
Ashton, K. 1966. The analysis of flow data from karst drainage systems.
Trans. Cave Res. Grp Gt Br., 7:161-203.
Aub, C. F. (in press a). Some observations on the karst morphology of
Jamaica. Congr. 5Cint. Speleol.
-------- (in press b). The nature of cockpits and other depressions in the
karst of Jamaica. Congr. 5Cint. Speleol.
Aubert, D. 1966. Structure, activite et evolution d’une doline. Bull. Soc.
Neuchateloise Sei. Nat., 89:113-20.
-------- 1969. Phenomenes et formes du karst jurassien. Eclogue geol. Helv.,
62:325-99.
Baläzs, D. 1968. Karst regions in Indonesia. Karszt-es Barlangkutatds,
1963-67:3-61.
Balch, E. S. 1900. Glacieres or Freezing Caverns. Philadelphia.
Balchin, W. G. V. and Lewis, W. V. 1938. The chalk water table south-east
of Cambridge. Pp. 20-4 in The Cambridge Region, ed. H. C. Darby.
Cambridge.
Barrett, P. J. 1963. The development of Kairimu Cave, Marakopa District,
South-West Auckland. N.Z. Jl Geol. Geophys., 6:288-98.
Bauer, F. 1964. Kalkabtragungsmessungen in den Österreichischen Kalk-
hochalpen. Erdkunde, 18:95-102.
Bird, E. C. F. 1968. Coasts. Canberra.
Birot, P. 1949. Essai sur quelques problemes de morphologie generale.
Lisbon.
-------- 1954. Problemes de morphologie karstique. Annls Geogr., 63:161-
92.
Blanc, A. 1958. Repertoire bibliographique critique des etudes de relief
karstique en Yougoslavie depuis Jovan Cvijic. Mem. Docums. Cent.
Docum cartogr. geogr., 7:135-228.
Bögli, A. 1956. Grundformen von Karsthöhlenquerschnitten. Stalactite,
6:56-62.
-------- 1960. Kalklösung and Karrenbildung. Z. Geomorph.-Suppl., 2:4-21.
-------- 1961a. Karrentische, ein Beitrag zur Karstmorphologie. Z. Geo-
morph., 5:185-93.
-------- 1961b. Der Höhlenlehm. Mem. Rass. speleol. ital., 5(2): 11-29.
-------- 1964a. Mischungskorrosion—ein Beitrag zur Verkarstungsproblem.
Erdkunde, 18:83-92.
-------- 1964b. Le Schichttreppenkarst. Revue Beige Geogr., 88:63-82.
-------- 1969. Neue Anschauungen über die Rolle von Schichtfugen und
Klüften in der karsthydrographischen Entwicklung. Geol. Rdsch.,
58:395-408.
229
230 Bibliography

Bretz, J. H. 1942. Vadose and phreatic features of limestone caverns.

J. Geol., 50:675-811.
--------- 1949. Carlsbad Caverns and other caves of the Guadalupe Block,
New Mexico. /. Geol., 57:447-63.
--------- 1960. Bermuda: a partially drowned, late mature Pleistocene karst.
Bull. geol. Soc. Am., 71:1729-54.
Brink, A. B. A. and Partridge, T. C. 1965. Transvaal Karst: some
considerations of development and morphology, with special refer­
ence to sinkholes and subsidence on the Far West Rand. S. Afr.
geogr. J., 47:11-34.
Brown, E. H. 1969. Jointing, aspect and orientation of scarp-face dry
valleys, near Ivinghoe, Buckinghamshire. Trans.lnst.Brit.Geogr.,
48:61-74.
Burdon, D. J. and Safadi, C. 1963. Ras-el-ain: the great karst spring of
Mesopotamia. Jnl Hydrol., 1:58-95.
Burke, A. R. and Bird, P. F. 1966. A new mechanism for the formation of
vertical shafts in Carboniferous Limestone. Nature, Loud., 210:
831-2.
Burns, K. L. 1964. Geological Survey Explanatory Report Devonport.
Tasmania Dept. Mines, Hobart.
Butcher, A. L. and Railton, C. L. 1966. Cave surveying. Trans.CaveRes.
Grp Gt Br., 8(2): 1-37.
Caumartin, V. and Renault, P. 1958. La corrosion biochimique dans un
reseau karstique et la genese du mondmilch. Notes biospeol., 13:87-
109.
Chandler, R. H. 1909. On some dry chalk valley features. Geol. Mag.,
6:538-9.
Chardonnet, J. 1948. Les calanques provengales, origine et divers types.
A mils Geogr., 57:289-97.
Chevalier, P. 1944. Distinction morphologique entre deux types d’erosion
souterraine. Revue Geogr. alp., 32:475-92.
Cholley, A. and Chabot, G. 1930. Notes de morphologie karstique. Annls
Geogr., 39:270-85.
Cleland, H. F., 1910. North American natural bridges, with a discussion of
their origin. Bull. geol. Soc. Am., 21:313-38.
Coleman, A. M. and Balchin, W. G. V. 1959. The origin and development
of surface depressions in the Mendip Hills. Proc. Geol. Ass., 70:
291-309.
Coleman, J. C. 1945. An indicator of waterflow in caves. Geol. Mag.,
82:138-9.
Collingridge, B. R., 1969. Geomorphology of the area. Pp. 42-58 in The
Caves of Northwest Clare, Ireland, ed. E. K. Tratman. Newton
Abbott.
Conrad, G., Geze, B., and Paloc, H. 1968. Phenomenes karstiques et
pseudokarstiques du Sahara. Congr. 4' int. Speleol., 3:411-16.
Corbel, J. 1952. A comparison between the karst of the Mediterranean
region and of northwestern Europe. Trans. Cave Res. Grp Gt Br.,
2:1-26.
--------- 1954. Les phenomenes karstiques en climat froid. Erdkunde, 8:
119-20.
--------- 1958. Les karsts du Nord-Ouest de TEurope. Inst. Et. Rhod. Mem.
et Doc., 12.
--------- 1959a. Erosion en terrain calcaire. Annls Geogr., 68:97-120.
 Bibliography 231

Corbel, J. 1959b. Les karsls du Yucatan et de la Fl'oride. Bull.Ass.Geogr.

France, 282-3:2-14.
--------- 1960. Nouvelles recherches sur les karsts arctiques Scandinaves.
Z.Geomorph.-Suppl., 2:74-80.
Cramer, H. 1933. Höhlenbildung im Karsten. Petermanns geogr. Mitt.,
79:78-80.
--------- 1941. Die Systematik der Karstdolinen. Neues Jb. Miner. Geol.
Paläont., 85:293-382.
Cullingford, C. H. D. 1962. British Caving. 2nd ed. London.
Curl, R. L. 1958. A statistical theory of cave entrance evolution. Bull. natn.
speleol. Soc., 20:9-22.
--------- 1960. Stochastic models of cave development. Bull. natn. speleol.
Soc., 22:66-76.
-------- 1966a. Scallops and flutes. Trans. Cave Res. Grp Gt Br., 7:121-60.
--------- 1966b. Caves as a measure of karst. /. Geol., 74:798-830.
Cvijic, J. 1893. Das Karstphänomen. Geogr. Abh., 5:217-329.
-------- 1918. L’hydrographie souterraine et 1’evolution morphologique du
karst. Revue Geogr. alp., 6:375-426.
--------- 1925. Types morphologiques des terrains calcaires. Le Holokarst.
Le Merokarst. Types karstiques de transition. C.r.hebd. Seanc. Acad.
Sei., Paris, 180:590-4, 757-8, 1038-40.
--------- 1960. La geographic des terrains calcaires. Academie serbe des
Sciences et des Arts monographies, 141.
Danes, J. V. 1908. Geomorphologische Studien im Karstgebiete Jamaikas.
9e Int. geogr. Congr., 2:178-82.
--------- 1910. Die Karstphänomene im Goenoeng Sewoe auf Java. Tijdschr.
K. tied, aardrijk.sk. Genoot, 27:247-60.
--------- 1916. Karststudien in Australien. Sber. K. böhm. Ges. Wiss., 7:1-75.
Davies, J. L. 1969. Landforms of Cold Climates. Canberra.
Davies, W. E. 1949. Features of cave breakdown. Bull. natn. speleol. Soc.,
11:34-5.
--------- 1960. Origin of caves in folded limestone. Bull. natn. speleol. Soc.,
22:5-18.
Davis, W. M. 1930. Origin of limestone caverns. Bull. geol. Soc. Am.,
41:475-628.
Deike, G. H. 1960. Origin and geologic relations of Breathing Cave,
Virginia. Bull. natn. speleol. Soc., 22:30-42.
--------- 1967. The Development of Caverns in the Mammoth Cave Region.
Ph.D. thesis, Pennsylvania State University.
Douglas, I. 1964. Intensity and periodicity in denudation processes with
special reference to the removal of material in solution by rivers.
Z. Geomorph., 8:453-73.
--------- 1965. Calcium and magnesium in karst waters. Helictite, 3:23-36.
--------- 1968. Some hydrologic factors in the denudation of limestone
terrains. Z. Geomorph., 12:241-55.
--------- 1969. Field Methods of Water Hardness Determination. Br. Geo­
morph. Res. Grp Tech. Bull. 1.
Drew, D. P. and Smith, D. I. 1969. Techniques for the Tracing of Sub­
terranean Drainage. Br. Geomorph. Res. Grp Tech. Bull. 2.
Dunham, R. J. 1962. Classification of carbonate rocks according to deposi-
tional texture. Am. Ass. Petrol. Geol., 1:108-21.
Duplessy, J. C., Labeyrie, J., Lalou, C., and Nguyen, H. V. 1970.
Continental climatic variations between 130,000 and 90,000 years
B.P. Nature, Lond., 226:631-3.

I
232 Bibliography

Ek, C. 1961. Conduits souterrains en relation avec les terrasses fluviales.

Annls Soc. geol. Belg., 84:313-40.
-------- 1964. Notes sur les eaux de fonte des glaciers de la Haute
Maurienne. Revue Beige Geogr., 88:127-56.
—------- 1969. L’effet de la loi de Henry sur la dissolution du CCE dans
les eaux naturelles. Pp. 53-5 in Problems of the Karst Denudation,
ed. O. Istelcl. Brno.
-------- Delecour, F., and Weissen, F. 1968. Teneur en C 0 2 de Pair de
quelques grottes beiges. Technique employee et premiers resultats.
Ann. Speleol., 23:243-57.
Emig, W. H. 1917. Travertine Deposits of Oklahoma. Oklahoma Geol.
Surv. Bull. 29.
Fagg, C. C. 1923. The recession of the chalk escarpment. Proc. Trans.
Croydon nat. Hist, scient. Soc., 9:93-112.
-------- 1954. The coombes and embayments of the chalk escarpment.
Proc. Trans. Croydon nat. Hist, scient. Soc., 12:117-31.
Feininger, T. 1969. Pseudokarst on quartz-diorite, Colombia. Z. Geo-
morph., 13:287-96.
Fenelon, P. (ed.) 1968. Vocabulaire fran^ais des phenomenes karstiques.
Mem. Docums. Cent. Docum. cartogr. geogr., 4:193-282.
Flathe, H. and Pfeiffer, D. 1965. Grundzüge der Morphologie, Geologie
und Hydrogeologie im Karstgebiet Guning Sewu, Java (Indonesien).
Geol. Jb., 83:533-62.
Folk, R. L. 1959. Practical petrographic classification of limestones. Bull.
Am. Ass. Petrol. Geol., 43:1-38.
Ford, D. C. 1964. Origin of closed depressions in the central Mendip Hills,
Somerset, England. Paper at Karst Symposium, 20th International
Geographical Congress, England, 1964.
--------- 1965a. Stream potholes as indicators of erosion phases in limestone
caves. Bull. natn. sp leol. Soc., 27:27-32.
-------- 1965b. The origin of limestone caverns: a model from the central
Mendip Hills, England. Bull. natn. speleol. Soc., 27:109-32.
--------- 1966. Calcium carbonate solution in some central Mendip caves,
Somerset. Proc. speleol. Soc. Univ. Bristol, 11:46-53.
-------- and Stanton, W. L. 1968. The geomorphology of the south-central
Mendip Hills. Proc.Geol.Ass., 79:401-28.
Ford, T. D. and King, R. J. 1966. The Golconda Caverns, Brassington,
Derbyshire. Trans.Cave Res.Grp Gt Br., 7:81-114.
-------- and --------- 1969. The origin of the silica sand pockets in the
Derbyshire limestone. Mercian Geologist, 3:51-69.
Frank, R. M. 1971. Cave sediments as palaeoenvironmental indicators,
and the sedimentary sequence in Koonalda Cave. Pp. 94-104 in
Aboriginal Man and his Environment in Australia, ed. D. J.
Mulvaney and J. Golson. Canberra.
Frear, G. and Johnston, J. 1929. Solution of calcium carbonate in aqueous
solutions at 25°C. J. Am . Chem. Soc., 51:2082-93.
Gams, I. 1962. Slepe doline Slovenije v Primerjalni Metodi. Geogr.Zb., 7.
-------- 1963. Die Einfluss der Schichtenlage auf die Richtung der Höhlen­
gänge und auf die Querschnitte in den längsten Höhlen Sloweniens.
Congr. 3e int. Speleol., 2:215-20.
--------- 1965. Types of accelerated corrosion. Pp. 133-9 in Problems of
the Speleological Research, ed. O. Stelcl. Prague.
-------- 1966. Faktorji in dinamika Korozije na karbonatnih Kameninah
Slovenskeya dinarskegain alpskegu Krasa. Geogr. Vestnik, 38:11-68.
 Bibliography 233

Gams, I. 1968. Über die Faktoren, die die Intensität der Sintersedimentation
bestimmen. Congr. 4l int. Speleol., 3:107-15.
--------- 1969. Some morphological characteristics of the Dinaric karst.
Geogrl /., 135:563-72.
Gavrilovic, D. 1969. Kegelkarst-Elemente im Relief des Gebirges Beljanica,
Jugoslavien. Pp. 159-66 in Problems of the Karst Denudation, ed.
O. Stelcl. Brno.
Geliert, J. 1962. Der Tropenkarst in Südchina im Rahmen der Gebirgs-
formung des Landes. Verb. dt. Geogr. Tags., 33:376-84.
Gerstenhauer, A. 1960. Der tropische Kegelkarst in Tabasco (Mexico).
Z. Geomorph.-Suppl., 2:22-48.
--------- 1964. Blatt 3 (Nord-Puerto Rico) des Internationalen Karstatlas.
Erdkunde, 18:148.
Geze, B. 1953. La genese des goulfres. Congr. 1er int. Speleol., 2:11-23.
--------- 1958. Sur quelques caracteres fondamentaux des circulations karst-
iques. Ann. Speleol., 13:5-22.
--------- 1965. La speleologie scientifique. Paris.
Gilevska, S. 1964. Fossil karst in Poland. Erdkunde, 18:124-35.
Glernie, E. A. 1948. Some points relating to Ogof Ffynnon Ddu. Trans.
Cave Res. Grp Gt Br., 1( 1): 13-25.
--------- 1950. Further notes on Ogof Ffynnon Ddu. Trans. Cave Res. Grp
Gt Br., 1(3): 1-47.
--------- 1954a. Artesian flow and cave formation. Trans. Cave Res. Grp
Gt Br., 3:55-71.
--------- 1954b. The origin and development of cave systems in limestone.
Trans. Cave Res. Grp Gt Br., 3:75-83.
--------- 1958. Nameless streams: proposed new terms. Cave Res. Grp Gt
Br. Newsl., 72-77:22-3.
--------- 1963. Flow markings. Cave Res. Grp Gt Br. Newsl., 87:8-9.
Gregory, J. W. 1911. Constructive waterfalls. Scott, geogr. Mag., 27:
537-46.
Groom, G. E. and Coleman, A. 1958. The geomorphology and speleo-
genesis of the Dachstein Caves. Cave Res. Grp Gt Br. Occ. Pub., 2.
--------- and Williams, V. H. 1965. The solution of limestone in South
Wales. Geogrl ] . , 131:37-41.
Gross, M. G. 1964. Variations in the 0 18/ 0 16 and C 13/C 12 ratios of
diagenetically altered limestones in the Bermuda Islands. J. Geol.,
72:170-94.
Grund, A. 1903. Die Karsthydrographie. Geogr. Abh., 7:3.
--------- 1910a. Zur Frage des Grundwassers im Karste. Mitt. geol. Ges.
Wien, 53:606-17.
--------- 1910b. Das Karstphänomen. Geol. Charakterbilder, 3.
--------- 1914. Der geographische Zyklus im Karst. Z. Ges. Erdk. Berl.,
1914:621-40.
Halfday, W. R. 1957. The Origin of the Limestone Caves of the Sierra
Nevada of California. Bull. Western Speleol. Surv. Misc. Ser., 3.
--------- 1960. Changing concepts of speleogenesis. Bull. natn. speleol. Soc.,
22:23-9.
Hancock, P. L. 1968. Joints and faults: the morphological aspects of their
origins. Proc. Geol. Ass., 79:146-51.
Harrison, J. V. 1930. The geology of some salt plugs in Laristan (Southern
Persia). Q.Jl geol. Soc. Land., 86:463-520.
Haserodt, K. 1969. Beobachtungen zur Karstdenudation am Kluftkarren in
234 Bibliography

glazialüberformten alpinen Bereichen. Pp. 123-38 in Problems of the

Karst Denudation, ed. O. Stelcl. Brno.
Hem, J. D. 1959. Study and Interpretation of the Chemical Characteristics
of Natural Water. U.S. geol. Surv. Wat.-Supply Pap., 1473.
Hendy, C. H. 1969. Isotopic studies of speleothems. New Zealand Spei.
Soc. Bull., 4:306-19.
Hodgkin, E. P. 1964. Rate of erosion of intertidal limestone. Z.Geomorph.,
8:385-92.
Holland, H. D., Kirsipu, T. V., Huebner, J. S., and Oxburgh, U. M. 1964.
On some aspects of the chemical evolution of cave waters. /. Geol.,
72:36-67.
Holmes, J. W. and Colville, J. S. 1970a. Grassland hydrology in a karstic
region of southern Australia. Jnl Hydrol., 10:38-58.
-------- a n d ---------- 1970b. Forest hydrology in a karstic region of southern
Australia. Jnl Hydrol., 10:59-74.
Hooper, J. H. D. 1958. Bat erosion as a factor in cave formation. Nature,
Lond., 182:1464.
Horberg, L. 1949. Geomorphic history of the Carlsbad Caverns area, New
Mexico. J. Geol., 57:464-76.
Howard, A. D. 1963. The development of karst features. Bull. natn. speleol.
Soc., 25:45-65.
Hundt, R. 1950. Erdfalltektonik. Halle.
Hunt, G. S. 1970. The origin and development of Mullamullang Cave N 37,
Nullarbor Plain, Western Australia. Helictite, 8:3-22.
Ineson, J. 1962. A hydrogeological study of the permeability of the chalk.
J. Instn. Water Engns., 16:449-63.
Jennings, J. N. 1963. Geomorphology of the Dip Cave, Wee Jasper, New
South Wales. Helictite, 1(3): 43-58.
-------- 1964. Geomorphology of Punchbowl and Signature Caves, Wee
Jasper, New South Wales. Helictite, 2:57-80.
--------- 1966. Jiri V. Danes and the Chillagoe Caves District. Helictite,
4:83-7.
--------- 1967a. Some karst areas of Australia. Pp. 256-92 in Landform
Studies from Australia and New Guinea, ed. J. N. Jennings and
J. A. Mabbutt. Canberra.
-------- 1967b. Further remarks on the Big Hole, near Braidwood, New
South Wales. Helictite, 6:3-9.
-------- 1967c. The surface and underground geomorphology. Pp. 13-31 in
Caves of the Nullarbor, ed. J. R. Dunkley and T. M. L. Wigley.
Sydney.
--------- 1968. Syngenetic Karst in Australia. Pp. 41-110 in Contributions
to the Study of Karst, P. W. Williams and J. N. Jennings. Aust.
Nat.Univ.Dept.Geogr.Pub., G/5. Canberra.
-------- 1969. Karst of the seasonally humid tropics in Australia. Pp. 149-
58 in Problems of the Karst Denudation, ed. O. Stelcl. Brno.
-------- and Bik, M. J. 1962. Karst morphology in Australian New Guinea.
Nature, Lond., 194:1036-8.
-------- and Sweeting, M. M. 1963. The Limestone Ranges of the Eitzroy
Basin, Western Australia. Bonn, geogr. Abh., 32.
Jordan, G. F. 1954. Large sink holes in the Straits of Florida. Bull. Am.
Ass. Petrol. Geol., 38:1810-17.
Jordan, R. H. 1950. An interpretation of Floridian Karst. J. Geol., 58:
261-8.
 Bibliography 235

Katzer, F. von 1909. Karst und Karsthydrographie. Zur Kunde der Balkan­
halbinsel, 8. Sarajevo.
Kayser, K. 1934. Morphologische Studien in Westmontenegro II. Z.Ges.
Erdk. Berl., 1934:26-49, 81-102.
Keller, W. D. 1957. The principles of chemical weathering. Columbia, Miss.
Kendall, P. F. and Wroot, H. E. 1924. The Geology of Yorkshire. Vienna.
King, L. 1951. The geology of the Makapan and other caves. Trans. R.
Soc. S.Afr., 33:121-51.
King-Webster, W. A. and Kenny, J. S. 1958. Bat erosion as a factor in
cave formation. Nature, Lond., 181:1813.
Kirkaldy, J. F. 1950. Solution of the chalk in the Mimms Valley, Herts.
Proc. Geol. Ass., 61:219-24.
Klaer, W. von 1957. Karstkegel, Karst-Inselberg und Poljeboden am
Beispiel des Jezeropoljes. Petermanns geogr. Mitt., 101:108-11.
Krebs, N. 1929. Ebenheiten und Inselberge im Karst. Z. Ges. Erdk. Berl.,
1929:81-94.
Krejci-Graf, K. 1935. Felsen aus Salz im Rumänien. Natur Volk, 65:116-
20 .
Kukla, J. and Lozek, V. 1958. K problematice vyzkumu jeskynnich vyplni.
Cslky Kras, 11:19-59.
Kunsky, J. 1958. Karst et Grottes. Trans. Heintz. Nr. 1399. Service
dTnformation Geologique du Bureau de Recherches Geologiques et
Minieres, Paris.
Kyrie, G. 1923. Grundriss der Theoretische Speläologie. Vienna.
Lange, A. L. 1962. Water level planes in caves. Cave Notes, 4:12-16.
Lasserre, G. 1954. Notes sur le karst de Guadeloupe. Erdkunde, 8:115.
La Valle, P. 1967. Some aspects of linear karst depression development in
south central Kentucky. Ann. /Iss. Am. Geogr., 57:49-71.
Lehmann, H. 1936. Morphologische Studien auf Java. Stuttgait.
--------- 1953. Der tropische Kegelkarst in Westindien. Verh. dt. Geogr.
Tags., 29:126-31.
--------- 1959. Studien über Poljen in den venaziaschen Voralpen und im
Hochapennin. Erdkunde, 12:258-89.
--------- 1960. La terminologie classique du karst sous l’aspect critique de
la morphologie climatique moderne. Revue Geogr. Lyon, 35:1-6.
-------- , Krommelbein, K. and Lötschert, W. 1956. Karstmorphologische,
geologische und botanische Studien in der Sierra de los Organos auf
Cuba. Erdkunde, 10:185-204.
-------- , Roglic, L, Rathjens, C., Lasserre, G., Harrassowitz, H., Corbel, J.,
and Birot, P. 1954. Das Karstphänomen in den verschiedenen
Klimazonen. Erdkunde, 8:112-22.
Lehmann, O. 1927. Das Tote Gebirge als Hochkarst. Mitt. geogr. Ges.
Wien, 70:201-42.
--------- 1932. Die Hydrographie des Karstes. Leipzig.
Leighton, M. W. and Pendexter, C. 1962. Carbonate rock types. Mem. Am.
Ass. Petrol. Geol., 1:33-61.
Louis, H. 1956. Die Entstehung der Poljen und ihre Stellung im Taurus.
Erdkunde, 10:339-53.
Lowry, D. C. 1964. The development of Cocklebiddy Cave, Eucla Basin,
Western Australia. Helictite, 3:15-19.
--------- 1967a. The origin of Easter Cave doline, Western Australia. Aust.
Geogr., 10:300-2.
--------- 1967b. Halite speleothems from the Nullarbor Plain, Western
Australia. Helictite, 6:14-20.
236 Bibliography

Lowry, D. C. 1969. The origin of small cavities in the limestone of the Bunda
Plateau, Eucla Basin. Geol. Surv. IV. Aus. Ann. Rpt, 1968:34-7.
Lozek, V., Sekyra, J., Kukla, J., and Feyfar, O. 1956. Vyskum Velke
Jasovske Jeskyne. Antliropozoikum, 6:193-282.
Lozinski, W. von 1907. Die Karsterscheinungen in Galazisch-Podolien.
Jber. geol. Bundesanst., Wien, 1907:683-726.
Marker, M. E. and Brook, G. A. 1970. Echo Cave: a Tentative Quaternary
Chronology for the Eastern Transvaal. Dept. Geog. Env. Stud.
Univ. Witwatersrand, Occ. Pap. 3.
Martel, E. 1910. La theorie de la ‘Grundwasser’ et les eaux souterraines
du karst. Geographie, 21:126-30.
— -------- - 1921. Nouveau traite des eaux souterraines. Paris.
Martin, J. 1965. Quelques types de depressions karstiques du Moyen Atlas
central. Revue Geogr. maroc., 7:95-106.
Matschinski, M. 1962. Sur le probleme d’alignement de donnees apparem-
ment dispersees. C.r. hebd. Seanc.A cad.Sei.,Paris, 254:806-8.
Maucci, W. 1960. La speleogenesi nel Carso Triestino. Boll. Soc. adriat.
Sei. nat., 51:233-54.
Melik, A. 1955. Kraska Polja Slovenije v Pleistocenu. Slovensk Akad.
Znanosti in Umetnosti, 7, Inst. Geogr., 3.
Messines, J. 1948. Les eboulements dans les gypses. Inefficacite des travaux
de reboisement. C.r. hebd. Sea/ic.Acad.Sci.,Paris, 226:1295-6.
Miller, V. C. 1953. A quantitative geomorphic study of drainage basin
characteristics in the Clinch Mountain area, Virginia and Tennessee.
Dept. Geol. Columbia Univ. Tech. Rpt, 3, New York.
Mistardis, G. 1968. Investigations upon influences of sea level fluctuations
on underground karstification in some coastal regions of south
Greece. Congr. 4Cint. Speleol., 3:335-40.
Moneymaker, B. C. 1941. Subriver solution cavities in the Tennessee
Valley. J. Geol., 49:74-86.
Monroe, W. H. 1966. Formation of tropical karst topography by limestone
solution and reprecipitation. Caribb. J. Sei., 6:1-7.
--------- 1968. The karst features of northern Puerto Rico. Bull. natn.
speleol. Soc., 30:75-86.
--------- 1970. A glossary of karst terminology. U.S. Geol. Surv. Water-
supply Pap. 1899K.
Montoriol-Pous, J. 1951. Los processos clasticos hypogeos. Rass. speleol.
ital., 3:119-29.
--------- 1954. Resultado de nuevas observaciones sobre los processos clas­
ticos hypogeos. Rass. speleol. ital., 6:103-14.
Moore, G. W. 1954. The Origin of Helictites. Occ. pap. natn. speleol. Soc.,
1 .

----------- 1960. Geology of Carlsbad Caverns, New Mexico. Natn. Speleol.

Soc. Guide Book, 1:10-18.
--------- 1962. The growth of stalactites. Bull. natn. speleol. Soc., 24:95-106.
------------ and Nicholas, G. 1964. Speleology-—the Study of Caves. Boston.
Morawetz, S. 1967. Zur Frage der Karstebenheiten. Z. Geomorph., 11:1-13.
Morehouse, D. F. 1968. Cave development via the sulfuric acid reactions.
Bull. natn. speleol. Soc., 30:1-10.
Myers, J. O. 1948. The formation of Yorkshire caves and potholes. Trans.
Cave Res. Grp Gt Br., 1 ( 1): 26-9.
Muxart, R., Stchouzkoy, T., and Franck, J. C. 1968. Observations hydro-
karstologiques dans le bassin amont de la Seille (Jura). Congr. 4'
int. Speleol., 3:175-80.
 Bibliography 237

Oertli, H. von 1953. Karbonathärte von Karstgewässern. Stalactite, 4 (4):

8-18.
Oldfield. F. 1960. Studies in the Postglacial history of British vegetation:
lowland Lonsdale. New Phytol., 59:192-217.
Ollier, C. D. 1963. The origin of limestone caves. Helictite, 1(2): 8-12.
-------- 1964. McEachern Cave, Nelson. Viet. Nat., 81:195-7.
-------- 1969. Volcanoes. Canberra.
-------- and Harrop, J. F. 1963. The caves of Mont Hoyo, Eastern Congo
Republic. Bull. natn. speleol. Soc., 25:73-8.
-------- and Holdsworth, D. K. 1968. Caves of Kiriwina, Trobriand Islands,
Papua. Helictite, 6:63-72.
-------- and --------- 1969. Caves of Vakuta, Trobriand Islands, Papua.
Helictite, 7:50-61.
-------- a n d -------- 1970. Sonic caves of Kitava, Trobriand Islands, Papua.
Helictite, 8:29-38.
-------- and Tratman, E. K. 1969. Geomorpliology of the caves. Ch. 4 in
The Caves of North-West Clare, Ireland, ed. E. K. Tratman.
Newton Abbot.
Ongley, E. D. 1968. An analysis of the meandering tendency of Serpentine
Cave, N.S.W. Jnl Hydrol., 6:15-32.
Panos, V. and Stelcl, O. 1968. Physiographic and geologic control in
development of Cuban mogotes. Z. Geomorph., 12:117-65.
Parde, M. 1965. Influences de la permeabilite sur le regime des rivieres.
Colloquium Geographicum, 7. Bonn.
Parker, G. G. 1964. Piping: a geomorphic agent in landform development
of the drylands. Int. Ass. Sei. Hydrol. Pub., 65:103-13.
Penck, A. 1924. Das unterirdische Karstphänomen. Pp. 175-97 in Zbornik
Radova Posvecen Jovanu Cvijicu, ed. P. Vujevic. Belgrade.
Pengelly, W. 1864. The introduction of cavern accumulations. Rep. Trans.
Devon. /Hs. Advmt. Sei., 3:31-41.
Pevalek, I. 1935. Der Travertin und die Plitvice Seen. Verb. int. Verein,
theor. angew. Limnol., 7:165-81.
Pfeffer, K-H. von 1967. Neue Beobachtungen im Kegeikarst von Jamaica.
Verb. dt. Geogr. Tags., 36:345-58.
-------- 1969. Charakter der Verwitterungsdecken im tropischen Kegelkarst
und ihre Beziehung zum Formenschatz. Geol. Rdscb., 58:408-26.
Pigott, C. D. 1962. Soil formation and development on the Carboniferous
Limestone of Derbyshire: parent materials. J. Ecol., 50:145-56.
Pinchemel, P. 1954. Les Plaines de Craie du Nord-ouest du Bassin Parisien
et du Sud-est du Bassin de Londres et leurs Bordures. Paris.
Pitty, A. F. 1966. An Approach to the Study of Karst Water: Illustrated
by Results from Poole’s Cavern, Buxton. Univ. Hull Occ. Paper
Geogr., 5.
-------- 1968. The scale and significance of solutional loss from the lime­
stone tract of the Southern Pennines. Proc. Geol. Ass., 79:153-78.
Playford, P. D. and Lowry, D. C. 1966. Devonian Reef Complexes of the
Canning Basin, Western Australia. Bull. Geol. Surv. W.Aust., 118.
Pluhar, A. and Ford, D. C. (1970). Dolomite Karren of the Niagara
escarpment, Ontario, Canada. Z. Geomorph., 14:392-410.
Pohl, E. R. 1955. Vertical Shafts in Limestone Caves. Occ. pap. natn.
speleol. Soc., 2.
--------- and White, W. B. 1965. Sulfate minerals; their origin in the central
Kentucky karst. Am. miner., 50:1461-5.
Price, H. J. 1959. Mechanics of jointing rocks. Geol. Mag., 96:149-67.
238 Bibliography

Priesnitz, K. von 1969. Über die Vergleichbarkeit von Lösungsformen auf

Chlorid-, Sulfat- und Karbonatgestein-Überlegungen zu Fragen der
Nomenklatur und Methodik der Karstmorphologie. Geol. Rdsch.,
58:427-38.
Prinz, W. 1908. Les cristallisations des grottes de Belgique. Brussels.
Rathjens, C. 1954. Zur Frage der Karstrandebene im dinarischen Karst.
Erdkunde, 8:114-15.
--------- 1960. Beobachtungen an hochgelegenen Poljen im südlichen
dinarischen Karst. Z. Geomorph., 4:141-51.
Reams, M. W. 1968. Cave sediments and the geomorphic history of the
Ozarks. Ph.D. thesis, Washington University, St Louis, Miss.
Reid, C. 1887. On the origin of dry chalk valleys and of coombe rock.
Q. Jl geol. Soc. Lund., 43:364-73.
Renault, P. 1958. Elements de speleomorphologie karstique. Ann. Speleol.,
13:23-48.
--------- 1959. Le karst Kouilou (Moyen Congo, Gabon). Revue Geogr.
Lyon, 1959:305-14.
--------- 1967-8. Contribution ä l’etude des actions mecaniques et sedimen-
tologiques dans la speleogenese. Ann. Speleol., 22:5-21, 209-67;
23:259-307.
Rhoades, R. and Sinacori, M. N. 1941. Pattern of groundwater flow and
solution. /. Geol., 49:785-94.
Riedl, H. R. 1961. Gedanken zur methodischer Intensivierung der speläo-
genetischen Forschung. Petermanns geogr. Milt., 105:264-8.
Roglic, J. 1939. Morphologie der Poljen von Kupres und Vukovsko.
Z. Ges. Erdk. Berl., 1939:299-316.
--------- 1940. Geomorphologische Studien von Duvanjsko Polje (Polje von
Duvno) in Bosnien. Mitt. geogr. Ges. Wien, 83:152-76.
--------- 1954. Korrosive Ebenen im dinarischen Karst. Erdkunde, 8:113-14.
--------- 1964a. Karst valleys in the Dinaric Karst. Erdkunde, 18:113-16.
--------- 1964b. Les poljes du karst dinarique et les modifications climatiques
du Quaternaire. Rev. beige Geogr., 88:105-25.
Roques, H. 1964. Contribution ä l’elude statique et cinetique des systemes
gaz carbonique-eau-carbonate. Ann. Speleol., 19:255-484.
--------- 1969. A review of present-day problems in the physical chemistry
of carbonates in solution. Trans. Cave Res. Grp Gt Br., 11:139-64.
Sanders, E. M. 1921. The cycle of erosion in a karst region (after Cvijic).
Geogr. Rev., 11:593-604.
Sawicki, L. R. von 1909. Ein Beitrag zum geographischen Zyklus im
Karst. Geogr. Ztschr., 15:185-204, 259-81.
Schmid, E. 1958. Höhlenforschung und Sedimentanalyse. Ein Beitrag zur
Datiemng des Alpinen Paläolithikums. Institut für Ur- und Frühge­
schichte der Schweiz. Schriften 13, Basel.
Schoemaker, R. P. 1948. A review of rock pressure problems. Mining
Technology, 12 (TP2495): 1-14.
Shaw, T. R. and Tratman, E. K. 1969. Mainly historical. Pp. 15-32 in The
Caves of North-West Clare, Ireland, ed. E. K. Tratman. Newton
Abbot.
Siffre, A. and Siffre, M. 1961. Le fa^onnement des alluvions karstiques.
Ann. Speleol., 16:73-80.
Siffre, M. 1959. Ecoulement nappo-laminaire et morphologie souterraine.
Stalactite, 4:49-55.
Silar, J. 1965. Development of tower karst of China and North Vietnam.
Bull. natn. speleol. Soc., 27:35-46.
 Bibliography 239

Smith, D. I. 1965. Some aspects of limestone solution in the Bristol region.

Geogrl J., 131:44-9.
--------- 1969. The solution erosion of limestone in an arctic morphogenetic
region. Pp. 99-110 in Problems of the Karst Denudation, ed.
O. Stelcl. Brno.
-------- , High, C., and Nicholson, F. H. 1969. Limestone solution and the
caves. Pp. 96-123 in The Caves of North-West Clare, Ireland, ed.
E. K. Tratman. Newton Abbot.
--------- and Mead, D. G. 1962. The solution of limestone. Proc. speleol.
Soc. Univ. Bristol, 9:188-211.
Smyk, B. and Drzal, M. 1964. Research on the influence of microorganisms
on the development of karst phenomena. Geographia Polonica, 2:
57-60.
Sparks, B. W. 1961. Geomorphology. London.
--------- and Lewis, W. V. 1957. Escarpment dry valleys near Pegsdon,
Hertfordshire. Proc. Geol. A.«., 68:26-38.
Stelcl, O., Vlcek, V., and Pise, J. 1969. Limestone solution intensity in the
Moravian karst. Pp. 71-86 in Problems of the Karst Denudation,
ed. O. Stelcl. Brno.
Stringfield, V. T. and LeGrand, H. E. 1969a. Hydrology of carbonate rock
terranes—a review. Jnl Hydro!., 8:349-413.
--------- and --------- 1969b. Relation of sea water to fresh water in
carbonate rocks in coastal areas, with special reference to Florida,
U.S.A. and Cephalonia (Kephallinia), Greece. Jnl Hydrol., 9:
387-404.
Sunartadirdja, M. A. and Lehmann, H. 1960. Der tropische Karst von
Maros und Nord-Bone in SW-Celebes (Sulawesi). Z.Geomorph.-
Suppl., 2:49-65.
Sweeting, M. M. 1950. Erosion cycles and limestone caverns in the Ingle-
borough District of Yorkshire. Geogrl J., 115:63-78.
--------- 1953. The enclosed depression of Carran, County Clare. Ir. Geogr.,
2:218-24.
--------- 1955. The landforms of North-West County Clare, Ireland. Trans.
Inst. Br. Geogr., 21:33-49.
--------- 1958. The karstlands of Jamaica. Geogrl J., 124:184-99.
--------- 1965. Denudation in limestone regions, I Introduction. Geogrl J.,
131:34-7.
--------- 1966. The weathering of limestones. With particular reference to
the Carboniferous limestones of northern England. Pp. 177-210 in
Essays in Geomorpliology, ed. G. H. Dury. London.
-------- 1968. Karstic morphology. In The University of Edinburgh British-
Honduras-Yucatan Expedition. Geogrl J., 134:49-54.
Swinnerton, A. C. 1929. The caves of Bermuda. Geol. Mag., 66:79-84.
--------- 1932. Origin of limestone caverns. Bull. geol. Soc. Am., 43:662-93.
Symposium on Cave Surveying 1970. Trans. Cave Res. Grp Gt Br., 12(3):
1-245.
Symposium on Cave Hydrology and Water Tracing. Trans. Cave Res. Grp
Gt Br., 10(2): 1-125.
Thomas, T. M. 1954. Swallow holes on the Millstone Grit and Car­
boniferous Limestone of the South Wales Coalfield. Geogrl J., 120:
468-75.
-------- 1963. Solution subsidence in south-east Carmarthenshire and
south-west Breconshire. Trans. Inst. Br. Geogr., 33:45-60.
240 Bibliography

Thorp, J. 1934. The asymmetry of the Pepino Hills of Puerto Rico in

relation to the Trade Winds. J. Geol., 42:537-45.
Thrailkill, J. V. 1960. Origin and development of Fulford Cave, Colorado.
Bull. natn. spe'leol. Soc., 22:54-65.
--------- 1968. Chemical and hydrologic factors in the excavation of lime­
stone caves. Bull. geol. Soc. Am ., 79:19-45.
Tratman, E. K. 1969. The rate of solution of carboniferous limestone in
the caves of north west Clare, Ireland. Pp. 87-98 in Problems of the
Karst Denudation, ed. O. Stelcl. Brno.
Tricart, J. 1955. Modele karstique et modele periglaciaire dans les Causses.
Revue geomorph. dyn., 6:193-201.
--------- 1968. Notes geomorphologiques sur la karstification en Barbade
(Antilles). Mem. Docums. Cent.Docum. cartogr. geogr., 4:329-34.
--------- and Silva, T. C. da 1960. Un exemple devolution karstique en
milieu tropical sec: Le morne de Bom Jesus da Lapa (Bahia,
Bresil). Z. Geomorph., 4:29-42.
Trimmei, H. 1950. Beobachtungen zur Frage der Raumbildung in der
Badlhöhle im Mittelsteirischen Karst. Mitt. geogr. Ges. Wien, 92:
26-30.
--------- 1951. Morphologische und genetische Studien in der Salzofenhöhle.
Höhle, 2:2-7.
--------- (ed.) 1965. Speläologisches Fachwörterbuch. Congr. 3° int. Speleol.,
C.
--------- 1968. Höhlenkunde. Brunswick.
Trombe, F. 1952. Trait e de Speleologie. Paris.
Twidale, C. R. 1971. Structural Landforms. Canberra.
Varnes, D. J. 1958. Landslide types and processes. Pp. 20-47 in Landslides
and Engineering Practice, ed. E. B. Eckel. Highway Research Board
Spec. Ref., 29.
Vaumas, E. de 1968. Phenomenes karstiques en Mediterranee orientale.
Mem. Docums. Cent.Docum. cartogr. geogr., 4:193-282.
Veeh, H. H. and Chappell, J. 1970. Astronomical theory of climatic
change: support from New Guinea. Science, N .Y., 167:862-5.
Verstappen, H. Th. 1960a. On the geomorphology of raised coral reefs
and its tectonic significance. Z. Geomorph., 4:1-28.
--------- 1960b. Some observations on karst development in the Malay
Archipelago. J. trop. Geogr., 14:1-10.
--------- 1964. Karst morphology of the Star Mountains (central New
Guinea) and its relation to lithology and climate. Z. Geomorph.,
8:40-9.
Wagner, G. 1935. Vom Salz des Toten Meeres. Natur Volk, 65:108-16.
Wall, J. R. D. and Wilford, G. E. 1966. Two small-scale solution features
of limestone outcrops in Sarawak, Malaysia. Z. Geomorph., 10:90-4.
Warwick, G. T. 1950. The reef limestone caves of the Dove and Manifold
Valleys. Cave Res. Grp Gt Br. Newsl., 31:2-6.
--------- 1953. Polycyclic swallow holes in the Manifold valley, Stafford­
shire, England. Congr. l ,r int. Speleol., 2:59-68.
--------- 1958. The characteristics and development of limestone regions in
the British Isles with special reference to England and Wales. Congr.
2C int. Speleol., 1:79-105.
--------- 1960. The effect of knick-point recession on the watertable and
associated features in limestone regions, with special reference to
England and Wales. Z. Geomorph.-Suppl., 2:92-7.
 Bibliography 241

Warwick, G. T. 1962. Cave formations and deposits. Pp. 83-119 in British

Caving, ed. C. H. D. Cullingford. London.
--------- 1964. Dry valleys in the southern Pennines, England. Erdkunde,
18:116-23.
--------- 1968. Some primitive features in British caves. Congr. 4' int.
Speleol., 3:239-52.
Weyl, P. K. 1958. The solution kinetics of calcite. J. Geol., 66:163-76.
White, E. L. and Reich, B. M. 1970. Behaviour of annual floods in lime­
stone basins in Pennsylvania. Jnl Hydrol., 10:193-8.
White, W. B. 1960. Termination of passages in Appalachian caves as
evidence for a shallow phreatic origin. Bull. natn. speleol. Soc.,
22:43-53.
--------- 1969. Conceptual models for carbonate aquifers. Ground Water,
7:15-21.
--------- , Jefferson, G. L., and Haman, J. F. 1966. Quartzite karst in south­
eastern Venezuela. Int. J. Speleol., 2:309-14.
Wiiford, G. E. 1964. The geology of Sarawak and Sabah Caves. Bull. geol.
Surv. Borneo Region, Malaysia, 6.
--------- and Wall, J. R. D. 1965. Karst topography in Sarawak. J. trap.
Geogr., 21:44-70.
Williams, P. W. 1963. An initial estimate of the speed of limestone solution
in County Clare. Ir. Geogr., 4:432-41.
--------- 1966a. Limestone pavements with special reference to Western
Ireland. Trans. Inst. Br. Geogr., 40:155-72.
-------- - 1966b. Morphometric analysis of temperate karst landforms. Ir.
Speleol., 1:23-31.
--------- 1968. An evaluation of the rate and distribution of limestone
solution in the River Fergus basin, western Ireland. Pp. 1-40 in
Contributions to the Study of Karst. Aust. National Univ. Dept
Geogr. Pub. G /5. Canberra.
--------- 1969. The geomorphic effects of ground water. Pp. 269-84 in
Water, Earth and Man, ed. R. J. Chorley. London.
--------- 1970. Limestone morphology in Ireland. Ch. 7 in Irish Geographi­
cal Studies, ed. R. Glosscock and N. Stephens. Belfast.
Wissmann, H. von 1954. Der Karst der humiden heissen und sommer­
heissen Gebiete Ostasiens. Erdkunde, 8:122-30.
--------- 1957. Karsterscheinungen in Hadramaut. Ein Beitrag zur Mor­
phologie der semiariden und ariden Tropen. Petermanns Mitt. Erg.,
262:259-68.
Woodward, H. P. 1936. Natural Bridge and Natural Tunnel, Virginia.
J. Geol., 44:604-16.
Wright, F. J. 1936. The Natural Bridge of Virginia. Va. Geol. Surv. Bull.,
46G:50-78.
Würm, A. 1953. Der Salzberg bei Djelfa im Sahara-Atlas. Natur Volk,
83: 141-7.
Zötl, J. 1957. Neue Ergebnisse der Karsthydrologie. Erdkunde, 11:107-17.
--------- 1965. Tasks and results of karst hydrology. Pp. 141-5 in Problems
of the Speleological Research, ed. O. Stelcl. Prague.
Zotov, V. D. 1941. Pot-holing of limestone by development of solution
cups. J. Geomorph., 4:71-3.
 INDEX
N.S.W. = New South Wales; Vic. = Victoria; Qld = Queensland;
Tas. — Tasmania; S.A. = South Australia; W.A. = Western Australia;
N.Z. = New Zealand; N.G. = Territory of Papua and New Guinea;
Craven = Craven, Yorkshire; Peak = Peak District, Derbyshire;
Mendip = Mendip Hills, Somerset; Clare = County Clare, Ireland.
The names of the individual states and provinces of the United States
of America and the Federal Republic of Yugoslavia are employed alone.
Bold face indicates page reference to figures, plates, and tables.

accordance of summits, 219 bacteria, 24, 27, 38

acid: carbonic, 23-4, 53; nitric, 24; Bahia, Brazil, 52
organic, 24; sulphuric, 24 Baldocks Cave, Mole Creek, Tas.,
Aguada Limestone, Puerto Rico, 187 176
allochem, 13 Bar Pot, Craven, 153
Amiens, France, 120 Barber Cave, Cooleman Plain,
anastomoses, 157 N.S.W., 158
Anavalos, Kynouria, Greece, 78-9 Bas Queyras, France, 186
Angabara R , Central Highlands, Bathhouse Cave, Yarrangobilly,
N.G., 62 N.S.W., 85, 110-11, 116
anhydrite, 2, 16, 53 bats, 38, 174
Aniene R., Apennines, Italy, 66 bedding, 20, 34, 54. 153, 165, 178,
anthropogenic effects on landforms, 203
41-2, 47-8, 49, 50, 118-19, 123, bedding grike, 47
176 bedding plane passage, 160, 161,
aquiclude, 89, 97 200, 209
aquifer, 80, 89-90 Beljanica Mountains, Serbia, 213
aragonite, 10, 15, 23, 178 hellhole, 198
arch, 34-5, 36, 37; natural, 101-2 Bermuda, 198
Arch Cave, Abercrombie, N.S.W., Big Hole, Braidwood, N.S.W., 126,
102 128
Arize R., France, 77 bioherm, 12, 15, 20, 200
atolls, emerged, 197 biolithitc, 11, 19, 201
autocatalysis, 51, 122, 159 biomicritc, 19, 41
auto-desiccation, 121 biostrome, 12, 16
Austwick Beck Head, Craven, 75 Birkwith Cave, Craven, 75
aven, sec blind shaft Bisistal, Switzerland, 57
Aven Armand, Lozere, France, 173 Black Range Cave, Cooleman Plain,
Aven d’Orgnac, France, 178 N.S.W., 206, 209
Avcns du Pic St-Loup, Provence, Blato Polje, Macedonia, 137
France, 148 blind shaft, 151, 152. 157
243
244 Index

blockpile, 153, 165, 175, 205 cave: active, 146; between-, 146;
Blue Spring, Florida, 81 branchwork, 153, 222; cliff-foot,
Blue Waterholes, Cooleman Plain, 191, 192, 218; collapsed, 100;
N.S.W., 68, 81, 82, 83 dead, 146; deep phreatic, 155-6,
bone-breccia, see gash-breccia 176- 7, 214-15; epiphreatic, 159;
Bourbioz, L. Annecy, France, 78 in emerged reef, 197-8; inflow,
146; master, 159, 217; meander,
Bourne R., Vercors, France, 86
100, 101, 167; network, 156-7;
boxwork, 205 relict, 100; shallow phreatic, 158-
breakdown, see collapse 65, 177, 198, 215; outflow, 76,
bridge, 158; natural, 100, 101-6, 109, 146, 191; submerged. 198; tectonic,
113, 202 147-8; through-, 146, 218; vadose,
Buckhaw Brow, Craven, 77 151-5, 160, 161, 203
Buege R., France, 86, 87 cave breakdown, 175-7
Bungonia Gorge, N.S.W., 98, 99 cave chamber, 164-5
Bwabwatu Cave, Kaileuna I., Trob- cave cross-sections, 35, 37 , 166-7,
riand Is., N.G., 198 203, 205
cave deposits, 149, 156, 157, 160,
calanque, 219 161, 167, 171-6
calcarenite, 22, 201; aeolian, 37, 51, cave diving, 9, 77, 157
127, 198-9 cave entrance, 146, 166-7; facies,
calcilutite, 13 177- 8
calcirudite, 14, 201 cave ice, 174
calcite, 10, 13, 16, 19, 23, 171, cave loam, 177
172-4, 176; floe, 173 cave longitudinal sections, 147, 152,
calcite vein, 29, 205 160. 163, 203
calclithite, 14 cave meanders, 160 , 167
calcrete, 29, 45, 51, 184 cave morphometry, 165-7
calc-sinter, 106 cave plans, 102, 154, 202, 203
Cambridge, England, 96 cave stream deposits, 175-6
Cameron Cave, Missouri, 154 cave surveying, 9
Camooweal, Qld, 53 Celebes, Indonesia, 143, 192
Canterbury, N.Z., 41 cenote, 125, 126, 128
Cape Leeuwin, W.A., 74 Central Highlands, N.G., 41, 42,
carbon dioxide, 24-9, 171, 186; 50, 51, 133-5
partial pressure of, 24-5, 56, 193; Central Lowland, Ireland, 66, 69
biogenic, 27, 48, 56, 142, 185, Cerknica Polje, Slovenia, 139
193 Ceteniste uvala, Yugoslavia, 135
carbonate rocks, impurities of, 10 Cetina Plain, Dalmatia, 64, 217
Carboniferous Limestone, S. Wales, chalk, 10, 19, 35, 40, 51, 61, 95,
125, 151 96, 118-19. 144, 201
Carlsbad Cavern, New Mexico, 145, Cheddar Gorge, Mendip, 100, 222
214
chelate, 24
Carran, Clare, 221
chert bands, 204
Carrao Basin, Venezuela, 5
case-hardening, 29, 40, 192 Chillagoe, Qld, 200
Castile Formation, New Mexico Chilterns, England, 120-1
and Texas, 2 Chrome Hill, Peak, 4
Castle Flarbour, Bermuda, 198 Cikola R., Dalmatia, 219-20
Cathedral Cave, Wellington, N.S.W., Clapham Beck Head, Craven, 75,
163-4 78
 Index 245

Clare, 56, 57 , 59, 82-3, 148 , 167, Dark Cave, Bukit Batu, Malaysia,
169, 193, 202-3, 214, 221-2 36
cleavage, 20, 34, 46 Deepdene, Augusta, W.A., 199
Cliff Cave Spring, Cooleman Plain, degenerate karst hydrology, 93
N.S.W., 75 Devonport, Tas., 4
climatic change, 41, 119, 123, 124, dip, 54-5, 123; primary depositional,
176, 193, 213-15 15, 200
Clinch Mountains, Virginia, 64-5 Dip Cave, Wee Jasper, N.S.W.,
clint, 47, 51, 53, 54, 56 153, 157, 164, 175, 203
Cluj, Romania, 7 dismicrite, 13
cockpit, 63, 187-9 Djelfa, Algeria, 53
cockpit morphometry, 133-4 doline, 120-8, 194, 218-19; alluvial
collapse, 32-8, 100, 112-13, 125-6, streamsink, 127-8; collapse, 124,
153, 156, 160, 161, 165, 178, 125-6, 127, 131, 184; solution,
197, 198, 203 121-4, 130-3, 186, 189; subjacent
Colombia, 2 karst collapse, 125-6, 127, 128 ,
column, 162, 173 131; subsidence, 37 , 126-7, 130-3,
186, 227
conekarst, 4, 182- 3 , 187, 188
doline alignment, 135, 189
constructive waterfalls, 107
doline asymmetry, 123-4
Coolagh River Cave, Clare, 150
doline density, 128
Cooleman Plain, N.S.W., 40, 43,
71, 84, 109, 110, 111, 114, 118, doline depth-diameter ratio, 131,
132
149
coombe rock, 119 doline elongation ratio, 132-3
Coppermine Cave, Yarrangobilly, doline morphometry, 128-35
N.S.W., 85 doline product of symmetry, 135
coquina, 14 dolomite, 2, 10, 16, 52, 176, 201;
solution, 29
coral reef, 14-15, 61, 196-200;
emerged, 197 dome, 34; salt, 184
corrasion, 67-9, 84, 100, 149-50, dome pit, 151
155, 156 Douk Gill Cave, Craven, 75
corrosion: biological, 38; cooling, Dourbie R., Grandes Causses,
27-8; mixing, 27, 28, 157, 170; France, 98
river, 67-9; see also solution drainage, 61, 64
Cotswolds, England, 84 drainage density, 64, 65
Craigmore, S. Canterbury, N.Z., drainage inheritance, 118
131-2 dripstone, 168, 173
Craven, 39, 43, 49, 50, 51, 54, 56, Diidencay R., Antalya, Turkey, 76,
57, 58, 74-5, 118, 126, 181, 203, 226
206-7, 214, 215-17 Duvno Polje, Bosnia, 141
Croesus Cave, Mole Creek, Tas.,
173 Easter Cave, Augusta, W.A., 33, 37,
Cullaun Caves, Clare, 85, 86 127-8, 172
current markings, 154, 155, 164, ebbing and flowing well, 75, 77
198 Ebbor Gorge, Mendip. 222
cutter, 49 eclogite, 2
Efflux, Bungonia, N.S.W., 78
Dachstein Plateau, Austria, 55-6, 57, Eisriesenwelt, Austria, 174
94, 148 erosion surface, 140, 156, 182- 3 ,
Darcy’s Law, 88-90, 208 199, 215-17, 221, 222
246 Index

estaveile, 74, 139, 164 Gouffre Berger, France, 145

evaporation, 66, 171, 186 Gouffre de la Luire, France, 92,
evaporite, 16-17, 30, 184 163, 164
Exit Cave, Tas., 204 Gouffre Henne Morte, France, 151,
exsurgence, 74 152
Gouffre St Pierre Martin, France,
faults, 22, 86, 87, 91 , 96, 133, 140, 145, 226
147-8, 153, 164, 201, 202, 203-4 Goulden’s Hole, S.A., 125
Fergus R., Clare, 68 , 69, 181 gour, see rimstone dam
Flint Ridge Cave, Kentucky, 145 Gowlaun R., Clare, 84
flood pulse, 9, 81 Grand Arch, Jenolan, N.S.W., 102
Florida, 219, 226 Grand Canyon, Colorado, 181
flow, artesian, 89, 97; confined, 89, Grand Plan de Canjuers, Provence,
208; deep phreatic, 96-7, 155-6, France, 163, 164
219; diffuse, 208; free-surface, 92, Grandes-Chaumilles, Jura Moun­
160; pressure, 92, 93, 94 , 160; tains, Switzerland, 123
rate of underground, 168; uncon­ granite, 5, 192-3
fined, 89; vadose, 97, 153-5 Green R., Kentucky, 68
flowstone, 173 Greta R., Craven, 216
flutes (Curl), 155 grike, 35 , 46, 49, 51, 52, 54, 60, 62
foiba, see blind shaft Grotta Gigante, Trieste, Italy, 145,
folds, 62, 80, 86, 87 , 96. 140, 147, 166
164, 218 Grotte de la Cigalere, France, 164
fossils, 13, 14, 16, 19 Grotte de Moulis, Ariege, France,
Fountain of Vaucluse, Provence, 176
France, 76, 81, 112 Grotte des Sourciettes, France, 168
Franconian Jura, Germany, 201 groundwater, 83, 88-9, 96-7, 198
Fred Cave, Auckland, N.Z., 154 Guadalupe Range, New Mexico, 16,
frost shattering, see thaw-freeze 214-15
action Guadeloupe, 66
frost wedging, see thaw-freeze guano, 51, 174-5, 191
action Gulf of Edremit, Turkey, 78
Gulf of Sibenik, Dalmatia, 219
Gunung Sewu, Java, 187, 189, 213
G.B. Cave, Mendip, 168, 170 gypsum, 2, 16, 31, 35, 51, 53, 173,
Gambier Limestone, S.A., 63-4 184
Gaping Gill Hole, Fell Beck, gypsum solution, 30
Craven, 152-3
gash-breccia, 33, 175 Haggas Hole, Waitomo, N.Z., 135
Geikie Gorge, Fitzroy R., W.A., 99 half-tube, 157, 160, 161
‘giant grikeland’, 200 halite, 2, 16, 35, 53, 54. 173, 184
glacial erosion, 49, 54, 56, 60, 184- halite solution, 30
6, 214, 221 head (of water), 88-9, 92. 93, 94,
glacial moraine, 49, 51, 58. 63, 66, 160, 164
126, 177, 222 helictite, 172, 173
glaciere, 174, 186 Hells Hole, S.A., 128
glacioeustatic sea level changes, 79, Henry’s Law, 24
197-8, 219 liojo, 136
glade, 188 Hölloch Cave, Muototal, Switzer­
Golconda Cave, Peak, 4 land, 78, 145, 203
Gölling Falls, Austria, 78 holokarst, 200
gorge, 86, 98-9; depositional, 115, Honeycomb Cave, Mole Creek,
199, 222 Tas., 153
 Index 247

Hranic:, Moravia, Czechoslovakia, 21; subsoil, 3; syngenetic, 198-9;

145 tower, 187, 190-3, 195, 200, 202,
Hull P)t, Craven, 203 214; tropical humid, 61, 133, 174,
Hull R, England, 69 179, 182- 3, 186-92, 194, 213,
hum, 139, 211 220
Huon, N.G., 197 karst barre, 4
karst conduit, 92, 93, 97
karst corridor, 120, 190, 200, 201,
Iehi, Kkori R., N.G., 182-3 204
Igue des Landes, Tarn, France, 147 karst cycle, 179, 189, 210-12
karst denudation rate, 179-80, 184,
index: of pitting, 130; vadose, 79 193, 196
Indonesia, 57, 181 karst immunity, 184, 196
infiltraiion, 33, 61, 63, 66 karst literature, 7-8
Inglebc rough, Craven, 73 karst relief ratio, 133
Inglebcrough Cave, Craven, 78 karst terminology, 8
interior valley, see valley, inland karst well, 46
intrachst, 13, 15, 16, 19 katavothre, 74
Kentucky, 132-3, 151, 152, 205
Kestrel Cavern, Nullarbor Plain,
Jamaica, 57, 179, 187, 189. 193, Australia, 182-3
211 Kikori R„ N.G., 189
Jesero Lake, Dalmatia, 139 King Country, N.Z., 21, 149
Jewel Cave, Augusta, W.A., 165 Kinta R., Malaysia, 66, 143, 190,
191, 204
joints, 19, 20, 34, 40, 54, 62, 86,
Kirk Göz, Antalya, Turkey, 76
96, 118, 119, 121, 130, 133, 153,
160, 164, 165, 189, 192, 196, 202, Kissimee R., Florida, 181
203, 204. 218 Kistanje Plain, Dalmatia, 217-18,
Jonte R.. Grandes Causses, France, 219
98 Konya, Turkey, 126
Jura Mountains, 57, 196, 218-19, Koonalda Cave, Nullarbor Plain,
222 Australia, 176
Kouilou, Gabon, 192
Krka R., Dalmatia, 64, 108, 217,
Kairimu Cave, Auckland, N.Z., 219
220-1 Kupres Polje, Bosnia, 137, 141
kankar, 29 Kweichow, China, 214
Karlovac-Sluin Plain, Croatia, 217 Kweilin, Baisha, China, 218
Karren, 39, 41, 53, 186
karst: altitudinal zonation of, 179; Lagaip R., Central Highlands, N.G.,
arid, 4), 64, 71, 122, 125, 144, 76
177, 131-4, 196, 214-15; bare, 3, Lake Cave, W.A., 66
41, 42 buried, 4; coastal, 3, 185,
lakes, 91. 107, 110, 115, 121, 125,
186, 197; cockpit, 187, 189, 192,
202, 211; cold climate, 112, 119, 126, 138. 139
122, 144, 174, 178, 184-5, 213- Langkawi Is., Kedah, Malaysia, 192,
14; co/ered, 3, 47, 48, 62; defini­ 220
tion of, 1; doline, 129, 130, 186, lapiez, 39
187, 95, 211-12; exhumed, 4, Lapland, Norway, 57
222; fossil, 4; free, 5, 64; glacier lateral solution undercutting, 100,
thaw, 5; hydrothermal, 144-5; 101, 141. 191-2, 201
impomded, 5; mantled, 3; relict, lime boundstone, 11
4; subjacent, 3; submerged, 219- lime grainstone, 14
248 Index

lime mud, 11 meander niche, 176

lime mudstone, 13 meander spur, 102, 153, 176
lime packstone, 13 megabreccia, 15, 200
lime wackestone, 13 Mellte R„ S. Wales, 83, 181
limestone, 2, 10; crystalline, 14; meltwater, 26, 119, 170, 176, 186,
dune, 11, 35; magnesian, 10 213-14
limestone alluvium, 63, 66, 112, Mende Polje, Portugal, 141
142, 143, 170 Mendip. 27, 83, 131, 164, 168, 169,
limestone classification, 11 181, 203, 206, 222-3
limestone colluvium, 63 Meregill, Craven, 203
limestone pavement, 54-5, 214, 221 merokarst, 200-1
Metro Cave, Charleston, N.Z., 161,
limestone purity, 10, 17 162
Limestone Ranges, W. Kimberley, micrite, 13, 19
W.A., 16, 57, 98-100, 199 Mixnitz Cave, Austria, 174
limestone removal rate, 7, 68 mogote, see tower
limestone solution, 23-8 Mole Creek, Tas., 41, 86, 87, 126.
Livanjsko Polje, Croatia, 91-2, 137 136
London Bridge, Burra Creek, moonmilk, 38
N.S.W., 102, 104-5 Moraca R., Montenegro, 64
Long Kin East Pot, Craven, 203 Morecambe Bay, England, 41
Los Alamos, New Mexico, 181 morphology: climatic, 179-94;
Lost World, R. Mangapu, N.Z., structural, 196-209
136 morphometry, 71-4, 79. 112, 120,
Lot R., Grandes Causses, France, 125, 128-35, 155, 165-7
98 Mt Arthur, N.Z., 40, 45 , 185
Luray Caverns, Virginia, 29 Mt Etna, Qld, 145
Mt Hoyo, Congo, 215
Mt Kaijende, Central Highlands,
N.G., 134, 135, 194
magnesium, 29-30 Mt Pierre Gorge, W.A., 99
magnesium carbonate, 10, 29-30 Mullamullang Cave, Nullarbor
Mainomo, Central Highlands, N.G., Plain, Australia, 165
52 Muriraga, Central Highlands, N.G.,
Mair’s Cave, Flinders Ranges, S.A., 49
203, 206
Malham Cove, Craven, 91, 115
Malham Tarn, Craven, 69, 115
Mammoth Cave, Kentucky, 129 Napier Range, W.A., 15, 138
Manavgat R., Turkey, 80 Narrengullen Cave, Yass, N.S.W.,
Manifold R., Peak, 71, 72, 117-18 203, 221
Manus L, N.G., 4 Natural Bridge, Cedar Creek, Vir­
Marakoopa Cave, Mole Creek, Tas., ginia, 102, 103
147 Natural Tunnel, Stock Creek, Vir­
marble, 11, 40, 44, 185 ginia, 102
Mären Mountains, Switzerland, 59 Nera R., Apennines, Italy, 66, 67
marl, 10 Neretva R., Yugoslavia, 64, 143, 218
Masocha, Moravia, Czechoslovakia, Niagara, Ontario, Canada, 52
145 nickpoint, 118, 120
Maungawharawhara Cave, N.Z., Norber Crags, Craven, 58
100 , 102 North Downs, England, 119
May R., Konya, Turkey, 226, 227 Nullarbor Plain, Australia, 35, 36 ,
McKinstry’s Canyon, Guadalupe 38, 40, 46, 52, 57, 62, 64, 97,
Range, New Mexico, 107 125, 144, 165, 182- 3 , 184, 214
 Index 249

Officer's Cave Ridge, Oregon, 5 plasticity, 34-5

Ogof Ffynnon Ddu, S. Wales, 165 Plitvice Lakes, Korana R., Croatia,
Ombla, Dubrovnik, Dalmatia, 78, 106, 107, 108
226 plunge pool, 149, 150. 153
oolite, 13, 16 pocket, 157
oosparite, 14 Pointe de la Jaquette, Bas Queyras,
ooze, see lime mud France, 31 , 53
orthovacuum, 166 polje. 90-1, 93, 136-43, 163 , 190,
overland flow, 61, 63, 114 194, 200; open, 139; semi-closed,
Owen Stanley Range, N.G., 2 139; tectonic, 139, 141 ; erosional,
Owenterbolea, Clare, 84, 116 140 , 141-2, 200; karst margin,
oxbow, 153 142-3
polje lake, 91
Ozark Mountains, Missouri, 177
ponor, 139, 142
pool deposit, 173
Padirac, Causse de Gramat, France, Popovo Polje, Hercegovina, 139
145 porosity, 17, 53, 84, 192; inter­
palaeokarst, 4 granular, 19, 20, 61, 144
Pamukkale, Turkey, 228 Porta di Prada, La Grigna, Italy,
paravacuum, 166 102
partition, 157 Porth-yr-Ogof, S. Wales, 83
Pazinski Potok, Slovenia, 114-15 Postojna Cave, Slovenia, 6, 168-9,
Peak, 39-40, 117-18, 168, 181 171, 227
Peak Cavern, Peak, 168 pothole, 46-7, 120, 145, 146, 147,
pedimentation, 194, 199, 200 148 , 151-3, 164, 166, 185 , 186,
Pegsdon, Hertfordshire, England, 201, 203; absorption, 152; col-
119 lapse, 153, 165; resurgence, 92,
pellets, 13, 16 163 , 164
Pennsylvania, U.S.A., 66, 166 Poulnagallum Cave, Clare, 169
Perlis, Malaysia, 116, 192 precipitation (of calcite), 2, 28-9, 40,
permeability, 2, 17-18, 78, 90, 96; 171
Prices Falls, Falls Creek, Oklahoma,
primary, 19; secondary, 20 109
permafrost, 119, 184-5, 222 Provence, France, 219
Pescara R., Apennines, Italy, 66 pseudokarst, 5
phreas, 90 Puerto Rico, 188 , 191. 192, 201
Piccaninny Blue Lake, S.A., 125 Pulau Dayang Bunting, Langkawi
Picznice Cave, Hungary, 163 , 164 Is., Malaysia, 192
piezometric surface, 89 Punchbowl, Wee Jasper, N.S.W.,
Pine Creek, Arizona, 109 124
piping, 5, 30, 32 Punchbowl-Signature Caves, Wee
Pivka R., Slovenia, 69 Jasper, N.S.W., 147, 165
Piu Piu Spring, Takaka, N.Z., 76, Punkva R., Moravia, Czechoslova­
77 kia, 181
plain: alluviated limestone, 189 ,
190-3, 201; corrosion, 114 , 217; quartz diorite, 2
karst margin, 143, 190, 217 quartzite, 5
plan, 136 rainfall intensity, 61, 188
planes of weakness, 17, 19, 20, 61, rainpit, 42, 153
121, 144, 155, 166-7 Rak R., Slovenia, 70
Planina Polje, Slovenia, 138 Rakova Kotlina, Slovenia, 100. 102
250 Index

Ras el ’Ain, Syria, 80 sandstone, 64, 114. 125, 151, 152

red clay, 66, 156, 176-7 Sarawak, 43, 51, 191, 201, 202
red earth, 188 Sarthe, France, 57
reef: backreef facies, 16, 142, 200; saturation equilibrium, 24-7, 56, 83,
forereef facies, 15, 199; primary 186, 193
relief, 196-7 scallops, see current markings
reef facies, 14-16 Schleichende Brunnen, Muototal,
reef knolls, 4 Switzerland, 78
rejuvenation, 117-18, 215-17 Sea of Argostoli, Greece, 75, 79
rejuvenation head, 117-18, 216, 217 seepage, vadose, 97
Reka R., Slovenia, 81 self-capture, 102, 104- 5. 146
relief: depositional karst, 196-9; Serpentine Cave, Jenolan, N.S.W.,
structural karst, 199-200 167
rendzina, 62-3, 123, 187 shakehole, 126
Rennell I., Solomon Is., 197 shale bands, 209
rest level, 78, 93 Shannon R., Ireland, 68
resurgence, 74, 83 Shatter Cave, N.S.W., 20
shillow, 47, 185
Rieseneishöhle, Austria, 174
Rift Pot, Craven, 203 Sierra de los Organos, Cuba, 23. 57
Rijeka, Croatia, 78 Sierra Nevada, California, 164
rimstone dam, 173 Silikatkarren, 5
rising, 76 Silver Spring, Florida, 81
rising coefficient, 79 sinkhole, see doline
siphon, 77, 160, 170
rising density, 79
river: allogenic, 64, 114, 117, 118; Slanic, Romania, 53
autogenic, 66; constructional slope processes, 32-3, 100, 121, 125,
action of, 85, 105-9; efficiency of, 176
66-7; sinking of, 69-71 Slovenia, 112, 168, 181
river capture, 102, 104-5 smoke, effect on solution, 56
river erosion, 67-8, 83, 99-100 snow, effect on solution, 40, 45,
river regime, 66-7 185
river solute load, 69-70 snow drifting, 123-4, 152, 174
river transport, 69 Sogöksu, Turkey, 113
rock mill, 149, 153 soil, 40. 47; residual, 3. 32, 62, 66,
rock pendant, 157 123, 126; transported, 63
rock salt, see halite soil air, 26-7
root depth, 66 soil erosion, 21, 49
root grooves, 51
root respiration, 27, 51 Solo R., Java, 115. 116
Rottnest I., W.A., 198 solution, 2, 23-8, 30, 149-51, 166;
phreatic, 155-64; rate of super­
runnel, 54, 55, 61; meandering, 45; ficial, 54-5; rate of underground,
rain solution, 44, 45; rounded 168-70; subsoil, 21, 124; vadose.
solution, 50, 54; solution, 44; 151-5
undercut solution, 48; wail solu­
tion, 45 solution bevel, 44
solution cup, 48
solution flute, 42, 43, 44. 45, 47,
Sadeng, Java, 115 48, 61
St Brendan’s Well, Clare. 84, 117 solution funnel step, 44, 45 , 186
St Cuthbert’s Swallet, Mendip, 170 solution lake, 159
salt crystallisation, 38 solution notch, 48, 49, 191, 192
 Index 251

solution pan, 47, 54, 62 Tanana R., Alaska, 181, 185, 186
solution pedestal, 56, 59 Tarn R„ Grandes Causses, France,
solution pipe, 33, 50-1, 53,122, 123, 67, 98
127, 199 Te Kuiti, N.Z., 41
solution pit, 53 tectonism, 197, 215-18, 219
solution ripple, 43, 45 Tennessee Valley, U.S.A., 97
Somerset I., Canada, 57, 181, 185 terra rossa, 63, 66
Source des Cent Fons, Herault, Thames R., England, 69
France, 86 thaw-freeze action, 23, 38, 40, 177-
South Algeria, 181 8, 185, 194
speleogen, 147, 148, 154-5, 158, 160- thermokarst, 5
1, 178 throughflow, 61, 63, 74, 114
speleothem, 162, 171-4, 184, 186,
Tiber R„ Italy, 67
199, 215; isotopic studies of,
178; submerged, 198 Tibera Lake, N.G., 140
spongework, 156 Timavo R., San Giovanni, Italy, 81
spore drift, 9, 93-4, 95 Totes Mountains, Austria, 94
spring sapping, 112, 119 tower, 190-3, 194, 200, 201, 202,
springs, 69-79, 82, 87; artesian, 80; 204, 218
boiling, 77, 81; location of, 78, Transvaal, S. Africa, 215, 226-7
91,96; solute contents of, 68 , 83-4, travertine, 106-9
169-70; sublacustrine, 78; sub­ triggering, 36
marine, 75, 78-9; vauclusian, 75, Trobriand Is., N.G., 197
76 tube, 161
stalactite, 162, 172-3; straw, 172 tufa, 105
stalagmite, 162, 172
tufa dam, 106-9
steephead, 112-13
Stella R., Castella Sacile, Italy, 81 Tumwalau Cave, Kiriwina I., Tro­
Stockyard Gully Cave, W.A., 100 briand Is., N.G., 198
streamsink, 70, 71, 72, 142, 153, 217 Tunnel, Napier Range, W.A., 142-3
streamsink density, 74 Tunnel Cave, Borenore Creek,
streamsink ordering, 71-4 N.S.W., 156
streamsink regime, 74 tunnelling, see piping
strength: mechanical, 17, 18-19, 35; Turner Falls, Honey Creek, Okla­
tensile, 18; compressive, 18; homa, 109
shear, 34
Strickland R., N.G., 98 underground geomorphology, 2
stylolites, 11, 22, 52-3 uvala, 135-6
submarine inflow, 6 8 , 79
subsidence, 5, 32-3, 126-7 valley: blind, 109-12, 120, 142; dry,
Svartisen, Norway, 181, 186 1, 19, 85, 114-20, 130, 131,
swallet, see streamsink 222; hanging, 118; inland, 143;
swamp slot, 48, 192 pocket, 112; semiblind, 109, 110,
Swildon’s Hole, Mendip, 170 116
system: climato-morphogenic, 179, valley threshold, 109, 1 16
194-5; CCb-H.O-CaCCL, 24-9; vegetation, effects of, 63-4, 105,
hydrogeologic, 207, 208, 209 168-9, 195
Velebit Mountains, Croatia, 122
Verandah Cave, Borenore Creek,
Tabasco, Mexico, 143, 192 N.S.W., 100, 101
tafoni, 41 Vercors, France, 181, 186
Takaka R„ N.Z., 70 Vidourle R., Provence, France, 81
252 Index

Vorfluter, 142, 190 West Virginia, 166

Vrana Lake, Cres, Croatia, 78 White Scar Cave, Craven, 75
Vung Ha Long Bay, Vietnam, 219- Windjana Gorge, Lennard River,
20 W.A., 99
Wonderland, Grampian Ranges,
Waitomo Caves, N.Z., 178 Vic., 5
Walkenried, Harz, Germany, 51 Wyanbene Cave, N.S.W., 126, 151
wang, 136
water tracing, 9, 168 Yucatan, Mexico, 78, 125, 181, 187
watertable, 88-9, 93 Yunnan, China, 214
watertable lowering, 119-20
watertable stream, 158 zanjones, 202
Watlowes, Craven, 115 zone: intermediate, 89, 90, 158; of
weathering, cavernous subsoil, 50 aeration, 88; of saturation, 78,
Wee Jasper, N.S.W., 27, 71, 164 88; phreatic, 90, 155, 158;
West Irian, 192 vadose, 89, 90, 151
 ABOUT THE AUTHOR
J. N. Jennings is a graduate of Cambridge University and for
nineteen years has been a member of staff of the Australian
National University where he is now Professorial Fellow in
Geomorphology. A past president of the Australian Speleological
Federation, he has examined caves and associated landforms in
many parts of Australia, in New Zealand, New Guinea, and
Malaysia, as well as in the United States, Europe, and Turkey.
His interest in this kind of country goes back to his boyhood
spent near the karst of Craven, Yorkshire.
Mr Jennings, who is co-editor of Landform Studies from
Australia and New Guinea, is general editor of the series, An
Introduction to Systematic Geomorphology. His own geomor-
phological papers range over many kinds of scenery besides
karst, including coasts, lakes, and cold climate landforms.

Book designed by Arthur Stokes.

Text set in 10/12 Intertype Times Roman and printed on Matt Prestige
Art paper by Halstead Press Pty Ltd, Sydney.
J. N. Jennings is a graduate of Cambridge
University and for nineteen years has been
a member of staff of the Australian National
University where he is now Professorial
Fellow in Geomorphology. A past president
of the Australian Speleological Federation,
he has examined caves and associated land-
forms in many parts of Australia, in New
Zealand, New Guinea, and Malaysia, as
well as in the United States, Europe, and
Turkey. His interest in this kind of country
goes back to his boyhood spent near the
karst of Craven, Yorkshire.
Mr Jennings, a co-editor of Landform
Studies from Australia and New Guinea,
is general editor of the series, An Intro­
duction to Systematic Geomorphology. His
own geomorphological papers range over
many kinds of scenery besides karst,
including coasts, lakes, and cold climate
landforms.
Jacket designed by Arthur Stokes. Jacket
photograph of Windjana Gorge, Western
Australia, is reproduced by courtesy of the
Western Australian Department of Lands
and Surveys.
Printed in Australia

An Introduction to
Systematic Geomorphoiogy
Seven volumes are at present planned for
this series.
1 Humid Landforms
I. Douglas (1972)
2 Desert and Savana Landforms
J. A. Mabbutt (1972)
3 Landforms of Cold Climates
J. L. Davies (published)
4 Coasts
E. C. F. Bird (published)
5 Structural Landforms
C. R. Twidale (published)
6 Volcanoes
Cliff Ollier (published)
7 Karst
J. N. Jennings (1971)
Further details are available from the A.N.U.
Press, P.O. Box 4, Canberra, A.C.T. 2600,
Australia.