! LIBRARY
UNIVERSITY OF
CALIFORNIA

EARTH

SCIENCES

LIBRARY

PLATE 1. Frontispiece

CHARACTERISTICS

EXISTING GLACIERS

WILLIAM HERBERT HOBBS

PROFESSOR OF GEOLOGY IN THE UNIVERSITY OF MICHIGAN

" The present is the key to the past." SIR CHARLES LYELL

!Nefo gorfe

THE MACMILLAN COMPANY
1911

Set up and electrotyped. Published May, 191 1.

J. S. Gushing Co. Berwick & Smith Co.
Norwood, Mass., U.S.A.

EARTH

SCIENCES

LIBRARY

PROFESSOR VICTOR GOLDSCHMIDT

OF THE

UNIVERSITY OF HEIDELBERG

A LEADER IN SCIENTIFIC RESEARCH

A GIFTED AND INSPIRING TEACHER

AND A NOBLE AND GENEROUS FRIEND

THIS BOOK IS AFFECTIONATELY DEDICATED

THE AUTHOR

334627

PREFACE

IT has been the common practice to treat the subject
of glaciation as if all ice masses having inherent motion
of whatever nature were governed by the same laws.
Thus the most recent and authoritative work upon the
subject has treated the glaciers of Greenland and Switzer-
land together. The aim of the present w^ork has been
rather to emphasize the wide differences in other than
dimensional respects which separate such bodies, and
to show that the laws which govern their nourishment
and depletion, and their reaction with the lithosphere as
well, are by no means identical.

The broad line of cleavage is found to lie between those
glaciers which completely cover a considerable portion
of the rock surface, and have the form of a flat dome or
shield, and the remaining types. These latter glaciers
being all restricted to mountain districts have been desig-
nated mountain glaciers, and they have been found to
bear very simple relations to each other, dependent upon
the measure of their nourishment and waste. Alimenta-
tion being in turn dependent upon climatic conditions,
all are brought in order within the cycle of changes
which correspond to a period of increasingly rigorous
climate followed in turn by more genial conditions the
cycle of glaciation. Throughout the attempt has been
to emphasize the broader physiographic elements of the
problem and to show the relations to alimentation and
depletion.

vii

viii PREFACE

No attempt has here been made to set forth the views
of that school of British geologists particularly which
holds that the denudational effect of glacier ice is nega-
tive, because it protects the basement from the process
of weathering. As will appear from the text, the writer
believes that protection from weathering on the cirque
floor combined with effective w r eathering at the base of
the cirque wall, explains the lateral migration of the
glacial amphitheatre. The doctrine of protection by ice
has been given so recent an exposition by an eminent
prophet of this school with the expressed approval of his
colleagues, that it is believed more is gained from setting
forth the evidence from one's own viewpoint than by
entering into controversy. Even the names "glacial
protection " and " glacial erosion " as applied to the two
schools to-day seem inappropriate.

The materials of this volume are three papers which
have been published at London, Philadelphia, and Berlin
during the year 1910. The first of the series appeared
in the Geographical Journal under the title " The Cycle
of Mountain Glaciation." In a greatly expanded form
it is Part I of the present volume. The remaining parts,
though originally published in technical journals, were
written with a view to their republication in book form,
and have in consequence been less altered. Of these the
earlier appeared in the Proceedings of the American Philo-
sophical Society under the title, " Characteristics of the
Inland-ice of the Arctic Regions" ; while the concluding
part was published a few months later at Berlin in the
international journal of glaciology bearing the title, " The
Ice Masses on and about the Antarctic Continent." To
the Royal Geographical Society of London, the American
Philosophical Society of Philadelphia, and the editor of
the Zeitschrift fur Gletscherkunde, the author is under

PREFACE ix

obligation for permission to republish the papers in their
present form. Although they contain original material
and of necessity make use of technical terms, it is thought
that the language will in the main be intelligible to the
general reader as well as to the specialist in glaciology.

ANN ARBOR, MICHIGAN,

November 2, 1910.

CONTENTS

INTRODUCTION

PAGK

The ancestry of glacial theories The factor of air temperature
Mountain versus continental glaciers Low level versus high level
sculpture References . 1

PART I
MOUNTAIN GLACIERS

CHAPTER I
THE CIRQUE AND ITS RECESSION

The glacial amphitheatre in literature Relation of cirque to berg-

schrund The schrundline Initiation of the cirque, nivation

References 12

CHAPTER II
HIGH LEVEL SCULPTURING OF THE UPLAND

The upland dissected Modification in the plan of the cirque as
maturity is approached Grooved and fretted uplands Charac-
teristic high relief forms of the fretted upland The col and its
significance The advancing hemicycle References ... 25

CHAPTER III

CLASSIFICATION OF GLACIERS BASED UPON COMPARATIVE
ALIMENTATION

Relation of glacier to its bed Ice-cap type Piedmont type
Transection type Expanded-f oot type Dendritic or valley type
Inherited basin type Tide-water type Radiating (Alpine)

type Horseshoe type References 41

xii CONTENTS

CHAPTER IV
Low LEVEL GLACIAL SCULPTURE IN MODERATE LATITUDES

PAGB

The cascade stairway Mechanics of the process which produces the
cascade stairway The U-shaped glacier valley The hanging
side valley References ......... 59

CHAPTER V
HIGH LATITUDE GLACIAL SCULPTURE

Variations in glacial sculpture dependent upon latitude Surface
features of Northern Lapland The flatly grooved valleys and
the scattered knobs The fjords of Western Norway The rock
pedestals bounded by fjords The Norwegian find References . 70

CHAPTER VI
GLACIAL FEATURES DUE MAINLY TO DEPOSITION

Abandoned moraines of mountain glaciers The tongue-like basin
before the mountain front Border lakes Stream action on the
mountain foreland The outwash apron Eskers and recessional
moraines Stream action within the valley during retirement of
the glacier Landslides and rock streams within the vacated
valley Rock flows from abandoned cirques References . . 81

PART II

ARCTIC GLACIERS

CHAPTER VII
THE ARCTIC GLACIER TYPE

Introduction North and south polar areas contrasted The fixed
areas of atmospheric depression Ice-caps of Norway Ice-caps
of Iceland Ice-covered archipelago of Franz Josef Land The
smaller areas of inland-ice within the Arctic regions The inland-
ice of Spitzbergen The inland-ice of Grinnell, Ellesmere, and
Baffin lands References 97

CHAPTER VIII
PHYSIOGRAPHY OF THE CONTINENTAL GLACIER OF GREENLAND

General form and outlines The ice face or front Features within
the marginal zone Dimples or basins of exudation above the

CONTENTS xiii

PAGE

marginal tongues Scape colks and surface moraines Marginal
moraines Fluvio-glacial deposits References .... 119

CHAPTER IX
NOURISHMENT OF THE GREENLAND INLAND-ICE

Few and inexact data Snowfall in the interior of Greenland The
circulation of air over the isblink Foehn winds within the coastal
belt Wind transportation of snow over the desert of inland-ice
Fringing glaciers formed from wind drift Nature of the sur-
face snow of the inland-ice Snowdrift forms of deposition and
erosion, sastrugi Source of the snow in cirrus clouds Refer-
ences . . 143

CHAPTER X
DEPLETION OF THE GREENLAND ICE FROM SURFACE MELTING

Eastern and western slopes compared Effect of the warm season
within the marginal zones of the inland-ice Differential surface
melting of the ice Moats between rock and ice masses Engla-
cial and subglacial drainage of the inland-ice The marginal
lakes Ice dams in extraglacial drainage Submarine wells in
fjord heads References 1(52

CHAPTER XI
DISCHARGE OF BERGS FROM THE ICE FRONT

The ice cliff at fjord heads Manner of birth of bergs from studies in
Alaska Studies of bergs born of the inland-ice of Greenland
References 178

PART III
ANTARCTIC GLACIERS

CHAPTER XII
THE ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE

General uniformity of conditions in contrast with the north polar
region Antarctic temperatures Geographical results of ex-
ploration The submerged continental platform The zone of
sea and pack ice The ice islands and ice-foot glaciers Refer-
ences 186

xiv CONTENTS

CHAPTER XIII
THE MARGINAL SHELF ICE

PAGE

Its nature and distribution The " Great Ross Barrier," Victoria Land

The " higher " and " lower " ice terraces off King Oscar Land
The u west-ice " of Kaiser Wilhelrn Land The ice barrier tongues
of Victoria Land The rectangular table berg of Antarctic waters

References 214

CHAPTER XIV
THE ANTARCTIC CONTINENTAL GLACIER WHERE UNCONFINED

The inland-ice margin on Kaiser Wilhelm Land The blue icebergs

of Antarctica Origin of the West-ice References . . . 245

CHAPTER XV

THE ANTARCTIC CONTINENTAL GLACIER WHERE BEHIND A
MOUNTAIN RAMPART

The inland-ice of Victoria Land Marginal sections along the outlets

Dimples of the ice surface above the outlets Ice aprons below
outlets Moats surrounding rock masses Mountain glaciers on
outer slope of the retaining ranges Ice slabs References . 253

CHAPTER XVI

THE NOURISHMENT OF ANTARCTIC ICE MASSES

The Greenland ice in its relation to the Antarctic continental glacier

Air temperatures, humidity, and isolation Nature of the snow

precipitated in Antarctica Winds upon the continental margins

The Antarctic continental (glacial) anticyclone Wind direc-
tion determined by snow-ice slope The southerly f oehn-blizzard
of the ice plateau Wind transportation of snow High level
cirrus clouds the source of the snow in the interior of Antarctica

Former extent of Antarctic glaciation References. . . 261

AFTERWORD

The two contrasted glacier types Physiographic form Denuding

processes Alimentation Marginal contours .... 285

LIST OF PLATES

1. The Bishops Glacier in the Bishops Range of the Selkirks, with

the Purity Range beyond Frontispiece

FACING PAGE

2. A. Summer snow bank surrounded by a brown border of finely

comminuted rock, Quadrant Mountain, Y. N. P. . . . 20
B. Snow bank lying in a depression largely of its own construc-
tion, Quadrant Mountain, Y. N. P.

3. A. View of the Yoho Glacier at the head of the Yoho Valley,

Canadian Rockies 26

B. Pre-glacial upland on Quadrant Mountain, invaded by the
cirque known as the " Pocket "

4. Maps to illustrate progressive dissection of an upland ... 28

i. Early stage of glaciation 2. Further investment of the
upland to produce a grooved upland. 3. Early maturity.
4. Complete dissection at maturity producing a fretted upland.

5. Multiple secondary cirques on the west face of the Wannehorn seen

across the Great Aletsch Glacier ....... 28

6. A. a. A grooved upland in the Bighorn Mountains, Wyoming;

b. A fretted upland, Alaska 30

B. Multiple cirque of the Dawson Glacier seen from the Asulkan
Pass, Selkirks

7. A. Fretted upland of the Alps as seen from the summit of Mont

Blanc 30

B. Map of a portion of one of the Lofoten Islands, showing a
fretted surface in part submerged and emphasizing the
approximate accordance of summit levels

8. A karling in North Wales 32

9. A. The Matterhorn from the Gorner Grat, near the Riffelhorn . 34
B. Col of the Overlook looking across the foot of the Illecillewaet

Glacier, Selkirks

10. A. Expanded forefoot of the Foster Glacier, Alaska ... 44
B. Type of piedmont glacier

11. Types of mountain glaciers ........ 48

12. A hanging glacieret, the Triest Glacier above the Great Aletsch

Glacier of Switzerland . 50

13. A. A hanging tributary valley meeting a trunk glacier valley

above the present water level on the " inside passage " to

Alaska . . .52

B. Irregularly bounded nves upon the volcanic cone of Mt. Ranier

xvi LIST OF PLATES

PLATE FACING PAGK

14. A. Series of hanging glacierets which extend the Asulkan Glacier

in the Selkirks 54

B. View of the Wenkchemna Glacier in the Canadian Rockies

15. Surface moulded by mountain glaciers near the ancient Lake Mono

in the Sierra Nevadas of California 60

16. A. The Little Cottonwood Canyon in the Wasatch Range trans-

formed at the bottom into the characteristic U -section . 64
B. Striated surface of glaciated valley floor near Loch Coriusk,
Skye

17. A. The Hardangerjb'kull and the Kongsnut nunatak ... 76
B. Upland glaciated by mountain glaciers and partially sub-
merged through depression

18. A. Development of tinds on the margin of the Jostedalsbraen . 78
B. Typical tinds on the margin of a fjord

19. A. Rock stream in a cirque of Greenhalgh Mountain, Silverton

quadrangle, Colorado 96

B. Rock stream at head of a cirque, in the silver basin, Silver-
ton quadrangle, Colorado

20. Map of a portion of the Jostedalsbraen ...... 102

21. Map of the margin of an Icelandic ice-cap 104

22. A. Fretted upland carved by mountain glaciers about King Oscar's

Fjord, Eastern Greenland 124

B. Front of the Bryant glacier tongue, showing the vertical wall
and the stratification of the ice

23. A. Portion of the southeast face of the Tuktoo glacier tongue,

showing the projection of the upper layers apparently as a

result of overthrust 128

B. Ice face at eastern margin of the inland-ice of Greenland in
latitude 77 30' N.

24. A. Normal slope of the inland-ice at the land margin near the

Cornell tongue 130

B. Hummocky moraine on the margin of the Cornell glacier
tongue

25. A. Lateral glacial stream flowing between ice and rock, Benedict

glacier tongue, Greenland ....... 170

B. The ice-dammed Lake Argentine in Patagonia

26. A. Ice-dammed lakes on the margin of the Cornell tongue of the

inland-ice .......... 174

B. Delta in one of the marginal lakes of the Cornell glacier
tongue

27. A. The fringing glaciers about Sturge Island, Balleny Group . 208
B. An ice-foot, with boat party landing

28. A. The ice-sheathed Bouvet Island, latitude 54 26' S., longitude

3 24' E. (after Chun) 208

B. Neve stratification in ice island (after Arctowski)

LIST OF PLATES Xv ii

PLATE FACING PAGE

29. A. The margin of the Great Ross Barrier 216

B. Near view of the Great Ross Barrier where highest 280 feet

30. A. Surface of the great shelf ice to the south of Minna Bluff . 220
B. Surface of the great shelf ice viewed from a balloon and show-
ing sastrugi

31. A. A new ice-face on the Great Barrier ...... 222

B. An old ice-face on the Great Barrier

32. View of the inland-ice of Kaiser Wilhelm Land from the top of

the Gaussberg 246

33. A. View of the Gaussberg surrounded by inland-ice in a depressed

zone 258

B. Moat surrounding rock which projects from the ice surface

34. A. View of the high surfaces of the Jotemheim from the Galdha,

Norway (after Fritz Machacek) 286

B. The Maelkevoldsbrae of the Jostef jeld, showing the develop-
ment of tinds about the borders of a Norwegian plateau
glacier (after Fritz Machacek)

ILLUSTRATIONS IN THE TEXT

FIG.

1. Ideal section across inland-ice ........ 7

2. Section across a mountain upland occupied by glaciers ... 7

3. View/of the ice-cap of the Eyriksjokull, Iceland .... 7

4. A glacial cirque excavated from the Pleistocene glaciated surface

of Norway . . . . . . . . . .14

5. Bergschruud below cirque wall on a glacier of the Sierra Neyadas,

California ........... 16

6. Schrundline near Mt. McClure, Sierra Nevadas of California . 18

7. Cross-section of a steep snowdrift site, showing recession by niva-

tion ............ 19

8. Characteristic form of drift sites on hillsides in Swedish Lapland . 21

9. Pre-glacial upland invaded by cirques, " biscuit cutting " effect,

Bighorn Mountains ......... 26

10. View of the scalloped tableland within the Uinta Range . . 27

11. Map of Quadrant Mountain, a remnant of the pre-glacial upland

on the flanks of the Gallatin Range, Y. N. P ..... 27

12. Series of semicircular glacial amphitheatres whose scalloped crest

forms part of the divide of the North American continent . 28

13. Diagram to illustrate the manner of dissection of an upland by

mountain glaciers . . . . . . . . .31

14. Position of the Aletschhorn and Dreieckhorn between the Upper,

Middle, and Great Aletsch neves ....... 33

15. Illustration of the formation of cols through the intersection of

cirques ............ 34

16. Map of a transection glacier ........ 45

17. The Baird glacier, a typical expanded-foot glacier . . .46

18. Outline map of the Hi spar glacier, Himalayas . . . .47

19. Outline map of the Tasman glacier, New Zealand . . 48

20. Outline map of an inherited basin glacier ..... 49

21. Outline map of a reconstructed glacier ...... 50

22. Outline plan of a radiating glacier ....... 53

23. Outline map of the Asulkan Glacier in the Selkirks ... 54

24. Outline map of the Wenkchemna Glacier in the Canadian Rockies 55

25. Longitudinal section along a glaciated mountain valley, showing

reverse grades and rock basin lakes in series .... 60

26. Rock bar with basin showing above ...... 62

27. Ideal cross-section of a U-shaped valley once occupied by a moun-

tain glacier . . . . . . . . . . .64

xix

XX ILLUSTRATIONS IN THE TEXT

FIG.

28. View in the glaciated Sierra Nevadas of California, showing the

sharp line which sometimes separates the zone of erosion from
that of sapping 65

29. Normal valleys from sub-aerial erosion 66

30. Glaciated and non-glaciated valleys tributary to a glaciated main

valley hanging valleys . . ... .07

31. Comparison of the longitudinal profile of a mature stream-cut

valley and its tributaries with a glacier-carved Alpine valley
and its tributaries . ., . t . v : . . . 68

32. Surface in Swedish Lapland moulded by continental glaciers and

subsequently grooved by sluggish mountain glaciers ... 71

33. Map of area in Swedish Lapland, showing cirques and kar-

lings . . . . . ... . ; 72

34. Map of area in Swedish Lapland moulded by sluggish glaciers

which succeeded continental glaciation . . . . . 73

35. Characteristic features due to glacial sculpture in Scandinavia . 74

36. Map of the vicinity of the Storf jord, showing the regular arrange-

ment of fjords and submerged valleys . . . . 75

36 a. Nunataks rising out of the surface of a Norwegian ice-cap near

its margin . . . . . . . . .76

37. Erosional surface due to ice-cap glaciation within the marginal

zone . . . . . . . ...... 76

38. The Seven Sisters, sharpened ice-cap nunataks in Northwestern

Norway due to overflow of glacier streams at margins /. . 77

39. Broad glacial trough overdeepened through uplift and subsequent

glaciation .... . .77

40. Circular tind with acute apex from the Lofoten Islands . . 78

41. Successive diagrams to illustrate a theory of the shaping of acute

circular tinds through exfoliation 79

42. Terminal and lateral moraines remaining from earlier mountain

glaciers . . . . . *. . . . .81

43. Sketch map of the morainic ridges near the mouth of Little Cot-

tonwood Canyon in the Wasatch Range 82

44. Convict Lake, a lake behind a morainal dam in a glaciated valley

of the Sierra Nevadas of California 82

45. Map of the moraines and drurnlins within and about the apron of

the piedmont glacier of the Upper Rhine . . . . . 83

46. Lake Garda in the southern gateway to the Alpine highland on

the apron site of the earlier piedmont glacier . . ... 84

47. Outline map of the northern border of the Alps, showing the

basins of former lakes 85

48. A braided stream flowing from the margin of a glacier ... 86

49. Ideal form of tongue-like basin remaining on the site of the apron

of a piedmont glacier 88

50. Gorge of the Albula river near Berkun in the Engadine . . 90

ILLUSTRATIONS IN THE TEXT xxi

FTG.

51. Ideal section showing successive slides from a canyon wall so as

to produce a staircase effect ........ 92

52. View of the succession of rock slides from the north rock wall of

the Upper Rhine near the town of Flims ..... 93

53. Map of two high glacial cirques now partially occupied by rock

streams ............ 95

54. Map showing the areas of fixed low barometric pressure and of

heavy glaciation in the Northern Hemisphere .... 100

55. Idealized section showing the form of "fjeld" and "brae" in

Norwegian ice-cap ......... 101

56. Maps of the Hofs Jokull and the Lang Jokull . . . .102

57. Map of the Vatna Jokull ......... 103

58. Cross-section of the Vatna Jokull from north to south . . . 104

59. Map of the ice-capped islands in the eastern part of the Franz

Josef Archipelago ......... 107

60. Typical ice cliff of the coast of Prince Rudolph Island, Franz Josef

Land ............ 108

61. Map of Nova Zembla, showing the supposed area covered by

inland-ice ........... 109

62. Map of Spitzbergen, showing the supposed glacier areas . .110

63. Inland-ice of New Friesland as viewed from Hinloopen Strait . Ill

64. Map of the southwestern margin of an extension of the inland-

ice of New Friesland ......... 112

65. Camping place in one of the "canals" upon the surface of the

inland-ice of North East Land ....... 114

66. Hypothetical cross-section of a glacial canal upon the inland-ice

of North East Land ......... 115

67. Map showing the supposed area of inland-ice upon Grinnell and

Ellesmere Lands ... ..... 115

68. View of the "Chinese Wall" on Grinnell Land . . . .116

69. Map showing the supposed area of inland-ice upon Baffin Land . 11?

70. Map of Greenland, showing the outlines of the inland-ice and the

routes of explorers . . . . . . . .120

71. Route of Garde across the margin of the inland-ice of South

Greenland ........... 121

72. Sketch of the east coast of Greenland, showing the inland-ice and

the work of marginal mountain glaciers ..... 122

73. Section across the inland-ice of Greenland near the 64th parallel

of latitude ........... 122

74. Comparison of the several profiles across the margin of the inland-

ice of Greenland . ....... 123

75. Map of the region about King Oscar's and Kaiser Franz Josef

Fjords, eastern Greenland ........ 124

76. Map of a glacier tongue which extends from the inland-dee of

Greenland down the Umanak Fjord .... 125

xxii ILLUSTRATIONS IN THE TEXT

FIG. PAGE

77. Tongues of ice which descend from the Foetal glacier . . . 126

78. Map of the Greenland shore in the vicinity of the Northeast

Foreland 127

79. A series of parallel crevasses on the inland-ice of South Green-

land 129

80. Rectangular network of crevasses on the surface of the inland-

ice of South Greenland 130

81. Map showing routes of sledge journeys in North Greenland . 133

82. a. Closure of the Neu-Haufen Dyke, Schiittau ; b. Scape colks

near Dalager's Nunataks ........ 136

83. Diagram to show the effect of a basal obstruction in the path of

the ice near its margin 139

84. Surface marginal moraines of the inland-ice of Greenland . . 139

85. Diagram to illustrate the air circulation over the isblink of

Greenland ........... 147

86. On the Sahara of snow ......... 151

87. Sastrugi on the inland-ice of North Greenland .... 155

88. Barchans in snow 156

89. Diagrams showing the distribution of temperatures within the

surface zones of the inland-ice . . . . . . .164

90. Map showing the superglacial streams within the marginal zone

of the inland-ice 165

91. Diagrams to show the effects on differential melting on the ice

surface '. 166

92. Fragments of rock of different sizes to show their effect upon

melting ........... 167

93. Section showing the so-called " cryoconite holes " upon the surface

of an ice hummock 168

94. Map showing the margin of the Frederikshaab ice apron extend-

ing from the inland-ice of Greenland, and showing the position

of ice-dammed marginal lakes ....... 171

95. Diagram showing arrangement of shore lines from marginal lakes

to the northward of the Frederikshaab ice tongue if its front
should be retired ......... 172

96. Sections from the inland-ice through the Great and Little Kara-

jak tongues to the Karajak Fjord 179

97. Origin of bergs as a result especially of wave erosion . . .180

98. Supposed successive forms of a tide-water glacier front . .181

99. Large berg floating in Melville Bay and surrounded by sea-ice . 182

100. Map of Antarctica, showing the principal points which have been

reached by exploring expeditions .194

101. Map of the Antarctic region, giving the tracks of vessels and

the margins of the continent , 195

102. Soundings over the continental platform to the westward of West

Antarctica . . . ... v 197

ILLUSTRATIONS IN THE TEXT xxiii

r &wi

103. Cracks formed on the free surface of an elastic block when crushed

in a testing machine 201

104. Open lane of water within the Antarctic pack-ice, showing the

minor elements of similar form which by separating yield

* 909

zigzag margins ..... ^ u ^

105. Lozenge-shaped lakes within the Antarctic pack arranged en

echelon ....... 20-

106. Sastrugi on pack-ice off Kaiser Wilhelm Land as seen from a

balloon 204

107. Pressure lines upon the surface of sea-ice 204

108. Pressure ridge formed on the shore of Victoria Land . . . 205

109. The Antarctica sinking after being crushed in the pack . . 205

110. Domed ice island off King Edward Land . . . . . . 209

111. King Edward Land with ice shelf in front 215

112. View of the shelf-ice of Coats Land 216

113. Map of the Ross Barrier edge 217

114. Section along the Ross Barrier edge, showing submerged portion

and the underlying water layer 217

115. a. Low margin of Ross Barrier on Balloon Inlet ; b. Relatively

high margin of the Barrier on Balloon Inlet .... 219

116. Outline map of the known portions of the Great Ross Barrier . 220

117. Map of the shelf -ice near King Oscar Land 225

118. West-ice seen from the " Gauss " off Kaiser Wilhelm Land . 228

119. Junction of the " West-ice " and the " sea-ice " 228

120. Diagram showing manner of formation of " West-ice " mass . 230

121. Map of the glaciers and shelf-ice tongues about the head of

Robertson Bay, Victoria Land 230

122. Map showing the shelf-ice tongues on the west of Ross Sea with

the glacier outlets above them 232

123. Ideal section through a shelf -ice tongue 233

124. The ice barrier breaking away to form a tabular and rectangular

berg 235

125. Rectangular and tabular berg of Antarctic waters . . . 236

126. Tabular Antarctic iceberg, showing perpendicular and rectangu-

lar jointing 236

127. View of a tilted tabular berg, showing the rectangular lines in

the plan 237

128. The inland-ice of Kaiser Wilhelm Land seen from the sea . . 246

129. Intersecting series of fissures in the surface of the inland-ice of

Kaiser Wilhelm Land 247

130. Section across the margin of the inland-ice of Victoria Land to

the westward of McMurdo Sound 253

131. a, b. Section across the Great Ross Barrier and up the Beardmore

Outlet to the ice plateau ; c. Section across the Drygalski ice-
barrier tongue and up the Backstairs Passage to the inland-ice 254

xxiv ILLUSTRATIONS IN THE TEXT

132. A comparison of sections across the margin of the Greenland

and Antarctic continental glaciers ...... 255

133. View from above the Ferrar Outlet, showing the dip of the sur-

face from indraught of the ice ....... 256

134. Map of the Beardmore Outlet 258

135. Map showing sastrugi on David's route to the south magnetic

pole 267

136. Lee side of a sand dune, showing curve of profile . , '..*' . 273

137. Profile across the ice-cap of the Vatna Jokuli .... 274

138. Section of one of the irregular ice grains enveloped in water

which was precipitated together with snowflakes upon inland-
ice of Northeast Land .. .. .. . 277

139. Sketch map showing glaciated and higher non-glaciated surfaces

of the rock masses which protrude through the ice in the
vicinity of McMurdo Sound 279

140. Diagram to illustrate the growth of an inland-ice mass through

the rhythmic action of the anti-cyclonic air-engine . . . 288

CHARACTEEISTICS

EXISTING GLACIERS

CHARACTERISTICS OF EXISTING
GLACIERS

INTRODUCTION

The Ancestry of Glacial Theories. If we are to gauge
the generally accepted hypotheses of any science and arrive
at individual conclusions respecting their value, we must
be prepared to inquire into the ancestry of each we must
trace out the route by which each has come to its present
position of eminence. It is a Scriptural saying that " we
see through a glass darkly ," and scientific reasoning, we
know, makes a demand upon the imagination. To a solid
basis of observation, which at best but half discloses the
truth, inductive reasoning is to be added if science is to
advance.

Psychological processes and the tendencies of scientific
thought are thus to be well considered by the more thought-
ful student of science in forming his opinions. Experience
has shown that whenever a new and more advanced view-
point has been gained to take the place of an earlier one,
and its superiority has come to be acknowledged, the ten-
dency has always been to sketch in from that one standpoint
even the more distant objects, rather than to move forward
to new and independent positions. This has been no less
true of glacier study than of the broader divisions of science.
This general fact is, perhaps, in part to be explained by the
optimism inherent in human nature; but account must also
be taken of the authority of a great name in science.

2 CHARACTERISTICS OF EXISTING GLACIERS

With the multiplication of workers which is characteristic
of present-day science, the number of authorities increases
and the servile attitude within the profession toward its
great leaders will gradually disappear. The student of
geology would, however, do well to take note of the early
dominant influence of Werner or von Buch in Germany, of
de Beaumont in France, of Murchison in England, or of
Agassiz, the " Pope of American Science."

It is a truism that the influence of things seen is more
potent than that of things merely heard of or read about.
The unconscious effect of the immediate environment, of
oft present scenes, in directing the trend of thought, and of
determining convictions, is an unwritten chapter in the
philosophy of science. It would be easy to show how all
the accepted views of geological processes would have been
different had the seats of learning been located either in the
tropics or in polar, rather than temperate, latitudes. More-
over, it has not always been easy to say what observed phe-
nomena are of general and what are of only local importance.
Environment is, therefore, of the utmost importance in the
evolution of the " body of doctrine " of any science.

To apply these considerations to glacial theories, we
find that whereas existing glaciers are found in all latitudes,
but with the largest and most important types in polar and
sub-polar regions; the earliest and by far the largest number
of studies have been made in the Alps, where a single type
of small glacier is found. The reason is not far to seek.
The Alps have now for a good many years been the play-
ground of Europe easily reached and explored by her scien-
tific men.

Until the close of the eighteenth century there existed
a popular belief that the mountain highlands were bewitched.
The Alps were the montagnes maudits and in consequence
a terra incognita. It was de Saussure who both by precept

INTRODUCTION 3

and example as well as by offering a generous prize, stimu-
lated interest in exploring the Alps and thus dispelled the
illusions which had so long clung to them. We may here
pass over his scientific conclusions, as we may over those of
Scheuchzer, Hugi, Venetz, and other early workers, impor-
tant as they were; for it was not until the early forties of
the nineteenth century when Agassiz l and Charpentier 2
published their important monographs upon the physiog-
raphy, the structure, the mechanical work, and the former
extensions of the Alpine glaciers, that a lively interest was
excited in them.

This sudden interest in glaciers on the part of geologists
arose, not so much because of an interest in the Alpine glaciers
themselves, as for the reason that on the basis of these studies
Agassiz soon founded his theory of the ice age, and was thus
for the first time able satisfactorily to explain the origin of
the erratic blocks which are found strewn over the Alpine
foreland, the North German plain, the British Isles, and
Northern North America. The great continental glaciers
which he thus hypothecated were from a thousand to a
millionfold greater than those ice masses which had been
seen and studied, from which ice masses they must have
differed most widely. This is particularly true, as we now
know, as concerns their physiographic development and
their alimentation. No continental glaciers being then
known, it was but natural that the attributes of Alpine
glaciers should have been carried over to the continental
type thus reconstructed in imagination upon the basis merely
of its carvings, its gravings, and its deposits.

It is one of the strange coincidences of science that almost at
the moment when the epoch-making studies of Agassiz were
being made upon Swiss glaciers, three great scientific explor-
ing expeditions were independently discovering the greatest
of existing continental glaciers, that of Antarctica, but with-

4 CHARACTERISTICS OF EXISTING GLACIERS

out being able to set foot upon it or to learn aught of its
characters. Thus it happened that the views concerning
continental glaciers took shape before any had been visited,
and one result is that even to-day in university and college
texts we find the attributes of continental, Alpine, and other
glacier types classified together as though all were necessarily
identical in origin.

The first attempt to arrive at observational knowledge of
" inland-ice " was the expedition of Otto Torell to Spitz-
bergen in 1858. It was the unsuccess of the Swedish polar
expedition of 1872-1873 which made the journey by Norden-
skiold and Palander across Northeast Land (Spitzbergen)
the first successful, comprehensive attempt to observe any
considerable area of inland-ice. It will, however, hardly
be claimed that the results of this expedition are well known,
or that they have in any important way influenced glacial
theories. The later discoveries of Nordenskiold, Nansen,
and above all Peary on the great continental glacier of
Greenland, rich as they are in results, are, moreover, not as
well known as they should be, and are only beginning to
modify the views held concerning continental glaciers.
In fact, it is only toward the beginning of the twentieth
century that former continental glaciers have begun to be
studied on the basis of any other model than the Alps.

The Factor of Air Temperature. With the advance of
knowledge concerning the sequence of conditions affecting
glaciers, it has come to be quite generally recognized that
for any given district the factor of supreme importance in
initiating glaciation is temperature; a very moderate change
in the average annual temperature being sufficient to trans-
form a district, the aspect of which is temperate, and to
furnish it with snow-fields and mountain glaciers. Thus
it has recently been estimated that a fall of but 3 F. in the
average annual temperature of Scotland would result in the

INTRODUCTION 5

formation of small glaciers within the area of the Western
Highlands, while a like fall of 12 F. within the Laurentian
Lake district of North America would be sufficient to bring
on a period of glaciation.

It is further of interest that such temperature changes
affect the distribution of air pressure over the continents and
in this or in other ways directly modify the precipitation of
moisture. Statistics have shown that cold periods corre-
spond to high precipitation and warm periods to smaller falls
of snow and rain. 3 It is further found that the larger cli-
matic changes are common to very large areas of the earth 4
and are probably world wide in their extent. 5

In climates such as now prevail on the borders of Ant-
arctica, it is true that most of the snow falls in the warmer
season. Gourdon has apparently been misled by this into
believing that warm rather than cold climates promote
glaciation. 6 As we shall see, glaciers are under these condi-
tions nourished by a different process, which is in a large
measure independent of local evaporation.

With the probability that such progressive climatic
changes would be initiated slowly, 7 the first visual evidence
of the changing condition within all districts of accentuated
relief would probably be a longer persistence of winter snows
in the more elevated tracts; which accumulation of snow
would eventually contribute a remnant to those of the suc-
ceeding winters, and so bring on a period of glaciation. Such
a change of air temperatures with resultant changes in snow
precipitation may be otherwise expressed as a depression of
the snow-line of the district. All are familiar with the fact
that as we ascend in the atmosphere we pass into succes-
sively colder strata. Mountains which even in tropical
regions push up their heads to great altitudes, are in conse-
quence capped with snow throughout the year. The snow-
line is the lower limit of this " perpetual " snow, and it is

6 CHARACTERISTICS OF EXISTING GLACIERS

evident that any refrigeration of the atmosphere will cause
the line to descend toward the lower levels. 8

From this beginning the process is an advancing one until
a culmination of glaciation is attained corresponding to the
most rigorous of the climatic conditions. A resumption of
a more genial climate would bring about a reverse series of
changes, a waning of the glaciers setting in so soon as the
winter's fall of snow is insufficient to contribute a remnant
to succeeding seasons. It is, therefore, proper to speak of
advancing and receding hemicycles of glaciation. 8o

This use of the expression cycle of glaciation carries with
it no idea of lapse of time except such as is implied in the
completion of a progressive series of climatic changes, and a
return to the initial condition. In any given district the
time may have been insufficient to accomplish the complete
normal series of denudational results indicated in neighboring
districts which were more favored in respect to glacier nour-
ishment. The term " cycle of glaciation " is, therefore,
not the equivalent of " glacial cycle " used by Professor
Davis, 9 since in our use the cycle is measured in climatic
changes rather than in the attainment of certain denudational
effects within the glaciated valleys. Russell's earlier discus-
sion of the " Life History of a Glacier " 10 takes account of
this alternation of sequential climatic changes a climatic
episode with resultant changes in the size and physio-
graphic forms of glaciers.

Mountain versus Continental Glaciers. Those glaciers
which are developed in mountain districts differ from the
ice masses of the interiors of continents or islands in several
important particulars. As respects their physiographic
forms, they are as different as possible. 11 Inland-ice assumes
a form the visible surface of which is largely independent
of the basement upon which it rests, while there is no definite
model to which the glaciers of mountains conform, they

INTRODUCTION 7

being moulded with reference to the irregularities of their
beds. It is characteristic of inland-ice that no portion of the
lithosphere is exposed above its higher levels. The glaciers
of mountains, on the contrary, always have rock exposed above

FIG. 1. Ideal section across inland-ice.

their highest levels. The physiographic form assumed by
inland-ice is invariably that of a flat dome or shield, and all
visible projections of the lithosphere within the area of the
ice are restricted to the marginal zone (see Fig. 1). The
glaciers of mountains, as already stated, conform to no definite
model, and rock projections may appear at any level, but are
always to be seen above the highest levels (see Fig. 2 and pi. 1).

FIG. 2. Section across a mountain upland occupied by glaciers with the glaciers

in black (after Hess).

The unique exception to this law is the small ice-cap or plat-
eau glacier which is transitional between inland-ice and
mountain glaciers (see Fig. 3). In size more nearly allied

FIG. 3. View of the ice-cap of the Eyriksjokull, Iceland, seen from the West

(after Grossman 12 ).

to the glaciers of mountains, in form the ice-cap resembles
the masses of inland-ice it is developed as a flat dome or

g CHARACTERISTICS OF EXISTING GLACIERS

shield. As regards the processes by which they are nour-
ished, ice-caps are, however, as will be seen, quite different
from true inland-ice; and they should in consequence be
considered separately and in order between the others, so
as to call attention to their intermediate position. Their
size is usually a direct consequence of the limitations of the
circumscribed area of the rock platform upon which they
rest usually either a small island or a limited portion of
a high plain or plateau. The regular surface form common to
inland-ice and ice-caps is due to the fact that the irregularities
of the base are small when compared with the dimensions of
the ice mass. The ice-caps of Norway or Iceland have in
common with the glaciers of mountains, a considerable
elevation above the sea, but the variations of their base
from a horizontal plane are small by comparison with the
other dimensions. Curiously enough there is to this rule a
single exception, and here it is not the flatness of the base but
the precipitousness of platform slope which is the determining
factor. This special case is of ice-caps on the high volcanic
peaks of low latitudes, which on excessively steep slopes
push their summits far into the upper atmospheric strata.

Low Level versus High Level Sculpture. In part the
failure to note the essential difference between mountain
glaciers and inland-ice is due to the peculiar evolution of
glacier studies which has been outlined in the introduction,
but in part it is to be explained by a rather general tendency
to treat the subject of erosion by glaciers in mountains from
studies made especially in the lower altitudes. 13 A quite
general neglect of those special conditions of denudation
which are operative in high-level areas of glaciers is, it is
believed, responsible for an over-emphasis laid upon the
U-shaped trunk valley and the hanging tributary valley,
important as these features are. 14 This over-emphasis can,
perhaps, be best illustrated by reference to a series of three

INTRODUCTION ! 9

successive idealistic sketches, executed with great skill by
an eminent American geographer, and intended to develop
especially the erosion forms which result from mountain
glaciers. 15 The low-level sculpturing expressed by these
sketches is, in the opinion of the writer admirable and a
true rendering of nature. It is the failure to recognize any
additional process of erosion operative in higher altitudes
which destroys the value of the high-level sculpturing dis-
played.

So far as low-level mountain glaciation is concerned, the
erosive processes are pretty well understood to be identical
with those of continental glaciers, namely, abrasion and
plucking. The former process is a wearing away of the rock
surface which is in every way analogous to the abrasion of a
facet upon a gem by a lapidary, the stones frozen into the
mass of the ice corresponding to the diamond dust imbedded
in the lap. The product of glacial abrasion is rock flour.
The plucking process, on the other hand, is a removal of the
rock in larger masses aided often by the fracture planes
already present, which so often bound the dislodged masses.
In parts of a glacier bed recently uncovered near the glacier
foot, the dislodged blacks may sometimes be fitted into the
rock floor from which they have been extracted. 16 With
respect to the direction of movement of the ice, abrasion is
particularly developed on obstructing rock masses on the
side from which the ice comes stoss side, and plucking
upon the side away from which it moves the lee side.
The two sides of an obstruction in the bed have therefore
been called the " scour " side and the " pluck" side. 17 The
plucking process is no doubt in some cases much facilitated
by a ready separation of the rock along planes parallel to
the surface, these planes being due to the strains set up in
the rock parallel to its free surface.

To these processes of abrasion and plucking there is in the

10 CHARACTERISTICS OF EXISTING GLACIERS

case of mountain glaciers a third important denuding process
which may locally be more important than both the others
acting together. It is this process of head-wall erosion which
as regards reaction with the lithosphere differentiates all
types of mountain glaciers from continental ones. This
distinguishing process is responsible for the development of
the cirque (Ger. cirkus), which is known by a variety of
names in different glacier districts. In Scotland it has been
generally referred to as the come, in Wales as the cwnij and in
Scandinavia as the botn or kjedel (kessel). In the scientific
literature of the subject the Bavarian- Austrian word "kahr"
has been used with increasing frequency for the same
topographic feature.

In view of this diversity in resultant topography, and
despite their close genetic relationships, we would do well to
sharply separate in our discussions continental glaciers from
the other types, which latter we may include under the broad
term of " mountain glaciers."

REFERENCES

1 L. Agassiz, "Etudes sur les glaciers," Neuchatel, 1840, pp. 1-346.
Accompanied by an atlas of 32 plates. An even more comprehensive
monograph Agassiz published in 1847 under the title, "Nouvelles etudes
et experiences sur les glaciers actuels, leur structure, leur progression, et
leur action physique sur le sol," Paris, 1847, pp. 1-598. With an atlas
of 3 maps and 9 plates (generally referred to as "System Glaciare").

2 Jean de Charpentier, " Essai sur les glaciers et sur le terrain erratique
du bassin du Rhone," Lausanne, 1841, pp. 1-363. Map and plates.

3 Eduard Bruckner, " Klimaschwankungen und Volkerwanderungen im
xix. Jahrhundert," I tern. Wochensch. f. Wissenschaft, Kunst und Technik,
March 5, 1910, p. 6.

4 Siegfried Passage, "Die Kalihari," Berlin, 1904, p. 662. A. Penck,
"Climatic Features of the Ice Age," Geogr. Jour., vol. 22, 1906, pp. 185-
186. Ellsworth Huntington, "Some Characteristics of the Glacial Period
in Non-glaciated Regions," Bull. Geol. Soc. Am., vol. 18, 1907, pp. 351-
388, pis. 31-35. Ellsworth Huntington, "The Pulse of Asia," New York
and Boston, 1907, pp. i-xxi, 1-415. Ellsworth Huntington, "The Libyan
Oasis of Kharga," Bull. Am. Geogr. Soc., vol. 42, 1910, pp. 660-661.

5 Frank Leverett, "Comparison of North American and European
glacial deposits," Zeitsch. f. Gletscherk, vol. 4, 1910, pp. 241-316.

INTRODUCTION 11

6 "It is in fact the proportion of water vapor in the air which con-
trols the greater or less abundance of snow precipitation, so that, as
Tyndall has remarked, it is the solar action which is necessary to bring
on the initial condition; the conclusion, which appears paradoxical at
first, is the following, that the warmest periods determine more active
evaporation of the ocean water ; it is to them that the greatest extensions
of glaciation correspond. Cold plays the merely passive role of condenser."
(Gourdon, Exped. Ant. Frang., 1903-1905, Glaciologie, 1908, p. 70.)

7 1. C. Russell, "Climatic Changes indicated by the Glaciers of North
America," Am. Geol., vol. 9, 1892, p. 336.

8 A. Penck, "Climatic Features of the Pleistocene Ice Age," Geogr.
Jour., vol. 27, 1906, pp. 182-187.

8a William Herbert Hobbs, " The Cycle of Mountain Glaciation," Geogr.
Jour., vol. 36, 1910, pp. 146-163, 268-284, 36 figs.

9 W. M. Davis, "Glacial Erosion in France, Switzerland, and Norway,"
Proc. Bos. Soc. Nat. Hist., vol. 29, 1900, pp. 294-300.

10 1. C. Russell, "Glaciers of North America," 1897, pp. 190-206.

11 E. v. Drygalski says : "The difference between glaciers and inland ice
is essentially a quantitative one. Glacier forms are small, inland ice
masses great glaciations. . . . Inland ice masses are ice overflows of
entire earth surfaces, glaciers are branching outflow systems for snow
deposits guided by the features of the earth's surface." In Keilhack's
"Lehrbuch der praktischen Geologie," 1908, p. 269.

12 Karl Grossmann, "Observations on the glaciation of Iceland," Gla-
cialists' Magazine, vol. 1, No. 2, 1893, pi. 3, fig. 2.

13 "The visitor replied that he was a valley climber, not a mountain
climber. He found sufficient pleasure at the mountain base, and such
was my case also. Mountain tops are indeed worthy objects of a climb-
er's ambition, but if one wishes to get at the bottom facts, let him ex-
amine the valleys." (W. M. Davis, "Glacier Erosion in the Valley of
the Ticino," Appalachia, vol. 9, 1901, p. 137.)

14 On hanging valleys, see especially W. M. Davis, Proc. Bos. Soc. Nat.
Hist., vol. 29, 1901, pp. 273-322; and G. K. Gilbert, "Glaciers," Harri-
man Alaska Expedition, vol. 3.

15 W. M. Davis, "The Sculpture of Mountains by Glaciers," Scot.
Geogr. Mag., vol. 22, 1906, figs. 1-3.

16 Ed. Bruckner, " Die Glacialen Ziige im Antlitz der Alpen," Naturw.
Wochensch., N. F., vol. 8, 1909, p. 792.

17 Penck, "Glacial Features in the Surface of the Alps," Jour. GeoL,
vol. 13, 1905, p. 6.

PART I

MOUNTAIN GLACIERS

CHAPTER I
THE CIRQUE AND ITS RECESSION

The Glacial Amphitheatre in Literature. It is safe to
say that no topographic feature is more characteristic of the
mountains which have been occupied by glaciers than is the
cirque. Approaching a range from a considerable distance,
there is certainly no feature which so quickly forces itself
upon the attention. The U-shaped valley and the hanging
side valley, important as these are, are here decidedly less im-
pressive. Yet the great majority of works upon the subject,
by ignoring the significance of the cirque, allow the reader to
assume that the glaciers discovered the cirques ready formed
to gather in the snows for their nourishment. Even the
standard work of Chamberlin and Salisbury is open to this
objection. 1

Despite the attitude of the general texts, which so largely
determine what might be called the accepted body of doc-
trine of a science, there are a number of papers dealing with
the origin of the cirque. One of the first to recognize the
cirque as a product of glacial erosion was Tyndall, whose keen
mind has so illumined the page of mountain glaciation. 2 In
opposition to his view, Bonney published in 1871 a somewhat
elaborate article, in which the line of argument was : (1) that
the Alpine cirques must have been produced by the agency

THE CIRQUE AND ITS RECESSION 13

which shaped the valleys below them; (2) that the valleys
were not moulded by glaciers; and hence, (3) the cirques
must have been retained from the pre-glacial land surface. 3
The published discussion of this paper developed no oppo-
sition to the view, though Doctor, now Sir Archibald, Geikie
stated that he could not see his way to account for the vertical
walls surrounding the cirque. On the other hand, the Italian
Professor Gastaldi recognized the work of the ice in the shap-
ing of cirques in the Italian Alps, 4 as Helland did in those of
Norway. The latter believed that excessive weathering in
the rock above the neve played an important role, though ab-
rasion by the ice upon the floor was the larger factor. 5 Later
Russell in America, 6 Wallace in England, 7 and de Martonne
upon the continent, 8 further advocated the glacial origin of
cirques. Penck has explained the development of cirques
as the result of sub-glacial weathering alternate thawing
and freezing beneath glaciers during the incipient stage
particularly (" hanging glaciers "). 9 This eroding process,
he considered, would be greatest toward the middle of the
glacier, so that the original concavity of the slope beneath it
would be more and more deepened. It must be evident
that this explanation does not properly account for the steep-
ness of the cirque walls, which it will be remembered could
not be accounted for by Geikie.

Attention was again directed to the process of cirque shap-
ing by an important paper of Richter's published in 1896. 10
His studies having been made in Norway, where a country
rounded and polished by the continental glacier had been
only partly invested by mountain glaciers, the cirques from
the latter formed individual " niches " in the uplands. Fol-
lowing Gastaldi, the form of these niches was happily likened
to that of an armchair (see Fig. 4). 11 Richter observed that
the steep walls of the cirque were the only surfaces ungla-
ciated, and hence he concluded that they were not to be

14 CHARACTERISTICS OF EXISTING GLACIERS

ascribed to ice-abrasion, but to weathering. The moulding
of the cirque floor he ascribed to abrasion, and, referring to
the cirque walls, said -

The material loosened by weathering is removed by the gla-
cier or slides off over the neve to form either actual moraines,
or, at least, neve moraines. These walls do not bury themselves
in their own debris, and in consequence continually offer fresh
surfaces for attack. Finally, the wearing away of the cirque
floor by the glacier cooperates to keep the cirque walls on a
steep angle and facilitates avalanching.

FIG. 4. Cirque excavated in the glaciated surface of Norway, Northern Kjedel
on Galdhopig (after E. Richter).

In a more extended and later paper, 12 treating especially
the formation of cirques, Richter has explained that his view
differs from that of Helland only in ascribing greater impor-
tance to weathering upon the cirque walls and less to abrasion
upon the cirque floor. Inasmuch as the excessive weather-
ing of cirque walls, as maintained by Richter, is above the sur-
face of the neve, a horizontal plane of denudation should
develop at that level. No evidence of this plane being dis-
covered, its absence is explained by Richter through abra-
sion from the snowbank which would collect upon it so
soon as formed. This is the fatal weakness of the Richter
hypothesis.

Relation of Cirque to Bergschrund. Up to the beginning
of the twentieth century, as we have seen, few geologists had

THE CIRQUE AND ITS RECESSION 15

greatly concerned themselves with the erosion conditions at
high levels, the work of Richter being on the whole the most
comprehensive. The whole subject of cirque erosion was
rather generally ignored, as it is, indeed, to-day. Sir Archi-
bald Geikie, referring to the corries of the Scottish High-
lands, 13 wrote

The process of excavation seems to have been mainly carried
on by small convergent torrents, aided, of course, by the power-
ful cooperation of the frosts that are so frequent and so potent
at these altitudes. Snow and glacier ice may possibly have had
also a share in the task.

Writing in the same year, Reusch ascribed the Norwegian
cirques to the action of surface water descending through the
crevasses over falls in the continental glacier which, in Pleis-
tocene times, overrode the country ; 14 and the following year
Bonney reiterated his view that cirques were the product of
water-erosion. 15 Only a few years before, Gannett had curi-
ously explained the origin of cirques through the wear of
avalanched snow and ice upon the cirque floor, likening the
erosive process to that which takes place beneath a water-
fall. 16

The discovery of the method by which the glacier exca-
vates its amphitheatre must be credited to a keen American
topographer-geologist, Mr. Willard D. Johnson of the United
States Geological Survey. 17 In fact, to him and to another
American topographer, Mr. Francois E. Matthes, we owe the
most of what is known from observation concerning the
initiation and developme^fof the glacier cirque. Reasoning
that abrasion was incompetent to shape the amphitheatre,
Johnson early surmised thatVthe great gaping crevasse which
so generally parallels the cirque wall and is termed the
Bergschrund (Fr., rimaye) went down to the rock beneath
the neve, and that it was no accident that glaciated moun-
tains alone " abound in forms peculiarly favorable to snow-

16 CHARACTERISTICS OF EXISTING GLACIERS

drift accumulation " (see Fig. 5). These observations were
made as early as 1883, and in order to test his theory, John-

FIG. 5. Bergschrund below cirque wall on a glacier of the Sierra Nevada, Cali-
fornia (after Gilbert).

THE CIRQUE AND ITS RECESSION 17

son allowed himself to be lowered at the end of a rope 150
feet into the Bergschrund of the Mount Lyell glacier until he
reached the bottom. He found a rock floor to stand upon,
and rock extended up for 20 feet upon the cliff side. We
may here quote his terse sentences, since too little attention
has been accorded this important observation. 18

The glacier side of the crevasse presented the more clearly
defined wall. The rock face, though hard and undecayed, was
much riven, the fracture planes outlining sharply angular masses
in all stages of displacement and dislodgment. Several blocks
were tipped forward and rested against the opposite wall of ice ;
others quite removed across the gap were incorporated in the
glacier mass at its base.

Everywhere in the crevasse there was melting, and thin
scales of ice could be removed from the seams in the rock.
The bed of the glacier, elsewhere protected from frostwork,
was here subjected to exceptionally rapid weathering. By
maintaining the rock wall continually wet, and by admitting the
warm air from the surface during the day, diurnal changes of
temperature here resulted in very appreciable mechanical effects,
whereas above the neve only the seasonal effects were important.

This observation of Johnson is, it will be observed, in con-
trast with the suppositions of Richter, who believed that the
maximum sapping upon the cirque wall occurred above the
surface of the neve. The function of the Bergschrund, which
separates the stationary from the moving snow and ice within
the neve, is thus found to be of paramount importance in the
shaping of the amphitheatre.

With the coming of winter this process halts and the
Bergschrund fills with snow, but the following spring it again
opens, though always a little higher up and nearer to the
cirque wall. In this way the blocks excavated from the base
of the wall are the more easily transferred to the moving
portions of the glacier. 19

18 CHARACTERISTICS OF EXISTING GLACIERS

The Schrundline. That a sharp line is observable in aban-
doned cirques separating the accessible from the non-scal-
able portions of the wall, has been pointed out by Gilbert,
who has given his support to the view of Johnson, and con-
firmed it by observations of his own 20 (see Fig. 6). Penck, on

FIG. 6. Schrundline near Mt. McClure in the Sierra Nevadas of California.
Above the Schrundline it is too steep for snow to rest, and the drifts are ac-
cordingly below this level (after Gilbert).

the other hand, the following year revived the view of
Richter that excessive sapping occurs upon the cirque walls
above the neve surface, 21 though he calls in the Bergschrund in
order to gather in and remove the rock fragments which fall
from the cliff. 22

Initiation of the Cirque, Nivation. Johnson's studies upon
the processes of cirque shaping, had shown how a nearly
perpendicular cirque wall is steadily cut backward through
basal sapping at the bottom of the Bergschrund. The prob-
lem of how the snowbank, which was the inevitable forerun-
ner of the glacier, had transformed the relatively shallow

THE CIRQUE AND ITS RECESSION 19

depression which it presumably discovered into the steep-
walled amphitheatre, he did not attempt to solve. Yet the
nourishing catchment basin is a prerequisite to the existence
of the normal glacier. The solution of this problem has been
suggested by another American topographer, Mr. F. E.
Matthes. 23 In the Bighorn mountains of Wyoming he has
found exceptional opportunities for this study. Owing to
the low precipitation within the region and the consequently
inadequate nourishment of glaciers, a large part of the
pre-glacial surface still remains. There is, therefore, repre-
sented within the district every gradation from valleys
which were occupied by snow during a portion only of the
year to those which were the beds of glaciers many miles in
length. Both small glaciers and high-level drifts of snow
still remain in a number of places.

Mr. Matthes has demonstrated that the snowbanks with-
out movement steadily deepen the often slight depressions
within which they lie by a process which he has called "niva-
tion " - excessive
frost-work about
the receding mar-
gins of the drifts
during the sum-
mer season. The
ground being con-
tinually moist in
this belt due to

the melting Of the FIG. 7. Cross section of a snowdrift site on a slope
cnnw tVi<=> wafpr showing formation of niche by nivation (after
LOW > Matthes).

penetrates into

every crevice of the underlying rock, so that it is rent dur-
ing the nightly freezing. Rock material thus broken up
and eventually comminuted, is removed by the rills of
water from the melting snow. 24 By this process the original

20 CHARACTERISTICS OF EXISTING GLACIERS

depression is deepened, and, if upon a steep slope, its wall
becomes recessed (see Fig. 7).

The occupation of a V-shaped valley by snow, as Matthes
has further shown, tends through the operation of this
process to transform it into one of U -sect ion, since the weath-
ered rock material upon the slopes is transported by the rills
and deposited upon the floor. All gradations from nivated
to glaciated forms are to be found in the Bighorn range.

During the field season of 1909 the writer took the oppor-
tunity to examine neve regions and high-level snowbanks in a
number of districts, with the result of confirming the im-
portance of the nivation process as outlined by Matthes.
In plate 2 A and B are shown two snowbanks which were
photographed on July 25 near the summit of Quadrant
mountain, in the Gallatin range of the Yellowstone National
Park. The gently sloping surface of this mountain repre-
sents the pre-glacial upland unmodified by Pleistocene gla-
ciation. Though between 9000 and 10,000 feet above sea, it
supports a rich herbage, and is a favorite grazing-ground of
the elk. In A of plate 2 the snow bank is seen surrounded
by a wide zone within which no grass is growing, but where a
finely comminuted brown soil is becoming a prey to the mov-
ing water. B of plate 2 exhibits another bank lying in the
depression which it has largely hollowed. At its lower end
(at the left) is seen an apron of fine brown mud deposited by
the overburdened stream as it issues from beneath the
drift.

Later the writer has had in Swedish Lapland opportunity
to observe the results of the nivation process under excep-
tionally favorable circumstances. Here in place of a pre-
glacial surface, such as has been dissected in the American
mountain districts already described, the surface of the
country has been planed down to softly rounded knobs of
rather large scale under the influence of the mantling con-

PLATE 2.

A. Summer snowbank surrounded by brown border of finely comminuted rock.
Quadrant Mountain, Y.N.P.

B. Snowbank lying in a depression largely of its own construction. Note stream,
outwash of fine mud at the left. Quadrant Mountain, Y.N.P.

THE CIRQUE AND ITS RECESSION

tinental glacier of Pleistocene time. Subsequent to this gen-
eral planation the higher areas have in favorable situations
been occupied either by mountain glaciers or by more or less
persistent snow-drifts. The drift sites are found upon the
hillsides as distinct niches in which the characteristic " arm-
chair " form of the incipient cirque is already apparent. It is
the scale particularly which distinguishes them from glacial
amphitheatres (see Fig. 8). It is of great interest to find

FIG. 8. The characteristic form of drift sites on hillsides in Swedish Lapland.
The form of the cirque is already discernible. On the floor a division into hexa-
gons indicates that the process of solifluction has played an important part. (After
a photograph by G. W. v. Zahn.) 26

that the quite remarkable but as yet little understood pro-
cess of rock flow (solifluction) has here played an impor-
tant part in shaping the incipient cirque. The floors of the
drift sites are in some instances at least divided into the
hexagonal pattern so characteristic of soil flow on relatively

22 CHARACTERISTICS OF EXISTING GLACIERS

flat surfaces. 26 Inasmuch as it is now recognized that melt-
ing snow is the immediate requisite for effective solifluction,
it is apparent that this process in some of its phases at least
is clearly allied to the process of nivation.

An interesting question is at what point the snow-field or
neve will, by taking on a motion of translation, assume the
functions of a glacier. At this stage of transition the Berg-
schrund should first make its appearance. Comparison of
nivated and glaciated slopes in the Bighorn mountains led
Matthes to think that upon a 12 per cent, grade the neve
must attain a thickness of at least 125 feet before motion is
possible. Another possible method of approaching this
problem has suggested itself to the writer. In mountains
like the Selkirks, with steep slopes terraced by the flatly dip-
ping layers in the rock, a peculiar type of small cliff glacier
is nourished high above the larger snow-fields of the range
and avalanched upon the lower shelves so as to leave vertical
sections open to study (see plate 3 A). Perhaps because of
their small size these cliff glaciers have not developed cirques,
though a Bergschrund parallels the generally straight head-
wall. Examined through a powerful glass, the snow in the
lower layers can be seen to have lost its brilliant whiteness,
though it does not yet appear as ice. A number were ex-
amined with a view to determine the approximate minimum
thickness of the glacier, but all exceed the minimum estimate
of Matthes by at least 100 feet. This is not regarded as in
any way discrediting his figure, but rather as suggesting the
possibility of more thorough examination along the same
line.

REFERENCES

1 "Geology," vol. 1 : "Processes and their Results," 1904, pp. 272-276,
and especially fig. 250. See also "College Geology," by the same authors,
1909, p. 256.

2 John Tyndall, "On the Conformation of the Alps," Phil. Mag., Ser.
4, vol. 24, 1862, pp. 169-173.

THE CIRQUE AND ITS RECESSION 23

* T. G. Bonney, "On the Formation of 'Cirques,' with their Bearing
upon Theories attributing the Excavation of Mountain Valleys mainly
to the Action of Glaciers," Quart. Jour. Geol. Soc., vol. 27, 1871, pp. 312-
324.

4 B. Gastaldi, "On the Effects of Glacier-erosion in Alpine Valleys,"
ibid., vol. 29, 1873, pp. 396-401.

6 Amund Helland, "Ueber die Vergletscherung der Faroer, sowie der
Shetland und Orkney Inseln," Zeitsch. d. Deutsch. Geol. Gesellsch., vol. 31,
1878, pp. 716-755, especially pp. 731-733.

6 1. C. Russell, "Quarternary History of Mono Valley, California,"
8th Ann. Rept. U. S. Geol. Surv., 1889, pp. 352-355.

7 A. R. Wallace, "The Ice Age and its Work," Fortnightly Review, vol.
60, 1893, especially p. 757.

8 E. de Martonne, "Sur la periode glaciaire dans les Karpates meri-
dionales," C. R. Acad. Sci. Paris, vol. 129, 1899, pp. 894-897 ; ibid., vol.
132, 1901, p. 362.

9 Albrecht Penck, "Morphologic der Erdoberflache," vol. 2, 1894, pp.
307-308, figs. 17-20.

10 E. Richter, " Geomorphologische Beobachtungen aus Norwegen,"
Sitzungsber. Wiener Akad., Math.-Naturw. KL, vol. 106, 1896, Abt. I.,
pp. 152-164, 2 pis. and 2 figs.

11 See topographic definition of the cirque by De Martonne ("La periode
glaciaire dans les Karpates meridionales," C.R., 9 e Cong. Geol. Intern.,
Vienna, 1903, pp. 694, 695).

12 E. Richter, "Geomorphologische Untersuchungen in den Hochalpen,"
Pet. Mitt., Erganzungsheft 132, 1900, pp. 1-103, pis. 1-6.

13 "Scenery of Scotland," p. 183 (revised in 1901).

14 H. Reusch, Norges Geol giske Undersogelse, No. 32, Aarbog for 1900,
1901, pp. 259, 260.

15 "Alpine Valleys in Relation to Glaciers," Quart. Jour. Geol. Soc., vol.
68, 1902, p. 699.

16 "The effect is precisely like a waterfall. The falling snow and ice
dig a hollow depression at the foot of the steep descent just as water does."
(Nat. Geogr. Mag., vol. 9, 1898, p. 419.)

17 W. D. Johnson, "An Unrecognized Process in Glacial Erosion" (read
before the Eleventh Annual Meeting of the Geological Society of America),
Science, N.S., vol. 9, 1899, p. 106; also "The Work of Glaciers in High
Mountains" (lecture before the National Geographic Society), ibid., pp.
112, 113. The first public formulation of the doctrine by Mr. Johnson
was in an address before the Geological Section of the Science Associa-
tion of the University of California, delivered September 27, 1892.

18 W. D. Johnson, " Maturity in Alpine Glacial Erosion," Jour. Geol.,
vol. 12, 1904, pp. 569-578 (read at Intern. Congr. Arts and Sciences, St.
Louis, 1904).

19 1. C. Russell, "Glaciers of North America," 1897, p. 193.

20 G. K. Gilbert, "Systematic Asymmetry of Crest-lines in the High
Sierras of California," Jour. Geol., vol. 13, 1905, pp. 579-588. See also
E. C. Andrews, ibid., vol. 14, 1906, p. 44.

24 CHARACTERISTICS OF EXISTING GLACIERS

21 Many European glacialists and among them apparently Garwood
(Geogr. Jour., vol. 36, 1910, p. 313), have failed clearly to understand
that the basal sapping occurs at and near the base of the Bergschrund.

22 Albrecht Penck, "Glacial Features in the Surface of the Alps," Jour.
Geol., vol. 13, 1905, pp. 15-17.

23 Francois E. Matthes, "Glacial Sculpture of the Bighorn Mountains,
Wyoming," 21st Ann. Rept. U. S. Geol. Surv., 1899-1900, pp. 167-190.

24 Mainly in later seasons.

25 The structure of the pavement in the foreground has been added from
another photograph.

26 See H. W. Feilden, "Notes on the Glacial Geology of Arctic Europe
and its Islands," Quart. Jour. Geol. Soc., vol. 62, 1896, p. 738 ; also O.
Nordenskiold, "On the Geology and Physical Geography of East-Green-
land," Meddelelser om Gronland, vol. 28, 1908, p. 273; also O. Nordenskiold,
*'Die Polarwelt und ihre Nachbarlander," Leipzig and Berlin, 1909, p. 63 ;
also W. H. Hobbs, "Soil Stripes in Cold Humid Regions," 12th Report
Mich. Acad. Sci., 1910, pp. 51-53.

CHAPTER II

HIGH LEVEL SCULPTURING OF THE UPLAND

The Upland 1 dissected. Having obtained a clear concep-
tion of the process of head-wall erosion through basal sap-
ping, Johnson was in a position to account for the topography
which he encountered in the High Sierras of California. This
topography is best described in his own words : 2

In ground plan the canyon heads crowded upon the summit
upland, frequently intersecting. They scalloped its borders,
producing remnantal table effects. In plan as in profile, the
inset arcs of the amphitheatres were vigorously suggestive of
basal sapping and recession. The summit upland the pre-
glacial upland beyond a doubt was recognizable only in
patches, long and narrow and irregular in plan, detached and
variously disposed as to orientation, but always in sharp tabu-
lar relief and always scalloped. I likened it then, and by way
of illustration I can best do so now, to the irregular remnants
of a sheet of dough on the biscuit board after the biscuit tin
has done its work.

In a portion of the region where Johnson's studies were
made, his views have received verification by Lawson in a
beautifully illustrated paper. 3 Davis has furnished an ex-
cellent example from the Tian Shan mountains of the opera-
tion of the same cirque-cutting process, recording his adhe-
sion to the Johnson doctrine, 4 though many of his later papers
would indicate that he did not ascribe large importance to

26 CHARACTERISTICS OF EXISTING GLACIERS

the discovery. 5 In 1909 two papers from his pen give, how-
ever, larger prominence to the process. 6

With little doubt the failure to generally recognize the im-
portance of this process of cirque recession, clearly here a
more effective agent than abrasion, is to be explained by the

fact that in parts of
Europe and in the
Alps in particular,
one looks in vain for
evidences of the ear-
lier and more signifi-
cant stages of the
process. Glaciation
was here so vigorous
as to cause the re-
moval of all summit
upland. Within the
arid regions of the
western United
States, a more fruit-
ful field for study is
to be found. Here the work of Johnson has been supple-
mented by that of Gilbert 7 and Matthes. 8 Perhaps no-
where are the early stages of the process so clearly revealed
as in the Bighorn Mountains of Wyoming (see Fig. 9).

A somewhat more advanced stage of the same process is to
be found in the Uinta mountains of Wyoming, recently de-
scribed in a valuable monograph by Atwood, though here
without consideration of the cirque-cutting process in ac-
counting for the present topography. 9 Yet nowhere, so far
as the present writer is aware, has a view been reproduced
which so well illustrates the remnantal tableland and the
"biscuit-cutting" process of cirque recession (see Fig. 10). 10
The present writer has photographed other examples of the

FIG. 9. Pre-glacial upland invaded by cirques
"biscuit-cutting" effect; Bighorn Mountains,
Wyoming.

PLATE 3.

A. View of the Yoho Glacier at the head of the Yoho Valley, showing to the right
a series of three small cliff glaciers. Canadian Rockies.

B. Pre-glacial upland on Quadrant Mountain, Y.N.P., invaded by the cirque
known as the "Pocket."

HIGH LEVEL SCULPTURING OF THE UPLAND 27

same type in the Yellowstone National Park (see plate 3 B
and Fig. 11). Remnants of the pre-glacial surface will, in

FIG. 10. View of the scalloped tableland within the Uinta range and near the
head of the west fork of Sheep Creek (after Atwood).

any given district, be large or small according as nourishment
of the glaciers has been insufficient or the reverse. The
Uinta range, which extends in an east-west direction, and,

FIG. 11. Map of Quadrant Mountain, a remnant of the pre-glacial upland on
the flanks of the Gallatin Range, Y. N. P.

CHARACTERISTICS OF EXISTING GLACIERS

like the Bighorn mountains, has a core of homogeneous
granitic rock, displays this fact. An examination of At-
wood's map n shows that to the eastward, where the precipi-
tation has been least, the remnants of the original upland are
more considerable. This qualifying condition of glacier
nourishment will be subject to some modification because of
peculiarities in snow distribution. As shown by Gilbert, the

FIG. 12. Series of semicircular glacial amphitheatres whose scalloped crest forma
part of the divide of the North American continent.

first glaciers within any mountain district will probably
appear upon that side of the divide which is in the lee of the
prevailing winds. This fact is particularly well brought out
in Fig. 12.

Such a condition as is here represented, gives a most de-
cisive answer to the question concerning the protective or
denuding action of glacier ice. To the west of the divide
the snow has been swept clear, and these sweepings lodg-
ing in the lee have produced the glaciers on this side only.

PLATE 4.

^
|1

H
ii

2 M S C8

'C e J2 W)

2 'S o c

fi-l

I5I&

nil

s!fl

'^3 *0 T3 "S

^ S >,

all

c ^ -c a

1-1 <N 00 T*

PLATE 5.

Multiple secondary cirques on the west face of the Wannehorn seen

across the Great Aletsch Glacier, to which it is tributary.

(After a photograph by I. D. Scott.)

HIGH LEVEL SCULPTURING OF THE UPLAND 29

While this glacial cover has no doubt protected its base from
the ordinary weathering process, the extraordinary weather-
ing in the Bergschrund combined with abrasion and plucking
upon the floor, has excavated some 2000 feet of rock material,
even if we were to assume that the " unprotected " surface
to the westward has not been lowered.

The cwms of Wales have not as yet entirely removed the
summit upland in which they have been recessed, and this
residual surface perhaps furnishes the best European example
of an earlier stage in the process of cirque recession. 12

Modification in the Plan of the Cirque as Maturity is Ap-
proached. Owing to the fact that the sapping process
within the cirque operates on all sides, its early plan, when
the upland surface is supplying snow from all directions, will
approach the circle (see Fig. 9 and plate 3 B). Moreover,
in this stage the cirque will be but little, if any, wider than
the deepened and widened valley below (see plate 4, Figs. 1
and 2). Later, with the continuation of the sapping process,
the cirque becomes enlarged to such an extent that its sides
form recesses in the walls of the valley. Thus, in the plan,
the glacial valley of this stage bears some resemblance to that
of a nail with a large rounded head.

As the upland is still further dissected, the cirque becomes
more irregular in outline and widens into a roughly elliptical
form, not infrequently allowing it to be seen that it is in
reality composite or made up of several cirques of a lower
order of magnitude (plate 5, plate 6 B and Fig. 13).

Grooved and Fretted Uplands. --The new emphasis put
upon topographic expression of character in the maps issued
by government bureaus during the past few years, has fur-
nished physiographers a tool of which they are hardly yet
fully aware. Before, the aim of topographers seemed to be
to suppress all character through a rounding off of angles and
an averaging of the data. Perhaps nowhere has the change

30 CHARACTERISTICS OF EXISTING GLACIERS

been more noteworthy than in the maps issued by the
United States Geological Survey, 13 and the later sheets par-
ticularly, when relating to glaciated mountain districts,
afford us the opportunity of tracing the successive steps in the
dissection of such upland districts by the cirques of moun-
tain glaciers. For plate 4, four areas have been selected to
represent successive stages in such a progressive dissection.
An early product, in which large remnants of the upland sur-
face still remain, may well be designated a " grooved or chan-
nelled " upland (see plate 6 A a).

As the hemicycle advances, it will be observed that on the
flanks of the range are found the largest remnants of the
original upland surface (see Fig. II), 14 owing to the tendency
of the cirque to push its side walls out beyond the limits of
the U-shaped valley below. With complete dissection of the
plateau no tabular remnants are to be discovered. The gen-
eral level of the district has now been lowered, but above this
irregular surface project one or more narrow pinnacled
ridges files of " gendarmes " separated by crevices or
" chimneys." These palisades at fairly regular intervals
throw off lateral palisades having crests which fall away in
altitude as they recede from the trunk ridge. In general
terms, and describing the major features only, we have here
to do with a gently domed surface, on which is a fretwork
of comb-like ridges projecting above it. This surface may
be designated a " fretted upland " (see plate 6 A 6). Such a
condition is realized in the Alps, and is seen to special ad-
vantage from the summit of Mont Blanc (see plate 7 A).

The transition from the grooved to the fretted upland is
well brought out in two views taken by Lawson in the High
Sierras of California (loc. cit., plate 45, A and B). The
fretted upland differs from the grooved upland of an earlier
stage of the cycle in the complete dissection of the surface.
The character of the fretted surface is well brought out by

PLATE 6.

A. (a) A grooved upland in the Bighorn Mountains, Wyoming. (6) A fretted up-
land, Alaska.

B. Multiple cirque of the Dawson glacier, having a major subdivision into halves,
which enclose, respectively, the Dawson and the Donkin neves. The view is from
the Asulkan Pass, Selkirk Mountains.

PLATE 7.

A. Fretted upland of the Alps as seen looking northeastward from the summit of
Mont Blanc, July 25, 1908. The cirque to the left is that of the Glacier de
Talefre, with the Jardin in its centre, and distant about 10 miles.

B. Map of a portion of one of the Lofoten Islands, showing a fretted surface par-
tially submerged and emphasizing the approximate accordance of summit levels.

HIGH LEVEL SCULPTURING OF THE UPLAND 31

the topography of the Lofoten Islands off the arctic coast of
Norway, where the effect is somewhat heightened through
the partial submergence and consequent obliteration of the
irregularities in the floor (see plate 7 B).

At this stage there is undoubtedly a general accordance
of level in the crests of the frets upon the domed surface, as
Daly, taking due account of the cirque-cutting process, has
claimed. 15 Moreover, the existence of such a series of frets
as are to be found in the Alps, forces us to conclude that such
an accordance of summits persists for a considerable time.
Were this not the case, we should find a larger number of low

FIG. 13. Diagram to illustrate the manner of dissection of an upland by mountain

glaciers.

cols and a longer persistence of the semicircular form of the
cirque. It seems probable, therefore, that a very definite
relationship obtains between the plan of the cirque and that
of the near-lying upland remnants that contribute snow to its
basin. So soon as cirques approach from opposite sides of a
divide, the portions of their basins which are more nearly ad-
jacent receive less snow, and, in consequence, accomplish less
sapping than the walls on either side where snow is lodged in

32 CHARACTERISTICS OF EXISTING GLACIERS

a quantity but slightly diminished. This self-regulating
process will tend to broaden the cirque and eventually give it
irregularities of outline dependent primarily upon the initial
positions and the individual nourishments of its near-lying
neighbours.

Characteristic higher Relief Forms of the Fretted Upland.
In the earlier stages of mountain glaciation the upland is chan-
nelled by valleys U-shaped in their upper stretches, and some-
what broadened into steep-walled amphitheatres at their
heads. With the complete dissection of the upland, the coa-
lescence of the many cirques at last cuts away every remnant
of the original surface and yields relief-forms which are de-
pendent mainly, as already stated, upon the initial positions
of the cirques. 16

If there be a highest area within the upland, the snow will
be carried farthest from it by the wind, and this will be in con-
sequence the last to succumb to the cirque-cutting process.
The dome of Mont Blanc in the midst of a forest of pinnacles,
no doubt owes its peculiar form to the fact that it dominated
the pre-glacial upland.

A high district whose area is not too large compared with
that of the individual cirques, when at last dissected by the
cirques may be designated a " karling." 17 A typical example
from Northern Wales is represented in plate 8.

Elsewhere within the upland the coalescence of cirques has
produced comb-like palisades of sharp rock-needles which
have long constituted the aiguille type of mountain ridge. In
the literature of physiography, such ridges have perhaps most
frequently been designated by the term "arete" (fishbone),
though in the Alps the term " grat " 18 (edge) has been applied
especially to the smaller and lateral ridges of this type. I pro-
pose to use for all such palisades of needles derived by this
process the name " comb-ridge ;;19 as the best English term
available. The frequent occurrence of lateral arms joined

PLATE 8.

A karling in North Wales (from the Bangor sheet of the British Ordnance Survey,

1907).

HIGH LEVEL SCULPTURING OF THE UPLAND 33

to the main palisade of needles suggests a differentiation into
main and lateral comb-ridges.

In every mountain district maturely dissected by glaciers,
are to be found sharp horns of larger base and especially of
higher altitude than the individual minaret-like teeth of the
comb-ridges. They are further in contrast with the latter
by having an approximately pyramidal form, and a base
most frequent-
ly a triangle
with flatly in-
curving sides.
They appear
most frequent-
ly at the junc-
tion points of
the comb-ridges
between three
or more impor-
tant snow-fields
(see Fig. 14).
Such forms are
generally termed " horns " in the Alps, and the word being
of the same form in English, it may well be retained as a
technical expression. The Matterhorn in Switzerland is the
type par excellence (see plate 9 A), though similar and
almost equally striking examples are numerous ; as, for
example, the Weisshorn and Gross Glockner in the Alps,
Mount Assiniboine in the Canadian Rockies, or Mount Sir
Donald in the Selkirks. The triangular base and pyramidal
form are so common to this feature that they have found
expression in the local names, as Dreieckhorn, Delta-form
peak, etc.

The Col and its Significance. --The prominent horns of any
glaciated mountain district no doubt occupy positions cor-

FIG. 14. Position of the Aletsch- and Dreieckhorns be-
tween the Upper, Middle, and Great Aletsch neves.

CHARACTERISTICS OF EXISTING GLACIERS

responding in the main to the more elevated areas in the
original upland surface, since such positions would be earliest
cleared of snow, and hence latest attacked by the cirques.
After complete dissection of the upland, the comb-ridges
which fret its surface will be attacked from opposite sides,
and their crests will be first lowered at the points of tan-
gency of the adjacent cirques generally near the middle

points of their curving
outlines. The sky-line of
the ridge will thus be
lowered in a beautiful
curve forming a pass or
col. Inasmuch as the
cirque approaches in its
form an inverted and
truncated cone of acumi-
nated type, the curve to
which the rim of the col
approximates will be fur-
nished by the intersection
of two cones of revolution
with the same apical angle
and having parallel axes
(see Fig. 15 and plate 9
B). This curve is approx-
imately a hyperbola, the
eccentricity of which will be largely dependent upon the
relative sizes of the cirques in question.

The corries of the Scottish Highlands, being generally of
small size, have coalesced to produce a very characteristic
scalloping of the horizon line seen to advantage in Ben Nevis,
or, better still, in the sculptured gabbro of the Cuchulin hills
in Skye. 20 To judge from views, also, such forms are found
in North Wales, features which in many respects are different

FIG. 15. Diagram to illustrate the forma-
tion of a col through the intersection of
cirques.

PLATE 9.

A. View of the Matterhorn from the Corner Grat.
(After a photograph by I. D. Scott.)

B. Col between Mt. Sir Donald and Yogo Peak in the Selkirks,
showing the characteristic hyperbolic profile.

(Copyright by the Keystone View Company.)

HIGH LEVEL SCULPTURING OF THE UPLAND 35

from those found in the Alps or in the North American
mountains. 21

It must be regarded as of deep significance that mountain
passes in areas which have supported glaciers are so generally
at high levels. Deep glacier-cut valleys available as high-
ways and transecting high ranges are extremely rare; so far
as the writer is aware, being known only from the Southern
Andes 22 and Alaska. 23 This fact must have its explanation,
it is believed, in a notable and abrupt retardation in the rate
of cirque-wall recession, following close upon the dissection
of the upland. Whether this is due to the reduced snow
accumulation immediately beneath the cirque wall owing
to the lack of a near-lying collecting ground, it is as yet too
early to say; but a comparison of the acclivities in the
marginal snow-slopes on neves of the Bighorn and Alaskan
districts might yield an answer to the question.

Though the sapping process at the base of cirque walls
up to maturity is doubtless far more potent than abrasion
and plucking upon the floor of the amphitheatre, it seems
likely that in the subsequent stage the reverse is the case.
This would at least explain the tendency of glacier valleys
to deepen rapidly in the higher altitudes, or, in Johnson's
phrase, to get " down at the heel:"

The Advancing Hemicycle. -- With the augmentation of
rigorous climatic conditions within any district where glaciers
already exist, the latter will be continually more amply
nourished, and must in consequence increase steadily in size.
Such climatic changes may even be conceived so considerable
that at last the entire range is submerged beneath snow
and ice, thus producing an ice-cap.

Direct observation of the successive stages through which
glaciers pass from their initiation to their culmination in an
ice-cap, is, of course, impossible, for the reason that we live
in a receding hemicycle in which practically all known gla-

36 CHARACTERISTICS OF EXISTING GLACIERS

ciers, instead of expanding, are drawing in their margins;
yet a synthetical reconstruction of the life-history is none
the less possible. To employ an illustration already used in
a different connection, in order to learn the life-history of a
particular species of forest tree, it would not be necessary
to sit down and observe an individual tree from the germina-
tion of its seed to the decadence of the full-grown tree. We
may with equal profit go into the forest and observe trees
of the same species in all stages of development. In the
study of glaciers our opportunity is hardly so fortunate
as this, for, as already stated, all glaciers appear to be in the
declining stage (if we ignore the short period variations)
whereas it is the advancing hemicycle with which we are now
concerned. The characters of glaciers as concerns their size
and shape depend, however, in so large a measure upon the
one element of alimentation, that if we neglect characters of
a second order of magnitude, we may by inference construct
the history with sufficient accuracy from existing examples.
The alimentation of mountain glaciers is dependent upon
the amount of precipitation and upon the temperature, the
former being in large measure determined by the adapta-
bility of the relief for local adiabatic 24 and contact refrigera-
tion of the air. The important factor, temperature, while
a function of many variables, yet in a broad way varies di-
rectly with latitude and altitude. The size and the form
of glaciers is, however, determined, not solely by nourish-
ment (mainly in the higher levels), but also to some extent
by losses (particularly in the lower levels). In the main,
however, the losses are controlled by the same factors as the
gains, and maintain to them a more or less determinate pro-
portional relationship. Exceptions to this definite proportion
occur when in high latitudes the glacier is attacked directly
by the sea (tidewater glaciers), when it is suddenly melted
by the heat of a volcanic eruption (Icelandic Jokulls), or

HIGH LEVEL SCULPTURING OF THE UPLAND 37

when disturbed by a heavy earthquake (Muir glacier in 1899).
In form glaciers will be in large measure determined by the
existing topography of the upland, which may generally
be assumed to be some product of sub-aerial erosion. * Start-
ing, therefore, with the puny glaciers of arid regions in low
latitudes, and ending with the high latitude glaciers within
areas of excessive precipitation, we run almost the whole
gamut of glacier alimentation.

The initial forms of glaciers may be described as snow-
bank, "new-born," or nivation glaciers, and will at first
be few in number and located with wide intervening spaces
of upland. The continuance of the nivation process will
deepen other intermediate small depressions upon the upland,
so that with increasing snowfall additional glaciers will
appear in the spaces between the first as the latter are devel-
oping their amphitheatres. These cirques, at first no wider
than the valleys below, will later cut recesses on either side,
at the same time that the glacier is pushed farther down the
valley and occupies its bed to a greater and greater depth.
The grooved upland of this stage, through additional cirque
recession in the highlands and through abrasion and plucking
in the intermediate levels, becomes at last transformed into
the fretted upland, with its network of projecting comb-
ridges. Up to this point the glacier ice has, perhaps, been
restrained within valleys, which it has discovered and has
progressively widened and deepened. If the annual tem-
perature continues to be lowered, there must come a time
when the ice-feet from the better-nourished glaciers, or from
those with the shortest route to the foreland fronting the
range, will debouch upon the plain, spreading as they do so
into fans or aprons (see plate 10 A). Later all neighboring
glaciers may arrive at this stage, and by spreading upon the
foreland, coalesce with one another to form a single broad
apron, such as may be seen in the Malaspina glacier of Alaska

38 CHARACTERISTICS OF EXISTING GLACIERS

(see plate 10 B). While the glaciers are thus pushing out
upon the foreland, they have been deepening in their valleys,
and eventually come to overtop portions of the lateral comb-
ridges of the fretted upland, thus moulding the sharpened
needles into rounded shoulders of rock. In places the
glaciers from adjacent valleys will flow together through the
irregular depressions separating peaks, thus producing islands
or nunataks.

But the increased size of the individual glaciers of the
range has corresponded to increased activity of cirque reces-
sion in the high altitudes, and this has resulted in the forma-
tion of cols or passes through the range. Snow which has
been divided at the summit, as has water by a divide, may
now be consolidated into glacier ice over the col before
the separation is made. Thus it comes about that without
a definite cirque, glaciers will transect the range flowing in
opposite directions from a central ice-field. Such a broad
central ice-field is found to-day between Mount Newton of
the St. Elias group and Mount Logan to the eastward. 25

The advance of the glacier ice up the sides of the valleys,
so as partially to submerge the lateral comb-ridges, may
not end until all are thus covered and the ice flows away
from the central broad area, radiating in many directions.
Here the process of cirque recession, which has mainly
sculptured the rock in the higher altitudes, comes to an end
as we reach the ice-cap stage of glaciation. Transitions
toward such ice-cap glaciers are to be found to-day in the
Elbruz and in the Kasbek region of the Causasus, where a
central elevated snow-field is the common neve of several
glaciers radiating in as many directions. 26 It is of consider-
able interest to note that in the Caucasus district, at least,
there is evidence that rocky comb-ridges are submerged
beneath the ice and make their appearance so as to separate
the marginal ice-tongues. The persistence of an ice-cap over

HIGH LEVEL SCULPTURING OF THE UPLAND 39

a mountain region, as is clear from study of the glaciated
mountains in Eastern Lapland tends to largely obliterate
relief forms characteristic of mountain glaciers as they are
replaced by the rounded shoulders of " rundlings " or the
smaller " roches moutonnees." As soon, however, as nourish-
ment has been so far reduced that the higher points once
more appear from beneath their snow cover, cirque recession
will begin again, and, if long continued, the evidence of the
ice-cap will disappear. Lack of glacial scratches or polish
in uplands sapped by this process should, therefore, not be
allowed to weigh too heavily in reconstructing the glacial
history of a district.

REFERENCES

1 The term " upland " is here used in a general sense to designate any
relatively elevated area of the land.

2 W. D. Johnson, Jour. Geol., loc. cit.

3 A. C. Lawson, "The Geomorphogeny of the Upper Kern Basin," Bull.
Dept. Geol. Univ. Calif., vol. 3, No. 15, especially p. 357, pis. 32, 45.

4 W. M. Davis, Appalachia, vol. 10, 1904, pp. 279-280.

5 E.g., cf. Scot. Geogr. Mag., vol. 22, 1906, pp. 76-89.

6 "Glacial Erosion in North Wales," Quart. Jour. Geol. Soc., vol. 65,
1909, pp. 281-350, pi. 14; also "The Systematic Description of Land
Forms," Geogr. Jour., vol. 34, 1909, p. 109.

7 Jour. Geol., loc. cit.

8 Ibid., loc. cit.

9 Wallace W. Atwood, "Glaciation of the Uinta and Wasatch Moun-
tains," Prof. Paper, U. S. Geol, Surv., No. 61, 1909, pp. 1-96, pis. 1-15.

10 Other apt illustrations have been furnished by Lawson in a photo-
graph taken in the Upper Kern region of the California Sierras (loc. cit.,
pi. 32 B), and by Davis in a sketch made in the Tian Shan mountains
(Appalachia, vol. 10, 1904, p. 279).

11 Loc. cit., pi. iv.

12 W. M. Davis, "Glacial Erosion in North Wales," Quart. Jour. Geol.
Soc., vol. 66, 1909, figs. 7, 27, 28.

13 See D. W. Johnson and F. E. Matthes, "The Relation of Geology to
Topography." Reprint from Breed and Hosmer's "Principles and Prac-
tice of Surveying," chap, vii., Wiley & Co., N.Y., 1908.

14 Other quadrangles of the U. S. Geological Survey which display the
upland surface more or less completely dissected by mountain glaciers are
the following : early stage: Younts peak (Wyoming), Marsh peak (Utah-
Wyoming), and Georgetown (Colorado); partial dissection: Mount Lyell
and Mount Whitney (California), Grand Teton (Wyoming), Gilbert peak

40 CHARACTERISTICS OF EXISTING GLACIERS

and Hay den peak (Utah- Wyoming), and Silver ton and Anthracite (Colo-
rado) ; complete maturity: Kintla Lakes (Montana).

15 R. A. Daly, "The Accordance of Summit Levels among Alpine Moun-
tains," Jour. GeoL, vol. 13, 1905, pp. 117-120.

16 The analogy with the forms produced by etching upon crystal faces
is so striking that it may be helpful to note it in comparison. The first
effect of a reagent in its attack upon the plane of a crystal face is the
excavation of deep pits which have a similar and wholly characteristic
form, though the surface in other places remains unchanged. These
pittings later increase in number, as they do in size, and eventually they
mutually coalesce, destroying utterly the original plane surface, and leav-
ing in relief a series of hills and ridges (etch-hills) projecting above a
somewhat irregular floor, whose average level is a measure of the average
depth of the excavations made by the process. The noteworthy differ-
ence between this process and that of cirque recession in glaciated uplands
is that the glacial etch-figures are relatively longer and narrower.

17 Penck und Bruckner, "Die Alpen im Eiszeitalter," vol. 1, Leipzig,
1909, pp. 284, et seq.

18 Very likely originally from grate, fishbone.

19 The use of combe in the Jura and the Cote d'Or for different types of
valley, or of coombe in the Southern Uplands of Scotland for a glacial
valley, being each essentially local and having further no relation to the
toothed article which suggests the name comb-ridge, does not constitute
a serious objection to this choice. Mr. Matthes (and possibly others)
have already used the expression comb-ridge in the above described sense.
(Appalachia, vol. 10, 1904, p. 260).

20 See Barker, "Glaciated Valley of the Cuchulins, Skye," GeoL Mag.
(fig. 4), vol. 6, 1899, p. 197 ; also "Ice Erosion in the Cuillin Hills, Skye,"
Trans. Roy. Soc. Edinb., vol. 40, 1901-1902, pp. 234-237.

21 This characteristic form of cirque, partly open at the head, is well
brought out in a view published by Sir Andrew Ramsey as early as 1852,
Quart. Jour. GeoL Soc., vol. 8, p. 375.

22 'Argentine-Chilian Boundary in the Cordillera de los Andes.' 5 vols.

23 R. S. Tarr, " Glaciers and Glaciation of Yakutat Bay, Alaska," Bull.
Am. Geogr. Soc., vol. 38, 1906, p. 149.

24 This term applied to change of temperature of a gas, implies that the
change is due to change of pressure and volume and not to the communi-
cation of heat from outside. The heating of a bicycle tire on pumping or
the cooling on emptying, may servo for illustration.

^Filippo di Filippi, "The Ascent of Mount St. Elias." Panorama at
end of volume (unnumbered) from an elevation of 16,500 feet.

26 H. Hess, "Die Gletscher," Braunschweig, 1904, pp. 65-68.

27 Penck und Bruckner, " Die Alpen im Eiszeitalter," vol. 1, Leipzig,
1909, pp. 286-287.

CHAPTER III

CLASSIFICATION OF GLACIERS BASED UPON COMPARA-
TIVE ALIMENTATION

Relation of Glacier to its Bed. From what has been said
in the preceding section concerning the changes of glaciers
in correspondence with a progressive augmentation of glacial
conditions, it must be evident that any attempt to use each
circumscribed body of snow and ice as a unit in name or in
type will lead to endless confusion. Ice bodies being ex-
tremely sensitive to changes in annual temperature, a differ-
ence of one degree may be sufficient to join many ice bodies
into one, or to differentiate one body into many. If, how-
ever, we examine the distribution of snow and ice masses
within the valley which they either wholly or partially
occupy, it will be seen that there are relatively few distinct
glacier types, and that the coalescence of smaller ice masses,
or the breaking up of larger ones, does not necessarily alter
the type exemplified.

The more important types called for by analysis on this
basis do not differ greatly from those in general use; but
the genetic relationships of these types are here first brought
out, together with distinct and intermediate transitional
forms. In the following table, excepting the initial type and
the glaciers with inherited basins, the arrangement is in the
main one of decreasing alimentation:

42 CHARACTERISTICS OF EXISTING GLACIERS

Nivation type (Bighorn glaciers).
Ice-cap type (Jokulls of Iceland).
Piedmont type (Malaspina glacier).
Transection type (Yakutat glacier).
Expanded-foot type (Davidson glacier).
Dendritic type, normal sub-type (Baltoro glacier).

Hanging glacierets (Triest glacier).

Cliff glacierets (Lefroy cliff glacieret).
Dendritic type, Tide-water sub-type (Harriman-Fjord

glacier).
Inherited basin type (Illecillewaet glacier).

Reconstructed type (Victoria-Lefroy glacier).

Volcanic cone type (Nisqually glacier).

Cauldron type (Caldera glacier).
Radiating (" Alpine") type (Nicolaithal glacier).
Horseshoe type (Mount Lyell glacier).

Nivation Type. This type of glacier has also been called
" new-born " or " snowbank " glacier, and represents the
initial stage of glaciation. Though small in size, such glaciers
differ markedly from those of the same dimensions which
cling to the steep walls of a large cirque (see horseshoe glaciers
below), which Tarr has referred to as "dying glaciers." 1
Numerous examples of snowbank glaciers are furnished by
the Bighorn mountains of Wyoming. Other known types
of mountain glaciers are all represented, and follow naturally
in sequence during a receding hemicycle of glaciation. In
their discussion we shall conceive a mountain district to pass
by slow stages from a culmination of glacial conditions
toward a comparatively genial climate.

Ice-cap Type. 2 Though in form and general characters
resembling so-called continental glaciers, the ice-caps by
reason of their smaller dimensions form a connecting-link
with mountain glaciers, and are usually developed upon
small plateaus or uplands. They correspond to conditions
of extremely heavy snow precipitation, and in consequence

CLASSIFICATION OF MOUNTAIN GLACIERS 43

have not been found fully developed outside the polar or
sub-polar regions (see inherited basin glaciers below).

The normal type of ice-cap glacier is represented by the
mantle over Redcliff peninsula, north of Inglefield gulf in
Greenland. 3 It suffers no interruption from mountain peaks,
but the ice creeps out in all directions from a central area,
and sends out marginal lobes and tongues which much resem-
ble, save for their whiter surface, the snouts of dendritic and
radiating glaciers (see below). The Jokulls of Iceland are
very similar, and form flatly arched or undulatory domes
of ice having short lobes about their margins (see plate 11, 1).
The largest of these, the Vatnajokull, has an area of 8500
square kilometres. 4 In Scandinavia the smaller plateau
glaciers with marginal tongues of proportionately greater
length, such as the Jostedalsbiaen, serve to connect this type
with that of the dendritic glaciers (see plate 11, 2). 5 The
Richtofeneis on Kerguelen island, recently described by the
German South-polar Expedition, seems to be very similar. 6
According to Meyer, the ice mass upon the summit of Kili-
mandjaro in Africa is an " ice carapace," having much re-
semblance to the ice plateaus of Scandinavia. 7

Piedmont Type. Piedmont glaciers, like ice-caps, corre-
spond to conditions of exceptionally heavy precipitation, and
are only known from sub-polar regions. In contrast to
small ice-caps, the existing examples are found in connection
with mountains of strong relief, so that the snow and ice
which in ice-caps find their way slowly out to the margin of a
flat or gently sloping plateau, are in the piedmont glacier
discharged through valleys from lofty highlands to debouch
upon the foreland at the foot of the range. The well-known
type is the Malaspina glacier of Alaska, explored and de-
scribed by Russell (see plates 10 B and 11, 3). 8 Near it and
farther to the west is the Bering glacier of about the same
size. 9 To the east of the Malaspina glacier is the Alsek, a

44 CHARACTERISTICS OF EXISTING GLACIERS

much smaller piedmont glacier. 10 In Chili south of 42 S.
lat. are found other piedmont glaciers, among them the San
Rafael. 11 During Pleistocene times piedmont glaciers
existed in many mountain districts, notably, however, the
Alps 12 and the Rocky mountains of North America. 13 An
imperfect transition from the piedmont type toward the
continental glacier is illustrated by the Friederickshaab
glacier in Greenland, which pushes its front out upon the
foreland as an extension of the inland ice of that continent
(see Fig. 94, p. 171).

Above the ice-apron and within the range, the piedmont
glacier bears a close resemblance to the dendritic type (see
below), though in general it may be said that its valleys
are filled to a much greater depth. The largest stream
feeding the fan of the Malaspina glacier has been named
the Seward glacier, while other tributaries are known
as the Agassiz and the Tyndall (see plates 10 B and
11, 3). It is interesting to note that however steep these
feeders to the ice-apron may be, the latter always shows
an exceedingly flat slope, and is, moreover, relatively
stagnant.

Transaction Type. In a late stage of augmenting glacial
conditions or in an early stage of the receding hemicycle,
what is essentially one body of ice may be divided over a pass
and flow off in opposite directions toward different margins
of the range. For this type, exemplified by the Nunatak
glacier of Alaska, Tarr has used the term " through glacier," 14
and Blackwelder has instanced the Yakutat glacier and
perhaps the Beasley within the same region. 15 Such glaciers,
which may be referred to as the transection type, are often
the highways which give readiest access to the hinterland.
A glacier of this type, which has been carefully mapped, is
the Sheridan glacier near the mouth of the Copper river in
Alaska (see Fig. 16). 16 ^ An excellent panorama of one of the

PLATE 10.

A. Expanded fore-foot of the Foster glacier, Alaska.

B. Type of piedmont glacier.

(From a photograph of the new model of the Malaspina glacier made under the
direction of Lawrence Martin.)

CLASSIFICATION OF MOUNTAIN GLACIERS

grandest transection glaciers has been furnished by Sella. 17
The glaciation of the Grimsel pass in Switzerland clearly
indicates that at one time a glacier of this type was parted
over the present divide, one stream passing down the Rhone

FIG. 16. Map of a transection glacier, the Sheridan Glacier near the Copper
River in Alaska (after G. C. Martin).

valley, and the other down the Hasli valley toward Meirin-
gen. Far grander exhibits of the same sort are to be found
in the Southern Andes. 18

Expanded-foot Type. When a piedmont glacier draws
in its margin as it shrinks with the coming of a warmer
climate, the several ice-streams which feed the apron of ice
upon the foreland end in smaller fans at the mouths of the
individual valleys. Perhaps the best known example of
such an expanded-foot glacier is the Davidson, on the Lynn
canal in Alaska, though the Foster and Mendenhall glaciers
of the same district are similar (see plate 10 A). The Miles
and Childs glaciers, near the Copper River, are also of this
type, and have been mapped by Martin. 19 The transection
glacier known as the Sheridan is in the same vicinity, and has

CHARACTERISTICS OF EXISTING GLACIERS

an expanded forefoot a good illustration of the combina-
tion of these two types in one (see Fig. 16). The type par
excellence of the expanded-foot glacier is the Baird glacier
on the Copper River (see Fig. 17). 20 A larger but less
perfect example of the expanded forefoot than any thus far
mentioned is the Klutlan, in the Yukon basin, whose foot
extends a number of miles beyond the front of the St. Elias

FIG. 17. The Baird glacier, a typical expanded-foot glacier (after Tarr and

Martin).

range/ 1 The Martin river glacier in the Copper river dis-
trict affords another example, since it expands for a dis-
tance of over 20 miles. It is, however, partially restrained
by a range of hills rising on its southern margin, and by
Martin has been considered intermediate between the pied-
mont and valley types. 22

CLASSIFICATION OF MOUNTAIN GLACIERS

Dendritic or Valley Type. Retiring within the range as
warmer temperatures succeed to more rigorous conditions,
glaciers are of necessity restricted to individual valleys and
their tributaries. They come thus to have a plan as truly
arborescent as that of water-drainage, and they may in this
stage be called " dendritic glaciers." Unfortunately, the term
" valley glaciers," in every way appropriate, has been gen-
erally applied to glaciers which occupy valley heads only,
and hence the term would have to be redefined in its natural
rather than its inherited significance. This glacier type
geographers are most familiar with in restorations of Pleis-

FIG. 18. Outline map of the Hispar glacier, Karakorum Himalayas (after

Conway) .

tocene glaciers, 23 but it is none the less a common form to-
day in districts more distant from commercial centres, and
hence less easily accessible for study. From the Karakoram
Himalayas, the Baltoro, Hispar, and Biafo glaciers, all of
this type, have been described and carefully mapped by Sir
Martin Conway. 24 An outline map of the Baltoro glacier is
reproduced in plate 11, 4 and one of the Hispar glacier in
Fig. 18. Other valley glaciers, generally less extensive,

CHARACTERISTICS OF EXISTING GLACIERS

have been mapped by Garwood 25 from the Kangchenjunga
Himalayas. In the Central Tian Shan mountains are other
glaciers of this type. 26 In the New Zealand Alps the Tasman
glacier furnishes another example of the same valley type 27
(see Fig. 19 and plate 11, 5). Still other examples have

FIG. 19. Outline map of the Tasman Glacier, New Zealand (after v. Lenden-

feld).

been described from the mountains of Alaska, such, for
example, as the Kennicott and Chistochina glaciers. 28

Comparison of a number of examples of valley glaciers
may illustrate as many different stages in the retreat of the
glacier from a position in which it occupied its entire valley,
to the retirement almost within the mother cirque at the
head. The examination of the vacated valley has taught
us that the tributary glaciers erode their beds less deeply
than the trunk stream lying in the main valley. It is the
surfaces of the ice-streams only that are accordant, and hence
a lack of accordance in the bed levels has yielded the so-called
hanging valleys with their characteristic ribbon falls. No-
where can the hanging valleys be observed in greater perfec-

PLATE 11.

36 Miles

TYPES OF MOUNTAIN GLACIERS.

1-2, ice-cap types from Iceland and Norway respectively ; 3, piedmont type, Alaska ;
4-5, dendritic types from the Himalayas and New Zealand respectively ; 6, dendritic
type (tidewater glacier), Alaska; 7-8, radiating types, Alps ; 9, radiating type, Hima-
layas ; 10-12, horseshoe types from Himalayas, Selkirks, and Canadian Rockies re-
spectively ; 13, horseshoe type, Colorado ; 1415, inherited basin types frcm Alps and
Selkirks respectively ; 16, inherited basin type (reconstructed glacier), Canadian Rockies.

CLASSIFICATION OF MOUNTAIN GLACIERS

tion or on a grander scale than in the troughs, now largely
abandoned of ice, which enter the great fjords of the "inside
passage " to Alaska (see plate 13 A). 29

Mile*

FIG. 20. Outline map of an inherited basin glacier, the Illecillewaet Glacier of the
Selkirks. The dotted line is the divide (after Wheeler).

As the foot of the trunk glacier retires up its valley, the
lateral tributaries which are nearest the mouth of the valley

50 CHARACTERISTICS OF EXISTING GLACIERS

are at first separated from it and develop their own front
moraines. Later they are left high above the main stream
as a series of "hanging glacierets" (see plate 12). 30 The
series of hanging glacierets, as will be observed in the maps
of the Baltoro and Hispar glaciers, often persist above the
main valley well below the foot of the trunk stream.

Inherited Basin Type. The dendritic type of glacier
hardly appears in the Alps at all, though the Great Aletsch
glacier might, perhaps, be regarded as a small and imperfect
example. The size and characters of the latter are, however,
for the district in which it lies, abnormal and to be accounted
for by the existence of a natural interior trough lying between
the Berner Oberland on the one side and the high range north
of the Rhone valley upon the other, from which basin small
outlets only are found through the southern barrier (plate 11,
14). A better example, however, of this special type of
glacier, in which the inherited topography has exercised a
greater influence upon the glacier form than has the auto-
sculpture, is furnished by the Illecillewaet glacier of the

Selkirks (see Fig. 20),
which, from a roughly
rectangular snow-ice field
lying between parallel
ridges, sends out short
tongues leading in differ-
ent directions. A glacier
of this type, with a mod-
erate increase only of
^ .' ;, -== ^Miie 9 alimentation, would pro-

FIG. 21. Outline map of a reconstructed duC6 a Small ice-Cap.

glacier, the Victoria and Lefroy glaciers in Another abnormal f Orm

the Selkirks (after Wheeler).

of glacier due to the pe-
culiarities of the basin which it inherited, is illustrated by
the Victoria glacier in the Canadian Rockies, a glacier

PLATE 12.

A hanging glacieret, the Triest glacier, above the lower stretch of the
great Aletsch Glacier, Switzerland.
(After a photograph by I. D. Scott.)

CLASSIFICATION OF MOUNTAIN GLACIERS 51

having no cirque, but only a couloir (the so-called " death-
trap ") in its stead (see Fig. 21). In this case the neve
which feeds the glacier is found high above upon the cliff
a true cliff glacieret and this neve avalanches its com-
pacted snow upon the surface of the Victoria glacier, which
thus well illustrates the reconstructed type. 31

Again, glaciers may develop, not upon a gently domed and
variously moulded pre-glacial upland such as we have thus
far had under consideration, but upon the sharply conical
volcanic peaks which in temperate and tropical regions push
their heads from the mountain upland at their base far up
above the snow-line. In such cases, regular cirques cannot
develop at the heads of the radiating ice-streams, but, on
the contrary, very irregular and mutually destructive forms
will result (see plate 13 B). 32 This is the more true because
of the loosely consolidated tuffs of which such cones are
always built up. If sufficiently lofty, the result may be a
small carapace or ice-cap such as is found to-day upon the
summit of Kilimandjaro in Africa. On the other hand, a
partially ruined crater may furnish a natural basin or caul-
dron for a small glacier cauldron type. 53

Tide- water Type. In high latitudes glaciers sometimes
descend to the level of the tide-water in fjords which con-
tinue their valleys. In such cases, the glacier front is
attacked mechanically by the waves and is further melted
in the water. In place of the convexly rounded nose, so
characteristic of the other types, there develops a precipitous
cliff of ice from which bergs are calved, and the glacier front
in consequence is rapidly retired (plate 11, 6). Unhappily,
the local term " living glaciers " has been applied to this
type in Alaska; " dead glaciers," in the same usage, being
applied to glaciers which yield no icebergs. The slopes of
the glacier surface and the measure of projection of the ice
above the water-level both render it probable that in most

52 CHARACTERISTICS OF EXISTING GLACIERS

cases, at least, the ice-foot everywhere rests on a solid base-
ment. On the other hand, the Turner glacier, debouching
into Disenchantment Bay, Alaska, shows a flat and relatively
low front section, which is separated from the remaining and
sloping portion of the glacier by a steep ice-fall. This has
led Gilbert to think that the lowest terrace is floated in the
water. 34

Radiating (Alpine) Type. A good deal of misunder-
standing is current in regard to alpine glaciers, often unhap-
pily referred to as valley glaciers. Examination of any good
map of Switzerland suffices to show that with the possible ex-
ception of the Great Aletsch, an abnormal type, Swiss glaciers
hardly extend into valleys at all. We have too long held the
alpine glacier close before the eye, and so have much exag-
gerated its importance. When Alaskan, Himalayan, and
New Zealand glaciers are brought into consideration, the
real position of the Swiss type becomes apparent. In reality
the glaciers of the Alps, far from occupying valleys, do not
even fill the mother cirques at the valley heads. Here they
lie, side by side, joined to one another like the radiating
sticks within a lady's fan, for which reason they have some-
times been called Zusammengesetzte Gletscher (see Fig. 22
and plate 11, 7). The mer de glace, next to the Great Aletsch
the largest in Switzerland, with its numerous tributaries, it
is true, completely fills a cirque, but only that of a tributary
valley (plate 11, 8). 35 Alpine glaciers are hence sheaves of
small glaciers or glacierets which start out from the second-
ary scallops of the mature cirque. They are wholly included
within the mother cirques, or fill and extend out from the
secondary or tributary cirques. In the Nicolai valley of
Switzerland, the Gorner glacier and its several tributaries
(see Fig. 22), with the Findelen and Langenfluh, the Theo-
dul, Furgen, and Z'Mlitt glaciers together, but partially fill
the mother cirque ofjwhich Zermatt is the centre. Lining

PLATE 13.

A. A hanging tributary valley meeting a trunk glacier valley above the present
water-level on the "inside passage" to Alaska.

B. Irregularly bounded neves upon the volcanic cone of Mt. Ranier.

CLASSIFICATION OF MOUNTAIN GLACIERS

the valley below upon either side are eighteen to twenty
glacierets, all resting upon the albs, or high mountain
meadows.

High up in the Chamonix valley, below the debouchure of
the mer de glace, similar glacierets are lodged upon the ledge

i.i.i

FIG. 22. Outline map of a radiating glacier, in the Nicolai valley, Switzerland.

below the sharp needles of de Charmoz, de Blatiere, du Plan,
and du Midi, their frontal moraines making a continuous
series of scallops above the shoulder of the valley. Similar
but smaller series are shown in Fig. 23 and pi. 14 A.

Horseshoe Type. The final representative type in our
series, unlike the alpine glacier, is no longer made up of ice-
streams joined together in sheaves. With further shrinking
of alpine glaciers corresponding to higher air temperatures,
the glacier front retires until it approaches the cirque wall.
It now takes on, either as an individual or as a collection of
small remnants, a broadly concave margin, which is in con-

CHARACTERISTICS OF EXISTING GLACIERS

trast to the convex or convexly scalloped front character-
istic of all other glacier types. This type of glacieret has
been sometimes described under the names " hanging " and
" cliff glaciers." 36 Reasons have been presented for restrict-
ing both these terms to special and different varieties of small
glaciers or glacierets. It is proposed to use here the term
" horseshoe glacier" for these last remnants of larger glaciers

FIG. 23. Outline map of a horseshoe glacier, the Asulkan glacier in the Selkirks.
The dashed line is the divide.

hugging the wall of the cirque. Most of the glaciers of
North America outside of Alaska belong in this class. As
already implied, they are generally broader than long, and
usually have concave frontal margins. Excellent examples
of this type are furnished by the Horseshoe glacier at
the head of the Paradise valley in the Canadian Rockies
and by the Asulkan glacier in the Selkirks (see Fig. 23 and
plate 14 A). The Mount Lyell glacier, long known and cited

PLATE 14.

A. Series of hanging glacierets which extend the Asulkan glacier in the Selkirks.

View of the Wenkchemna glacier at the head of the valley of the Ten Peaks in
the Canadian Rockies.

CLASSIFICATION OF MOUNTAIN GLACIERS

miles

from the High Sierras of California, is, however, an equally
good type. 37 For further illustration of the type the Wenk-
chemna glacier in the Canadian Rockies has been chosen (see
Fig. 23, plates 11, 2 and 14 B). The Asulkan and Wenk-
chemna glaciers
have both been de-
scribed by Scherzer
as belonging to the
piedmont type. The
former hugs a cirque
wall with an in-
curving frontal mar-
gin, and is extended
by a series of small
hanging glacierets
(see plate 14 A).
Unlike the piedmont
glaciers, it has no
foreland on which to expand, but lies in a cirque at the head
of a typical U-shaped valley. The Wenkchemna glacier oc-
cupies a similar position in the great cirque outlined by the
Ten Peaks at the head of a valley tributary to the Bow (see
Fig. 24) , 38

In plate 11 the various types of glacier are shown on
approximately the same scale, and from this it will be appre-
ciated that the size, directly dependent upon the alimenta-
tion of the glacier, must be a determining factor in classifica-
tion. The ice-cap and piedmont glaciers will in this respect
overlap, being differentiated by the accentuation of the relief
of the land. For the other types the proportion of the
glacier-carved valley which is still occupied by the ice will
determine the form and the more important characters of
the existing glacier. It is important, therefore, in order to
determine the type to which an individual glacier belongs,

FIG. 24. Outline map of the Wenkchemna glacier
in the Canadian Rockies. The dashed line is the
divide.

56 CHARACTERISTICS OF EXISTING GLACIERS

to map the divide surrounding the valley, as well as the
boundaries of the glacier which lies within it. It will be
shown later that in Antarctica, where melting of snow or
ice occurs only under exceptional and local conditions some
additional glacier types are encountered (see chapter xv).

REFERENCES

1 R. S. Tarr, " Valley Glaciers of the Upper Nugsuak Peninsula, Green-
land," Am. Geol, vol. 19, 1897, p. 265 and fig. 2.

2 More fully described under Part II.

3 T. C. Chamberlin, "Glacial Studies in Greenland," IV., V., Jour.
Geol., vol. 3, 1895, pp. 199, 470.

4 Th. Thoroddsen, "Island, Grundriss der Geographic und Geologie,
V. Die Gletscher Islands," Pet. Mitt., Erg. Bd. 32 (Nos. 152-153), 1906,
pp. 163-208, map, pi. xii.

5 H. Hess, 'Die Gletscher' (map 3).

6 Emil Werth, "Aufbau und Gestaltung von Kergulen." Sonderabd.
aus Deutsch. Siidpolar Expeditionen, 1901-1903, vol. 2, pp. 93-183, pis.
9-14, 3 maps.

7 Hans Meyer, 'Der Kilimandjaro, Reisen und Studien,' pp. 436.
Berlin, 1898 (reviewed by Rabot).

8 1. C. Russell, "An Expedition to Mount St. Elias," Nat. Geogr. Mag.,
vol. 3, 1891, pp. 52-204, pis. 2-20. See also Filippi, loc. cit.

9 Roughly outlined on map of Alaska to accompany "The Geography
and Geology of Alaska," by Brooks (Prof. Paper U. S. Geol. Surv., No. 45,
1906, plate in cover). For details of marginal portion and description, see
G. C. Martin, Bull. 335. U. S. Geol. Surv., 1908, pp. 46-48, and pis. 1, 2, and 5.

10 E. Blackwelder, "Glacial Features of the Alaskan Coast between
Yakutat Bay and the Alsek River," Jour. Geol., vol. 16, 1907, pp. 428-
432, map.

11 See Rabot, 'La Geographie,' vol. 3, 1901, p. 270. See also Hess, 'Die
Gletscher,' p. 63.

12 Penck u. Bruckner, 'Die Alpen im Eiszeitalter,' especially vol. 2, 1909,
map opposite p. 396.

13 Fred H. H. Calhoun, "The Montana Lobe of the Keewatin Ice-sheet,"
Prof. Paper No. 50, U. S. Geol. Surv., 1906, pp. 14-21, map, pi. 1.

14 Bull. Am. Geogr. Soc., vol. 38, 1906, p. 149. See also Prof. Paper No.
64, U. S. Geol. Surv., 1909, pp. 35-36, 105, pis. vii-viii.

15 Jour. Geol, vol. 16, 1907, p. 432.

16 G. C. Martin, Bull. 284, U. S. Geol. Surv., 1906, pi. 12.

17 Filippi, loc. cit.

18 Argentine-Chilian boundary, maps.

19 G. C. Martin, loc. cit.

20 Tarr and Martin, "The National Geographic Society's Alaskan Ex-
pedition of 1909," Nat. Geogr. Mag., vol. 21, 1910, p. 25.

CLASSIFICATION OF MOUNTAIN GLACIERS 57

21 C. W. Hayes, "An Expedition through the Yukon District," Nat.
Geogr. Mag., vol. 4, 1892, pp. 152. See also map of Mendenhall and
Schrader, Prof. Pap. U. S. Geol. Surv., No. 15, 1903, fig. 4, p. 41.

22 G. C. Martin, "Geology and Mineral Resources of the Controller Bay
Region, Alaska," Bull. No. 335, U. S. Geol. Surv., 1908, pp. 48-49, pi. i.
ii. and v.

23 One of the best maps of such a restored valley glacier of Pleistocene
age is that of the Kern valley of California (see Lawson, loc. cit., pi. xxxi.).

24 W. M. Conway, ** Climbing and Exploration in the Karakoram Him-
alayas," maps and scientific reports, 1894. See also Fanny Bullock
Workman and William Hunter Workman, "The Hispar Glacier," Geogr.
Jour., vol. 35, 1910, pp. 105-132, 7 pis. and map.

25 E. J. Garwood, "Notes on Map of the Glaciers of Kangchenjunga,
with remarks on some of the Physical Features of the District," Geogr.
Jour., vol. 20, 1902, pp. 13-24, plate.

26 Max Friedrichsen, " Die heutige Vergletscherung des Khan- Tengri-
Massives und die Spuren einer diluvialen Eiszeit in Tion Schan," Zeitf.
Gletscherk., vol. 2, 1908, pp. 242-257.

27 R. v. Lendenfeld, "Der Tasman Gletscher und seine Umrandung,"
Pet. Mitt., Erg. Bd., vol. 16, 1884, pp. 1-80, map, pi. 1.

28 W. C. Mendenhall and F. C. Schrader, "The Mineral Resources of the
Mount Wrangell District, Alaska," Prof. Pap. U. S. Geol. Surv., No. 15.
1903, pi. iv and ix. See also Brooks, Prof. Pap. U. S. Geol. Surv., No. 45,
map, pi. xxxiv.

29 R. S. Tarr, "Glacier Erosion in the Scottish Highlands," Scot. Geogr.
Mag., vol. 24, 1908, pp. 575-587.

30 The term "hanging glacier," now used in a variety of senses, is, it is
believed, best retained with the restricted meaning. The term "cliff
glacier," generally considered synonymous, may be restricted to the long
strips of incipient glacier ice which sometimes parallel the main valleys on
narrow terraces above precipitous cliffs which are primarily determined
by the rock structure (see ante, p. 54; and also Matthes, Appalachia,
vol. 10, 1904, p. 262). In the sense here employed, a hanging glacier is
generally the equivalent of the Kahrgletscher, a term quite generally em-
ployed in Germany. The term "horseshoe glacier" we have here sug-
gested for an essentially different type of glacieret (see p. 53).

31 See map and description of this glacier by Scherzer, "Glaciers of the
Canadian Rockies and Selkirks," Smith. Contrib., No. 1692, 1907, chaps.
2-3.

32 Cf. I. C. Russell, "Glaciers of Mount Ranier," 18th Ann. Rept. U. S.
Geol. Surv., 1898, pp. 329-423.

33 Hans Meyer, "Der Calderagletscher des Cerro Altar in Equador,"
Zeitsch. f. Gletscherk., vol. 1, 1906-1907, pp. 139-148.

34 G. K. Gilbert, * Harriman Alaska Expedition,' vol. 3, " Glaciers," 1904,
pp. 67-68. See also Tarr and Butler, " The Yakatat Bay Region, Alaska,
Physiography and Glacial Geology." Prof. Paper No. 64, U. S. Geol. Surv.,
1909, pp. 39, 40, pi. xa.

58 CHARACTERISTICS OF EXISTING GLACIERS

35 This valley is a large hanging valley tributary to the Chamonix val-
ley, which latter alone is comparable in size to those that form the beds
of the Baltoro, Hispar, and Tasman glaciers. If at first it seems that con-
fusion may result from the introduction of valleys of different orders of
magnitude, a second thought suffices to show that the difficulty is of
theoretical rather than of practical importance, at least so far as existing
examples of glaciers are concerned.

36 See footnote on p. 57.

37 1. C. Russell, "Existing Glaciers of the United States," 5th Ann.
Kept. U. S. Geol. Surv., 1885, pp. 314-328, pi. 40.

38 Sherzer, Smith. Contrib., No. 1693, 1907, chaps, iv. and vii. The only
resemblance to the piedmont glacier is in the shape. Neither glacier ex-
pands upon a foreland, but both lie in cirques at the heads of U-shaped
valleys. They have no appreciable tributaries, and, as already pointed
out, piedmont glaciers are necessarily of large size, corresponding to ex-
cessive precipitation.

CHAPTER IV

LOW LEVEL GLACIAL SCULPTURE IN MODERATE
LATITUDES

The Cascade Stairway. No one who has climbed a moun-
tain glacier to its neve has failed to be struck by the alter-
nation of plateau and precipitous slope, for the surfaces of
mountain glaciers are, with few exceptions, broken into
broad terraces. Each steep descent is well understood to
overlie a corresponding fall in the glacier-bed. Perched upon
the high cliffs which overlook the Pinnacle pass during his
first attack upon Mount St. Elias, the late Professor Russell
wrote of these terraces : 1 -

Were the snow removed and the rock beneath exposed, we
should find terraces separated by scarps sweeping across the bed
of the glacier from side to side. Similar terraces occur in glaciated
canyons in the Rocky Mountains and the Sierra Nevadas, but their
origin has never been explained. The glacier is here at work
sculpturing similar forms, but still it is impossible to understand
how the process is initiated.

The generalized description of uncovered glacier-beds
within the High Sierras of California perhaps as well as
any that has been penned lays the emphasis upon the
more essential and impressive characters : 2 -

" The amphitheatre bottom terminated forward in either a cross
cliff or a cascade stairway, descending, between high walls, to yet
another flat. In this manner, in steps from flat to flat, common

60 CHARACTERISTICS OF EXISTING GLACIERS

enough to be characteristic, the canyon made descent (see Fig.
25 and plate 15). In height, however, the initial cross cliff at the
head dominated all. The tread of the steps in the long stairway,
as far as the eye could follow, greatly lengthened in down-canyon
order.

\\zoo-

IOZOO'-

9SOO'-

FIG. 25. Longitudinal section along a glaciated mountain valley, showing re-
versed grades and rock basin lakes in series. Vertical scale about two and one
half times the horizontal (after Salisbury and Atwood 3 ).

The grade on the treads is often reversed, so that rock
ridges separate basins or colks, and these latter come to be
occupied by the characteristic glacial lakes. High up in the
valley, where the treads are relatively short, these lakes are
more ar less kettle-shaped, though relatively shallow, and
they usually rest directly upon the rock. They are, there-
fore, often referred to as rock frasw lakes, though a morainal
dam sometimes plays a part in impounding the water
(Fig. 44, p. 82). Often connected together like pearls upon a
thread, or, better still, like the larger beads in a rosary, they are
sometimes referred to as pater noster lakes 4 (see Fig. 12, p. 28).
Lower down in the valley and upon the longer treads, lakes
are more apt to be long and ribbonlike in form (see plate 15).

Mechanics of the Process which produces the Cascade
Stairway. Since Russell's meditation above the Pinnacle
pass, nearly a score of years ago, considerable study has been
given to the subject of erosion upon the glacier-bed. In
the Alps, Penck and Bruckner have enunciated their " law
of adjusted cross-sections." The glacier, on invading the
mature river-valley, characterized by uniformly forward
grades and by accordance of trunk with side valleys, will,
in general, be so modified that a small cross-section corre-
sponds to a deepening of the valley. 5 Thus may be brought

PLATE 15.

Land surface moulded by mountain glaciers near the ancient Lake Mono, east
of the Sierra Nevadas in California (after Russell).

GLACIAL SCULPTURE IN MODERATE LATITUDES 61

about the hanging side valley, and a local modification of,
and perhaps even a reversal of, direction in the grade of the
main valley.

If the rock be not homogeneous throughout, or if it be
unequally intersected by joint planes, further abrupt changes
in grade will result. The two processes which are effective
in deepening the bed of the valley are well recognized to
be abrasion and plucking. Greater softness in the rock
will correspond to greater depth of abrasion, while the
perfection of the parting planes will directly determine
the amount of quarrying in the rock by plucking. Abra-
sion being greatest on the upstream side of any irreg-
ularity in the bed, and plucking being largely restricted to
the downstream side, the tendency of these processes work-
ing together will be to produce steps of flat tread but steep
riser, the latter coinciding with the nearly perpendicular
planes of jointing. To quote de Martonne, 6 " the mass of the
ice does not rest everywhere upon its bed, and in particular
upon the risers of steps (Mer de Glace, Fiesch, Rhone glacier).
Speaking generally, the contact becomes closer with each
diminution of the down slope; it tends to be relaxed with
each increase of the slope."

It is further probable that the cliffs at the lower margins
of the terraces are in many cases, at least, considerably re-
cessed through the operation of a sapping process in every
way analogous to that which obtains at the base of the
Bergschrund, or Randspalte. So soon as the rock cliff has
been formed, either below a narrowing of the valley or
where a hard layer of rock transects it, the glacier will de-
scend over it in an ice-fall, showing gaping transverse cre-
vasses. These fissures in the ice may be sufficiently profound
to admit the warm air at midday to the rock joints, and so
bring about with the nightly fall of temperature a me-
chanical rendering of the rock.

CHARACTERISTICS OF EXISTING GLACIERS

Basal cliff sapping being downward as well as backward,
the reversed grades of the treads in the staircase could be
thus explained. In the Alps, Penck distinguishes especially
one larger cliff in the staircase which separates the head
cirque from the trough valley (Trogthal). 7

The extended studies of Penck and Bruckner upon the
Alps have shown that as a general rule the risers of the steps
are found just above the junction of the main valley with

FIG. 26. Rock bar with basin showing above, from the Upper Stubaithal near
the Dresdner Hiitte (after Bruckner).

its tributaries. 8 Thus the main glacier stream is here re-
inforced by large contributions of ice and accomplishes a
larger amount of excavating upon its bed. Sausage-like
the valleys widen below the steps so that deeper basins

GLACIAL SCULPTURE IN MODERATE LATITUDES 63

alternate with higher narrows and afford a certain corre-
spondence between the plan and the profile of the valley.

Owing to the backward tilt of the treads within this cascade
stairway, their outer edges rise from the sanded and in part
flooded floor in the form of a rocky bar which crosses the
valley from side to side. These bars are the well-known
Riegel 9 or verrous 10 of the Swiss Alps, which for want of a
better English designation we may term "rock bars." The
topographic form of such bars is well brought out in Fig.
26. In many Alpine valleys such bars are quite numerous,
no less than eight being encountered in a walk down the
Haslithal from the Grimsel to Meiringen. The largest and
best known of these is the famous Aarschlucht near Meirin-
gen. Many of the larger Riegel are found to correspond in
position to the outcropping of a zone of limestone, which,
being less easily eroded by the glaciers than is the sur-
rounding gneiss rock, has in consequence been left in relief. 11

The U -Shaped Glacier Valley. --To-day it is everywhere
recognized that one effect of the occupation of valleys by
mountain glaciers is to so transform them that the cross-
section has the form of a letter U. The steepness and the
height of the side walls will, in hard rocks at least, be to
some extent a function of the depth to which the valleys
have been filled by the glaciers. Thus the Little Cotton-
wood canyon on the western front of the Wasatch range, so
often cited and figured as a typical U-valley (see plate 16 A),
is one in which the ice-foot pushed out but a short distance
beyond the portal of the valley. At this point, therefore,
the valley was occupied by ice to a very moderate depth,
and it is the bottom portion only which betrays the curve
of the U-section. In the higher Alpine valleys, on the
other hand, which were once filled to a much greater depth,
the steep undercut side walls often complete the form of the
letter U. Their intersection with an earlier valley located

64 CHARACTERISTICS OF EXISTING GLACIERS

on the same general line, but at a higher level, has developed
rather sharp shoulders. These remnants of the earlier
valley are the albs, or high mountain meadows, so common
along the Swiss valleys (see Fig. 27).

The form of these remnants of older and now higher
valley floors, is not that of a water-worn valley, but gener-
ally is a relatively shallow
glacier-carved trough. They,
therefore, indicate that since
sluggish glaciers carved the
earlier valley, a new uplift
of the range has taken place. 12

FIG. 27. -Ideal cross-action of a U-shaped Subsequent to the Uplift,

valley once occupied by a mountain the glacier acquired a steeper

gradient and carved its bed

below the middle of the older U -valley, as are the Norwe-
gian valleys likewise to be explained through an uplift of
the land.

It is clear that the widening of the valley bottom is accom-
plished by the ice through the combined abrading and pluck-
ing processes. As is true of so many geological processes,
the direct attack is here through a limited range only, but
is extended upward and made more effective through under-
mining or sapping. The dividing line between the vertical
zones of direct and indirect action of ice erosion and of
undermining is often a sharp line. The upper zone
quarried by the undermining process, here always greatly
facilitated by frost rending, develops irregular but nearly
vertical (joint) surfaces. The lower eroded surface, on the
other hand, is rounded into shoulders (roches moutonnees),
and is further smoothed and scratched (see Fig. 28 arid
plate 16 B). This line of sharp separation may be con-
tinued up the valley and there be joined to the schrund line
of the cirque (see Fig. 6, p. 18).

PLATE 16.

A. The Little Cottonwood Canyon in the Wasatch Range transformed at the
bottom into the characteristic U section.

(After a photograph by Church.)

B. Striated surface of glaciated valley floor near Loch Coriusk, Skye.
(.From a photograph by B. Hobson.)

GLACIAL SCULPTURE IN MODERATE LATITUDES 65

The areas of the valley section taken at different levels
obviously stand in direct relation to the size of the glacier
at those levels. When the glacier ended within the valley
(dendritic glacier), ablation in the lower levels diminished

FIG. 28. View in the glaciated Sierra Nevadas of California, showing the sharp
line which sometimes separates the zone of abrasion from that of sapping (after
a photograph by Fairbanks.)

the width of the ice-stream as its foot was approached. In
some cases, the glacier never reached the margin of the up-
land, in which event the lower portion of the valley is
relatively narrow and reveals the characteristic section of
a river-carved valley. Even when widened by glaciation,
the widest section may be found considerably above the
lowest limits of the ice advance. This is illustrated by
Big Cottonwood canyon in the Wasatch range, which at its
portal is a narrow V-shaped valley, but which above has
been widened by glaciation. 13

Wherever, on the other hand, mountain glaciers have been

CHARACTERISTICS OF EXISTING GLACIERS

so amply nourished as to expand beyond the margin of the
upland, the valley is found to widen rapidly towards its
mouth and expands to the foreland in trumpet form. This
is well illustrated by the portals of the larger Alpine valleys,
which once supported piedmont glaciers (see Fig. 47, p. 85).
It has been urged, by those who regard the glacier influ-
ence as always protective to its bed, that the deep U -valleys
have in pre-glacial or in inter-glacial times, been cut down by
rivers, and that these narrow valleys the glaciers have sub-
sequently widened. Some part the gorges of mountain
streams must have played, particularly in inter-glacial
times when the glacier-formed rock bars had been sawed
through by the torrents which followed the retreat of the
glacier from portions of its valley.

While nearly all glacialists seem to be agreed that a widen-
ing of valley bottoms results from the occupation by moun-
tain glaciers, many are unwilling to admit that there is in
addition, a deepening of the valley through the action of the
same processes. To the present writer the evidence for the

overdeepening in-
herent in the cross
profiles of valleys
is convincing.

The Hanging Side
Valley. - - Whereas
under normal con-
ditions of sub-aerial
erosion, the indi-
vidual tributary
valleys meet the
main valley at a
common level, or
accordantly (see Fig. 29), this is not true of glaciated
valleys. Since the smaller tributary glaciers are unable to

FIG. 29. Normal valleys from sub-aerial erosion
accordant drainage (after Davis).

GLACIAL SCULPTURE IN MODERATE LATITUDES 67

erode their beds as effectively as the larger trunk streams,
when both have been vacated by the ice, side valleys are
found to have their beds standing above the general level
of the main valley they are not accordant as are the trib-
utary valleys of

Wt;^

^M//w

rivers and they
are in consequence
spoken of as "hang-
ing valleys " (see
Fig. 30). Unlike
those tributaries
which have never

been OCCUpied by FIG. 30. Glaciated and non-glaciated valleys tribu-

tary to a glaciated main valley. Both types of
side valley are hanging (after Davis).

glaciers, they are
found to be too
large for the streams which now flow in them. This stream
drops over the steep U-wall into the main valley in the
characteristic ribbon type of waterfall found in such num-
bers in every glaciated mountain district.

As pointed out by Penck, it is the surfaces only of main and
tributary glaciers that are accordant, or at common level.
It is, perhaps, profitable to consider for a moment why it 13
that the tributaries of water-streams should, under normal
conditions, be accordant, as was long ago pointed out to be
the rule by Play fair; whereas the beds of tributary glacier
streams enter the main valley above the level of its floor.
In both cases the tributaries are notably smaller than the
main stream. The abrading process by which the water-
stream lowers its bed is in no wise dependent upon the depth
or volume of water, for water-streams have a cutting power
directly determined by the gradient of their bed and increas-
ing at a marvellous rate with increase of slope. Now trib-
utary valleys in mountain districts have gradients which
are much steeper than that of the main valley near the point
of their junction (see Fig. 31).

CHARACTERISTICS OF EXISTING GLACIERS

In the glacier stream floor, gradient evidently plays a
much less important role in the abrasion of the bed, while
depth of ice would appear to be a determining factor, the

FIG. 31. Comparison of the longitudinal profiles of a mature stream-cut valley
and its tributaries with a glacier-carved Alpine valley and its tributaries. Note
how in both instances the average gradient of the tributaries is always in excess
of that of the main valley, near the junction (after scaled profiles prepared by
Nussbaum 14 ).

friction between the stones by which the glacier is shod and
the rock floor causing a correspondingly greater wear. The
hollowing of flagstones is proportional, not only to the num-
ber of footsteps which have come in contact with the stones,
but also upon the weight of the individuals and the number
of projecting nails upon their boot heels.

REFERENCES

1 1. C. Russell, "Expedition to Mount St. Elias," Nat. Geogr. Mag.,
vol. 3, 1891, pp. 132-133.

2 Johnson, Jour. GeoL, vol. 12, 1904, pp. 570-571.

3 The interpretation of topographic maps, Prof. Pap. No. 60, U. S.
GeoL Surv., 1908, p. 66.

4 Nussbaum, " Die Taler der Schweizeralpen," Bern, 1910, p. 28.

GLACIAL SCULPTURE IN MODERATE LATITUDES 69

6 A. Penck, Jour. GeoL, vol. 13, 1905, pp. 1-19.

6 Em. de Margerie, "Sur I'inegale repartition de 1'erosion glaciare dans
le lit des glaciers alpins," C. R. Acad. Science, Paris, December 27, 1909,
pp. 1-3 (reprint).

7 See also Nussbaum, "Die Taler der Schweizeralpen," Bern, 1910,
pi. 2, figs. 2-4.

8 Ed. Bruckner, "Die glazialen Ziige im Antlitz der Alpen," I.e., 1910,
p. 787 ; also Fritz Nussbaum, " Die Taler der Schweizeralpen, Eine
geographishe Studie," Bern, 1910, 3 pis. and 12 figs.

9 Bruckner, I.e., p. 787.

10 De Martonne, "Sur la genese des formes glaciares Alpines," C. R.
Acad. Sci., Paris, January 24, 1910, p. 1 (reprint).

11 Bruckner, I.e., p. 790.

12 Albrecht Penck, "The Origin of the Alps," Bull. Am. Geogr. Soc., vol.
41, 1909, p. 68.

13 w w. Atwood, "Glaciation of the Uinta and Wasatch Mountains,"
Prof. Pap. U. S. Geol. Survey, No. 61, 1909, pp. 85-88, pi. x.

"Nussbaum, I.e., 1910, final plate.

CHAPTER V

HIGH LATITUDE GLACIAL SCULPTURE

Variations in Glacial Sculpture Dependent upon Lati-
tude. Thus far we have considered mountain glaciers in
those districts which are most accessible for study, and hence
are better known mountain districts within moderate
latitudes, in which the snow-line is from 7000 to 12,000 feet
above sea, and where, in consequence, there is high relief
and correspondingly steep gradients. Moreover, in most
of these districts the surface upon which the mountain
glaciers whose handiwork we may study, began their carv-
ing process was a surface moulded by the water-streams of a
humid region the initial surface in the glacial cycle was
a product of sub-aerial erosion. The results are not in all
respects the same in those higher latitudes where the snow-
line descends to near the sea-level, and where in Pleistocene
times, a continental glacier largely planed away the irregu-
larities of earlier erosion periods, leaving a hard rock sur-
face, but slowly acted upon by the well-known weathering
processes.

The low level of the snow-line is here further responsible
for the development of the subsequent mountain glaciers
where there is only moderate relief, so that glacier streams
developed on low gradients were notably sluggish in their
movements. Inasmuch as the sub-polar regions particularly

HIGH LATITUDE GLACIAL SCULPTURE

have been characterized within the recent geological period
by rather remarkable uplifts of the land, this elevation has
had an important bearing on the origin of the surface fea-
tures there developed.

Surface Features of Northern Lapland. A visit to
Northern Lapland is in this regard most enlightening to one
who has observed glacial carving in lower latitudes only.
In the lower levels of these Northern regions, which lie to the
eastward, where the land was relatively low, and where the
prevailing winds had already given up much of their mois-
ture, mountain glaciers have, in consequence, found little
to nourish them. Here is found to-day a surface of low,
bare hills, rounded and smoothed and betraying the
sculpture of continental ice masses alone (see Fig. 35 a).

TIG. 32. Characteristic surface in Swedish Lapland which has been moulded
mainly by the continental glaciers of Pleistocene times. The low central trough
is, however, the work of a subsequent mountain glacier on a low gradient
the Karso trough valley (after O. Sjogren).

The surface which still shows the features moulded by
erosion beneath the continental glacier, would appear to
extend far to the northeastward. To quote Feilden * on the
Kola Peninsula : " As we sail eastward along the Murrnan

CHARACTERISTICS OF EXISTING GLACIERS

coast of Russian Lapland, we see on our right hand a bold
and precipitous country. Its highest summits appear to rise
to 500 or 600 feet. The hills are planed down to a general
level, and no peaked mountain breaks the monotony of the
scene." 2

The Flatly Grooved Glacier Valleys and the Scattered
Knobs. In the somewhat higher levels farther to the west-
ward, but before the high Norwegian plateau is reached, the
handiwork of mountain glaciers is recognized, though no ice
masses are here in evidence to-day. Locally, where were
centres of dispersion, the characteristic " arm-chair " form
of the glacial cirque is to be seen, and well developed karlings
are made out, though here the jagged pinnacles so common
in lower latitudes, are seldom seen (see Fig. 33). Out from

FIG. 33. Map of a portion of the area south of Tornetrask in Swedish Lapland
showing the cirques and karlings developed by mountain glaciers subsequent to
the continental glaciation.

these centres of late mountain glaciation, the sluggish glacier
streams have channelled broad and gently hollowed grooves
within the former undulating surface. These shallow val-
leys have but little in common with the deep U -channels

HIGH LATITUDE GLACIAL SCULPTURE

of the Alpine highland. Where these streams have been
numerous and of nearly equal size, they have coalesced and
so, to a large extent, have occupied the country, leaving a
smoothed floor out of which more or less elongated knobs
of rock rise on abrupt or even precipitous slopes (see Figs.
33 and 34). Where relatively large streams have channelled

FIG. 34. Glacial surface crossed by shallow channels carved by sluggish mountain
glaciers on a surface of slight relief. The area is south of Tornetrask in Swedish
Lapland (after O. Sjogren).

the surface guided by pre-existing valleys, the horizon lines
now show broad depressions resembling the bite of some
gigantic monster. The Lapporten, or Lapp's Gate, which is
seen in so many views about Tornetrask, furnishes a strik-
ing illustration (see Fig. 35 b).

The Fjords of Western Norway. In the plateau region of
Norway still further to the westward, where the heavier
precipitation and the much higher elevations have made it
possible for glaciers to persist to the present day, such flat
U -channels may now be seen high up upon the level of the
plateau (see Fig. 35 c). Similarly the steep-sided knobs
(nunataks) which are found to characterize the relatively

CHAEACTERISTICS OF EXISTING GLACIERS

FIG. 35 a. Characteristic surface con-
tours in Eastern Swedish Lapland due
to sculpturing by continental and moun-
tain glaciers.

6. Horizon line showing the modifica-
tion of surface moulded by continental
glaciers, through the later work of slug-
gish mountain glaciers. The deep ' ' bite
in the horizon line is the famous Lapp's
Gate.

c. Flat U-channel on plateau of West-
ern Norway; Geirangerfjord.

d. A steep rock knob rising from the
plateau of Norway ; Oxenelvene on the
Nordfjord.

low surfaces of Swedish
Lapland are here to be seen
rising out of the level of the
plateau (see Fig. 35 d).

In these sections of Scan-
dinavia, the problems of
glacial sculpture are much
more complex, and are not
to be solved by considera-
tion of the latest glaciation
only. As is now well recog-
nized, the Pleistocene glaci-
ation consisted of some
four distinct glacial cycles,
separated by inter-glacial
^ periods which were charac-
terized by relatively mild
climatic conditions. An
earlier submergence of the
coast regions (see plate 17,
B) has been followed by
large and rapid uplifts, so
that former strand lines are
now to be seen high up
upon the shores. Of the
origin of the fjords the
deep and now partially
submerged U -valleys we
know at least that their
present form was given
them when they were occu-
pied by glacier streams; 3
and their definitely oriented
arrangement further be-

HIGH LATITUDE GLACIAL SCULPTURE

trays the fact that the glacial excavation exercised a selec-
tive process on the lines of preexisting fractures within the
rocky basement, guided, perhaps, on these lines by earlier
rivers which had first discovered these special lines of weak-
ness within the earth's surface shell (see Fig. 36).

FIG. 36. Map of the vicinity of the Storf jord in Norway, showing the regular
arrangement of the fjords and submerged valleys in three principal parallel
series separated by sub-equal space intervals.

The Rock Pedestals bounded by Fjords. The late uplift of
the coast subsequent to the formation of these deep fjords
has raised veritable pedestals of rock surmounted by rela-
tively flat surfaces, on which are revealed under exceptional
circumstances the characteristic subdued forms found in the
lower country of Northern Lapland (see Fig. 32, p. 71).
Since these pedestals now lift their heads above the snow-line
of the region, ice-caps are amply nourished upon them, so as
often to more than cover the surface and spill over the edges
wherever the margin has been notched by the earlier sculp-
turing. Near the centre of the pedestal the process of sub-
glacial abrasion pares down the inherited irregularities,

CHARACTERISTICS OF EXISTING GLACIERS

whereas near the margins where the ice is thinner and the
gradients are steeper, the inherited knobs have been greatly

FIG. 36 a. Nunataks rising out of the surface of the Folgefond, an ice-cap of

Southern Norway.

increased in height. Such a knob enveloped in the marginal
portion of a Norwegian ice-cap is reproduced in plate 17 A

FIG. 37. Erosional surface left within the marginal zone of a Norwegian ice-
cap. The smoothly domed floor and the steep projecting knobs are characteris-
tic. A moraine is in the foreground.

PLATE 17.

A. The Hardangerjokull, a plateau glacier of Southern Norway, where at its
margin is seen the Kongsnut nunatak.

B. Upland sculptured by mountain glaciers and partially submerged through de-
pression. Part of one of the Lofoten Islands.

(From a photograph by Dr. L. M. Hollander.)

HIGH LATITUDE GLACIAL SCULPTURE

and others appear in Fig. 36 a. Figure 37 shows, on the
other hand, the site of such a margin to an ice-cap, after the

FIG. 38. View of the Seven Sisters in Northwestern Norway, a series of ice-cap
nunataks sharpened by the overflow of the glacier streams at the margin.

ice has retired. Here we find a smoothly polished surface
descending on low gradients toward the margin of the ped-
estal. From this gently domed surface rise numerous knobs

FIG. 39. Broad glacial trough overdeepened by the overflow glacier of later ice-
cap.

of rock, the marginal nunataks, though here the ice has not
reached the margin of the pedestal so as to descend into the

CHARACTERISTICS OF EXISTING GLACIERS

surrounding fjords. (See pi. 34 A.) In Fig. 38 is seen
another example where the ice has spread over the edge of
its base and has deepened and widened the troughs upon its
margins, thereby sharpening the intervening knobs.

FIG. 40. Circular tind with acute apex from the Lofoten Islands. The Ten-
naes Tind, Kirkefjord (after a photograph by Dr. L. M. Hollander).

The Norwegian Tind. Where the overflow streams on
the margin of the ice-cap are of notably smaller dimensions

PLATE 18.

A. The Fjaerlandsfjord on the margin of the Jostedalsbraen, showing the nu-
nataks inherited from an earlier cycle as they develop into tinds by the over-
flow ice streams deepening the channels which separate them.

B. Nykerne, Vesteraalen. Typical tinds formed on the margin of Norwegian
plateau glaciers. A later product of the process shown above, in A.

HIGH LATITUDE GLACIAL SCULPTURE 79

than their predecessors of an earlier cycle a sharp shoulder
has developed on either side of the valley (see Fig. 39).
The deepening of the channels of small overflow glacier
streams, if continued, so lowers the marginal channels as to
transform the inherited nunataks into high peaks sometimes
having the form of bee-hives, on which the sharp change in
slope marking the transition from the earlier to the later
sculpture can often be made out. In plate 18 A an existing
Norwegian glacier is seen modifying the nunataks in this man-
ner, while in B of the same plate the later effect of the process
is to be observed. These steep rounded peaks are the character-
istic tinds of the Norwegian coast. They are markedly circular
at their base, they rise at first on excessively steep slopes, but
at greater heights take on gentler gradients, often showing
a sharp change in curvature, as is indicated in plates 18 B
and 34 B.

In the Lofoten Islands, a western outlier of the Norwe-
gian plateau to the north of the Arctic circle, tinds have de-
veloped apparently by this process, though they have here
taken a sharply conical form with almost circular base, so
that they resemble in form the point of a well-sharpened pen-
cil (see Fig. 40). Inasmuch as these develop in a massive

SNOW CAP SNOW CAP

FIG. 41. Successive diagrams to illustrate a theory of the shaping of acute
circular tinds through exfoliation.

igneous rock, a gabbro, and the surfaces indicate clearly that
the forms are now being shaped as a result of heavy exfolia-
tion, a suggestion may be hazarded with regard to the latest

80 CHARACTERISTICS OF EXISTING GLACIERS

stages of their evolution. A tind shaped by overdeepening
on the margin of an ice-cap (see plate 18 B and Fig. 41) is
by reason of its steep sides able to support snow only upon its
summit. About its flanks the tind is scaled off and rendered
circular in plan. The protecting snow-cap 4 prevents this ac-
tion at the top, but this cap is melted at its margin where
warmed by radiation from the neighboring rock surface.
The water derived from this melting enters all cracks due to
exfoliation, thus greatly facilitating the process and prevent-
ing the formation of an overhanging rock cornice. The
stages of the process are suggested in the diagrams of Fig. 41.

REFERENCES

1 H. W. Feilden, "Notes on the glacial geology of Arctic Europe and
its Islands," Part II, Quart. Jour. Geol Soc., vol. 52, 1896, p. 726.

2 The italics are mine. W. H. H.

3 Fr. MachaSek, " Geomorphologische Studien aus dem norwegischen
Hochgebirge," Abh. d. k. k. geogr. Gesellsch. in Wien, vol. 7, 1908, pp. 1-
61, 11 pis. and a map.

4 In the long winter season.

CHAPTER VI

GLACIAL FEATURES DUE MAINLY TO DEPOSITION

Abandoned Moraines of Mountain Glaciers. Not only do
we find in valleys the marks of former occupation by moun-
tain glaciers in characteristic erosional forms the cirque,
the roches moutonnees, the U -valley, and the hanging side val-
ley but in many cases, at least, the evidence is supported
by characteristic glacial deposits. These deposits are natu-
rally less in evidence in the higher levels, where erosion has
been more active ; but toward the lower reaches the impor-
tance of glacial deposits rapidly increases. With the retire-
ment of the glacier up its valley, medial and ground moraines
come alike to occupy the valley floor, though the talus and
landslide conspire to cover the lateral portions from sight.

FIG. 42. Terminal and lateral moraines remaining from earlier mountain glaciers
which pushed out upon the flanks of the Sawatch range (after W. H. Holmes).

Wherever the glaciers have pushed out upon the foreland, and
there been halted for considerable periods, the lateral and
terminal moraines have been left as definite and often well
marked topographic features (see Fig. 42). Such terminal

G 81

CHARACTERISTICS OF EXISTING GLACIERS

moraines at the mountain front sometimes show the con-
tours of the expanded foot, and
quite generally also the first series
of recessional moraines (see Fig.
43 and plate 15).

Within the valleys and back
from the front of the range,
glaciers have also left behind
them series of recessional termi-
nal moraines to mark the princi-
pal halting places during their
retreat. These moraines in many
cases hold back the water of the
Bailey stream, forming morainal
of Little Cottonwood canyon lakes, as, for example, in Parker

of the Wasatch range (after ^ / J.T cc TVT i

Canyon of the Sierra Nevadas
(see plate 15), or Convict Lake
within the same general region

FIG. 44. Convict Lake, a lake behind a morainal dam in a glaciated valley of the
Sierra Nevadas in California (after a photograph by Fairbanks).

GLACIAL FEATURES DUE MAINLY TO DEPOSITION 83

The Tongue-like Basin before the Mountain Front.

Wherever with ampler nourishment glaciers have coalesced
and expanded to the proportions of the piedmont type, the
marginal moraine has acquired formidable dimensions and
may be miles in width, and of considerable height. 1 The
width of this ice deposit is extended toward the interior
of the ice-apron by a zone of drumlins cigar-shaped
hills of till whose longer axes are perpendicular to the mo-
raine. 2 Thus is built up a tongue-like basin often with sub-
ordinate marginal lobes, within which the glacier-apron rested
(Zungenbecken of Penck). Such a basin is shown in Fig.
49, p. 88.

Border Lakes. The study of the glaciers which in Pleis-
tocene times pushed out from the portals of the Alps upon the

FIG. 45. Map of the moraines and drumlins within and about the apron of the
piedmont glacier of the Upper Rhine (after Penck and Bruckner).

Swiss and Italian forelands, has proved most illuminating.
In Fig. 45 is reproduced a map of the moraines formed about
the apron of the great piedmont glacier which once occupied

CHARACTERISTICS OF EXISTING GLACIERS

the valley of the Upper Rhine and a portion of its foreland.
The central area of this basin is now occupied by the beautiful
Lake Constance, which in an earlier and higher stage ex-
tended past the border of the foot-hills into the Alpine valley.

Such lakes. are
found in many
similar sites of
piedmont ice-
aprons on the
borders of the
Alps, and have
been referred
to as border
lakes (Rand-
seen). Heavy
morainic accu-
mulations hem
them in upon
the outer mar-
gin, and their
waves lap the
rising slope of
the glacier val-
ley within its
gateway. An
instance in
which the plan
of the lake

brings out with especial clearness the relatively narrow
valley and the expanded form of the ice-apron without,
is Lake Garda (see Fig. 46). Geologically considered, lakes
are, however, notoriously short-lived, and the basins of
extinct lakes are found at the portals of most of the larger
Alpine valleys which have not existing lakes. 3 The con-

FIG. 46. Lake Garda in a southern gateway to the Alpine
highland expanded over the apron site of the earlier pied-
mont glacier (after Penck and Bruckner).

GLACIAL FEATURES DUE MAINLY TO DEPOSITION 85

ditions on the northern border of the Alps are brought
out in Fig. 47.

FIG. 47. Outline map of the northern border of the Alpine highland showing the
basins of former lakes. The trumpet-like widening of the valleys at their
mouths should be especially noted (based on a map by Bruckner).

Tongue-like basin lakes within the apron site of former
piedmont glaciers would appear to have been characteristic
also of the piedmont glaciation in the northern Rocky moun-
tains of Montana. 4

Stream Action on the Mountain Foreland. Wherever
glaciers are so large as to expand upon the foreland to the
range which furnishes their nourishment, they build up, as
has been seen, a broad rampart of morainic rock debris
which later marks the limit of their advance. The con-
stancy of occurrence and the magnitude of these deposits
from the ice near its margin, testify to the gradual change
from increasing rigor of climate to a progressive amelioration
of these conditions. No less significant in this respect are
the heavy deposits which outside the marginal moraine have
been distributed by streams of water from the glacier.

Most of this water emerges from beneath the ice, though
much of it may have flowed upon the glacier surface for
greater or less distances until permitted to descend through
crevasses to the bottom. Russell's study of the Malaspina
glacier of Alaska, the one existing example of a piedmont

86 CHARACTERISTICS OF EXISTING GLACIERS

glacier that has been carefully studied, showed that streams
of water from near the upper edge of the ice-apron there

SCALE : 1 in. = 1 mi.

FIG. 48. A braided stream which flows from the margin of the Vatnajokull in
Iceland. (From the new map by the Danish General Staff, 1905.)

GLACIAL FEATURES DUE MAINLY TO DEPOSITION 87

disappeared into tunnels within the ice to be lost to sight
until their reappearance at the outer margin.

The water of these streams is in the lower levels held
within the ice as within a pipe, and is in consequence under
strong hydrostatic pressure. Its flow is, therefore, much
more rapid than would be the case with a liquid having a
free surface. In fact, it could not otherwise ascend the slope
which we have found to be characteristic of the outer portion
of the tongue-like basin beneath the ice-apron.

The Outwash Apron. Emerging from beneath the ice, the
flow is suddenly checked and the stream being overloaded
with rock debris, this is quickly deposited as sediment, the
coarser materials nearer the ice margin and the finer ones
at greater distances. Thus is built up a broadly extended
outwardly sloping platform composed of water-deposited
materials, which platform surrounds the glacier and its
marginal moraine as an outwash plain or outwash apron.

Over the nearly level surface of the outwash apron, the
streams flow in ever shifting serpentine courses, and are
joined to their neighbors on either side only to divide the
waters of the combined streams at the first minor obstruction
that is encountered. Such composite streams may be
compared to the strands of a braid, and they have been de-
scribed as " braided streams" (see Fig. 48). The width of
such a stream, or perhaps better, series of streams, may
be as great or even greater than the individual length. Con-
stantly shifting their courses through lateral migrations,
such rivers grade the plain on which they flow to the even-
ness of a well-sanded floor. This peculiarity and their
location just without a marginal moraine (see Fig. 49)
make the later determination of such plains a relatively
easy matter.

Eskers and Recessional Moraines. To indicate their re-
lation to glaciers as well as to describe their deposition by

88 CHARACTERISTICS OF EXISTING GLACIERS

streams, out wash deposits are referred to as " fluvio-glacial."
They are sands and gravels imperfectly stratified and hav-
ing included lenticular masses of coarser and finer materials.

FIG. 49. Ideal form of a tongue-like basin remaining on the site of the ice apron
of a piedmont glacier and surrounded by the outwash apron. M, marginal
moraine at outer limit of ice advance ; D, drumlins ; C, basin usually occupied
by lake ; T, outwash plain of fluvio-glacial deposits (after Penck).

Russell has described streams which issue from the margin
of the Malaspina glacier with such velocity that the sudden
checking of their current causes the deposit of relatively
coarse materials in a steep apron resembling in form the dry
deltas of mountain fronts within semi-arid regions. He
points out that a continuance of such streams during a reces-
sion of the glacier would build up a serpentine ridge of water-
deposited materials whose average course would be per-
pendicular to the marginal moraine. 5 Such ridges, known as
"eskers," have not been reported from the sites of the Alpine
piedmont aprons, 6 though they appear to have formed
under somewhat similar conditions along the east front of
the Rocky mountains in Montana. 7

It should not be overlooked that while Russell's obser-
vations make it probable that eskers are now forming in
the tunnels beneath the ice apron and behind the alluvial
fans which block their outlets, the eskers do not appear
outside the ice front. Tarr, who has confirmed Russell's
conclusion, thinks that the active stream erosion at the ice
front would destroy the esker as fast as it was uncovered,
and that eskers have become visible only where the ice
ended in bodies of standing water or else had become rela-

GLACIAL FEATURES DUE MAINLY TO DEPOSITION 89

lively stagnant. 8 The former case would apply to the osar
of Sweden, and the frequent termination of eskers in delta-
like sand plains with relatively flat upper surface and
steeply sloping margins, would likewise favor this view.

With the commencement of the receding hemicycle of
glaciation, the ice front retires from its marginal moraine
and eventually enters the mountain valley, though usually
leaving behind it a series of smaller and so-called " reces-
sional moraines " to mark successive and relatively short
halting places during its retreat. The uncovered site of
the former ice-apron is, as we have seen, a basin, so that
this is filled with water from the melting of the ice during
the retreat to the mountain front. The water thus im-
pounded finds its outlet at the lowest level of the morainal
crest, and being already filtered of coarser material by
the lake itself, this outlet rapidly cuts a channel through
the loose materials of the moraine and its bordering plain
of fluvio-glacial deposits. Thus sections are exposed to
view revealing a history of the glacier whose episodes are to
be compared with those disclosed by the plan of the valley
above when it has likewise been laid bare.

Stream Action within the Valley during the Retirement of
the Glacier. The "glacier staircase" left by the ice, with
its rock-basin lakes high up in the valley and its morainal
lakes within the lower reaches, undergoes a rapid transform-
ation under the influence of running water so soon as the
ice has largely vacated the valley. Flowing from the waning
remnant of the glacier, this water is overburdened with
sediment. Its current is sluggish over the treads of the steps,
but develops cascades over the cliffs between. The coarse
debris which it carries is thus quickly dropped upon the
treads to fill the lake basins, and with the aid of the finer
material as tools, the rock obstructions are cut through in
narrow canyons and with a marvellous rapidity. Where a

90 CHARACTERISTICS OF EXISTING GLACIERS

barrier of more resistant rock has hemmed in a portion of
the valley (Riegel), narrow and picturesque gorges have
been cut, such as the Aarschlucht and the gorge of the

FIG. 50. Gorge of the Albula river, near Berkun, in the Engadine.

GLACIAL FEATURES DUE MAINLY TO DEPOSITION 91

Corner. Tyndall has described many of these interesting
gorges within the "European playground." 9 One of the
finest illustrations is furnished by the Albula River, near
Bergun, in the Engadine (see Fig. 50). Here the glaciated
valley with its characteristic U -section, is at the top of the
narrow gorge, which latter, therefore, represents the work
of the stream since the retirement of the glacier.

Landslides and Rock Streams within the Vacated Valley.
Perhaps the most general characteristic of regions which
have in recent times been sculptured by mountain glaciers
is the dominance of the precipitous rock face the walls of
the fretted upland. To-day, wherever rock climbing is
indulged in, there glaciers are to be seen, or the evidence
of their former presence is everywhere overwhelming. It
is the sapping process active in cirque recession and in
valley widening which has here developed the nearly ver-
tical rock face.

Obviously such steep surfaces are unstable under existing
conditions of weathering within humid regions, and can
long retain their forms only under the most favorable cir-
cumstances. Until the glacier vacated the valley, the walls
were in part supported at least toward the base by the ice
itself. On the Vernagt and Rhone glaciers a sliding down
of the walls has begun in the parts but lately left unsup-
ported. 10 If of weak or porous materials, or if intersected
by many planes of ready jointing or cleavage, such precipi-
tous faces become an easy prey to frostwork and rock
slide. For these reasons, glaciated valleys within mountain
districts have often been the scenes of disasters from ava-
lanche. Wherever the rock is of a porous nature or has
open structures, water gradually comes to fill all the spaces
within the material, at least along certain planes favorable
to its entry. After a time prodigious masses of rock sud-
denly descend under the influence of gravity, and within

92 CHARACTERISTICS OF EXISTING GLACIERS

the space of a few seconds, or at most minutes, they have
either partially or wholly blocked the valley, leaving great
scars to mark their former positions.

The landslide of Frank, Alberta, which occurred in 1903,
near where the southern line of the Canadian Pacific Rail-
road enters the Crows 7 Nest Pass of the Rocky Mountains,
was the movement of a mass of loose earth a half mile square
and between 400 and 500 feet in thickness. Only about
a minute and a half after this mass started from a shoulder
of Turtle Mountain, it had travelled two miles and a half
and been spread over a square mile of valley bottom. 11
Farther south in the Rocky Mountains, and in Colorado
particularly, are numerous relics of former great slides. 12
Here the insecure foundations of massive rocks and a
jointed and shattered condition of these rocks themselves
has facilitated the entrance of water within the rock
mass and greatly promoted avalanching.

FIG. 51. Ideal section showing successive slides from a canyon wall producing
a staircase effect with back-tilted treads (after Russell).

How important the vertical joint planes may be in the
settling away and eventual fall of the valley walls is shown
to advantage in the Otzthal of Switzerland, where, in the
angle between the Vernagthal and the Rosenthal a little
above the height of the glacier surface at its maximum,
the wall is now settling down in sections separated by joint
planes so as to produce the form of a staircase. 13 These
conditions are, moreover, common to all high vertical cliffs,
and Russell long ago pointed out that successive slides
take place from steep canyon walls in such a manner as to

GLACIAL FEATURES DUE MAINLY TO DEPOSITION 93

produce a staircase effect with back-tilted treads (see Fig.
51). 14

From the glacial valleys of Switzerland many examples
of great landslides have been supplied. In 1881 the town
of Elm in Canton Glarus was overtaken by a slide from the
Plattenbergkopf, which had been partly undermined in a
slate quarry. About twelve million cubic yards of rock fell
a distance of about 1500 feet, shot across the valley and up
the opposite slope to a height of 300 feet, and being there
deflected spread over a broad plain in a sheet which had an
area of a million square yards. Over the range from Elm

FIG. 52. View of the succession of rock slides from the north wall of the Upper
Rhine near the town of Flims.

and above the town of Chur in the valley of the Upper Rhine
is the site of a veritable succession of slides from the valley
wall, known far and wide as the Flimser Bergsturz (see Fig.
52).

Many of the apparent steps in the transverse sections of
Alpine valleys 15 are to be explained through landslides of
this nature.

94 CHARACTERISTICS OF EXISTING GLACIERS

Rock Flows from Abandoned Cirques. Long after the
waning horseshoe glaciers have disappeared from glacial
amphitheatres, the winter snows will there be collected and
persist through a portion, at least, of the summer season.
The same conditions of excessive frost weathering, which we
have become familiar with in the process of nivation and
of cirque recession within the same levels must, therefore,
long continue to exist. Essentially the same conditions
may be said to be characteristic of those other and vaster
inhospitable areas of the sub-polar regions which are un-
covered by ice and have but a thin covering only of snow.
For these districts the mechanical process of rock rending
and comminution is as characteristic as the chemical process
of decomposition within a warmer humid region.

After the rending of the rock materials has been
accomplished, gravity becomes effective to bring about a
transfer of material to the lower levels by a process of rock
flow which has been called "solifluction." 16 To this flow of
rock debris belong many of the properties either of water or
of ice-streams, and the moving masses have in different dis-
tricts been called " mud rivers/' " stone rivers/ 7 " rock
flows/ 7 " rock glaciers/ 7 etc. Together with this flow goes
also, under certain conditions, a peculiar striping of the sur-
face of the ground, 17 and as this occurs only below a drift
of snow, the function of the thaw water in giving the mass
its property of flow is at once apparent. It is well to empha-
size, then, that the thaw water from the melting snowdrift
determines both the nightly freezing and rending of the rock,
and the fluxion of the rock mass as well.

Outside the inhospitable sub-polar regions it is the aban-
doned glacier cirques which, largely because of their high
altitude and their peculiar form, best supply the conditions
requisite to solifluction. Within the San Juan Mountains
of Colorado, the high glacial cirques are many of them occu-

GLACIAL FEATURES DUE MAINLY TO DEPOSITION 95

pied by somewhat remarkable rock streams. 18 The position
of two of these rock streams relative to the neighboring
cirque walls is brought out in Fig. 53 and in plate 19, A

45*30"

|Q744'

Contour interval SOfeet

FIG. 53. Map of two high glacial cirques now partially occupied by rock

streams. The dotted areas are rock streams (after Howe).

and B. The rocks are here, by reason of their loose founda-
tion, of their open joints, and their steep forward grades,

96 CHARACTERISTICS OF EXISTING GLACIERS

most favorable to the entrance of water, and the subsequent
fall of the rock materials.

In the mountains of Alaska so-called " rock glaciers "
occur which have much in common with the rock streams
of Colorado. Rock glaciers are mixtures of ice and rock,
sometimes passing upward into glaciers of ice and having
in lower levels a surface coating only of angular rock
debris. 19

REFERENCES

1 Russell, "Malaspina Glacier," Jour. GeoL, vol. 1, 1893, pp. 228-238.

2 Penck und Bruckner, "Die Alpen im Eiszeitalter;'' also Calhoun,
I.e., Prof. Pap. U. S. GeoL Surv., No. 50, p. 20.

3 Bruckner, Z.c.,.1909, p. 793.

4 Calhoun, I.e., p. 16.

5 1. C. Russell, " Glaciers of North America," pp. 123-125.

6 Penck und Bruckner, "Die Alpen im Eiszeitalter."

7 Calhoun, I.e., p. 20.

8 R. S. Tarr, " Some phenomena of the glacier margins in the Yakutak
Bay Region, Alaska," Zeit. f. Gletscherk., vol. 3, 1909, pp. 96-97.

9 "Hours of Exercise in the Alps," pp. 224-230.

10 Ed. Bruckner, " Die glazialen Ziige im Antlitz der Alpen." Naturw.
Wochensch., N. F., vol. 8, 1909, p. 792.

" Howe, Prof. Pap. 67, U. S. GeoL Surv., 1909, p. 51. Also G. E.
Mitchell, Nat. Geogr. Mag., vol. 21, 1910, pp. 285-287.

12 Ernest Howe, " Landslides in the San Juan Mountains," Prof. Pap., 67,
U. S. GeoL Surv., 1909, pp. 1-58.

13 Bruckner, " Die glazialen Ziige im Antlitz der Alpen," I.e., p. 792.
See also Salisbury, "Physiography," N. Y., 1907, Fig. 98, p. 108.

14 1. C. Russell, "Topographic features due to landslides," Pop. Sci.
Month., vol. 53, 1898, pp. 480-489.

15 See E. J. Garwood, Geogr. Jour., vol. 36, 1910, p. 320.

16 J. G. Andersson, " Solifluction, a component of sub-aerial denuda-
tion," Jour. GeoL, vol. 14, 1906, pp. 91-112.

17 O. Nordenskiold, " Die Polarwelt und ihre Nachbarlander," 1909, pp.
60-65. Wm. H. Hobbs, "Soil Stripes in cold humid regions and a kin-
dred phenomenon," 12th Rept. Mich. Acad. Sci., 1910, pp. 51-53.

18 Howe and Cross, "Glacial phenomena of the San Juan mountains,
Colorado," Bull. GeoL Soc. Am., vol. 17, 1906, pp. 251-274. See also
Howe, I.e., pp. 31-55.

19 Stephen R. Capps, Jr., "Rock Glaciers in Alaska," Jour. GeoL, vol.
18, 1910, pp. 359-375, figs. 1-10.

PLATE 19.

A. Rock stream in a cirque on Greenhalgh Mountain, Silverton quadrangle, Colo-
rado (after Howe, U. S. Geol. Survey).

B. Rock stream at the head of a cirque in the Silver Basin, Silverton quadrangle,
Colorado (after Howe, U. S. Geol. Survey).

PART II

ARCTIC GLACIERS

CHAPTER VII
THE ARCTIC GLACIER TYPE

Introduction. As elsewhere pointed out, continental
glaciers are in other than dimensional respects sharply
differentiated from those types which have been described
as mountain glaciers. 1 The ice-cap glacier, while of smaller
dimensions than the true inland-ice or the continental glacier,
is physiographically allied with this type, and has few affini-
ties with mountain glaciers. The sharpness of the distinc-
tion has often been overlooked for the reason that true moun-
tain glaciers frequently exist within a fringe surrounding the
larger areas of inland-ice, both in the Arctic and Antarctic
regions. The distinguishing difference between mountain
glaciers and continental glaciers is one primarily dependent
upon the proportion of the land surface which is left un-
covered by the ice, and the position of this surface relative
to the margins of the snow-ice mass. With true mountain
glaciers land remains uncovered above the highest surfaces
of the glacier, where, in consequence, a special erosional pro-
cess cirque recession becomes operative. The smaller
ice-caps take their characteristic carapace form and cover
the surface of the land within their margins, because that
surface is relatively level. Had it been otherwise, the same
conditions of precipitation would have yielded mountain
H 97

98 CHARACTERISTICS OF EXISTING GLACIERS

glaciers in their place. The law above stated is none the
less applicable, since, because of this flat basement, no land
projects above their higher levels. 2

There are, as we shall see, other attributes which strikingly
differentiate the large continental glaciers from all other
bodies of land ice. These relate particularly to the nature
of the snow which feeds them, to changes which that snow
undergoes after its fall, to the manner of its transportation,
etc. Most of these differences are of such recent discovery,
or at least of such recent introduction into the channels of
dissemination of science, that they have not yet found their
way to the student of glacial geology. We shall profitably
begin our description of continental glaciers with the inter-
mediate ice-cap type, so as to establish connection with
mountain glaciers in the important condition of alimenta-
tion. Before doing so, it will be well to call attention to
some contrasts which exist between the north and south
polar regions in those conditions upon which glaciation
depends.

North and South Polar Areas Contrasted. A glance at a
globe, which sets forth the land and water areas of the earth,
is sufficient to show that the disposition of land and water
about the opposite ends of the earth's axis is essentially
reciprocal. About the north pole we find a polar sea, the
Arctic ocean, surrounded by an irregular chain of land
masses which is broken at two points, nearly diametrically
opposite. In the Antarctic region, on the contrary, it is a
high continent which is massed near the pole, and this is
bounded on all sides by a sea in which are joined all the
great oceans of the globe save only the Arctic. This polar
continent is deeply indented on two nearly opposite margins,
but to what extent is not yet known. The margins of the
continent are extended beneath the sea in a wide conti-
nental shelf or platform. The broad encircling ocean,

THE ARCTIC GLACIER TYPE 99

while to some extent invaded by the southern land tongues
of South America, Africa, Australia, and New Zealand, is
yet so little occupied by land masses that wind and ocean
currents are alike but slightly affected by them. The
Antarctic conditions are, therefore, oceanic character-
ized by uniformity and symmetry to a much larger extent
than is true of the northern polar region.

Within the northern hemisphere a large quantity of heat
from the tropics finds its way northward to the breaks in the
northern land chain, through the medium of great ocean
currents the Gulf Stream in the Atlantic, and the Japan-
ese Current in the Pacific. Cold return currents from the
Arctic region, and the widely different specific heats of
land and water, cooperating with the effect of the northward-
flowing warmer currents, result in a marked diversity in
temperature, winds, and precipitation at different longitudes
within the same latitudes. Lack of symmetry in distribu-
tion and wide variations in climatic conditions are, therefore,
characteristic of the north polar region; and it follows that
the present glaciation of the northern hemisphere is localized
within a few scattered areas where the land projects far-
thest toward the pole, and near where there are sea areas of
excessive evaporation to supply the necessary moisture.

The Fixed Areas of Atmospheric Depression. Examina-
tion of Fig. 54 will show that the areas of existing heavy
glaciation in the northern hemisphere lie close to the so-
called fixed areas of low barometric pressure, each of which
is a long, curved trough, concave to the northward, one
central over the Aleutian Islands' Arc at the northern
bight of the Pacific ocean, the other extending from the
southeastern extremity of Baffin Land past Cape Farewell,
and northeastward across Iceland, so as to occupy similarly
the northern bay of the Atlantic ocean. For such northern
latitudes, these areas of fixed low barometric pressure are

100 CHARACTERISTICS OF EXISTING GLACIERS

in consequence characterized by abnormally large evapora-
tion. Wherever the moisture-laden winds proceeding from
these areas are forced to rise by upland barriers, or to come

FIG. 54. Map showing the areas of fixed low barometric pressure in the northern
hemisphere (after Buchan). The areas of heavy glaciation have been added.

in contact with cold rock or snow surfaces, condensation and
precipation must inevitably take place.

The prevailing westerly winds from the Pacific, when they
encounter the high backbone of the Cordilleran System of
mountains in Alaska nourish the great mountain glaciers of
that region. The Cordilleras of Alaska are, however, compe-
tent to arrest but a small portion of these moisture-laden
clouds, for it is only in moderate latitudes that they bar the
way, and no highlands lie beyond them to the eastward until
the vicinity of Baffin Bay has been reached.

On the border of the Atlantic low pressure area are found
Greenland, Iceland, Spitzbergen, Norway, Franz Josef Land
and Nova Zembla, each with its upland areas and its exist-

THE ARCTIC GLACIEE TYPE 101

ing glaciation. In Norway, Iceland, and Franz Josef Land
we find ice-caps; in Spitzbergen, Nova Zembla, Baffin Land,
Grinnell Land and Ellsmere Land, the mantle of snow and ice
is best described by the name " inland-ice," while upon the
continent of Greenland the inland-ice has continental dimen-
sions, and forms one of the two continental glaciers that still
exist. 3

Of all the northern ice-sheets, with the exception of the archi-
pelago of Franz Josef Land, the rule holds that they are smaller
than the land masses upon which they rest, and this in part ex-
presses the difference between the northern and southern
types of inland-ice.

Ice-caps of Norway. In contrast with all save the pied-
mont type of mountain glaciers, the snow-fields of ice-caps

FIG. 55. Idealized section showing the form of "fjeld" and "brae" in Nor-
wegian ice-cap.

are much the larger. Speaking broadly, high and relatively
level plateaus, light winds, and low temperatures are favor-
able to the existence of ice-caps. To-day they are not to be
found in latitudes lower than 60. In Norway, within the
zone of heavy precipitation along the western coast, and
upon the remnants of the plateau separated by the fjords
are still to be found a number of small ice-caps. These
caps consist of a central carapace of snow and ice from the
borders of which narrow tongues descend into the fjords.
The largest of these ice-caps is the Jostedalsbraen, having
an area of 1076 square kilometers. Whereas with moun-
tain glaciers the neve is contained within a basin, the cirque,
we here find the so-called " fjeld " nearly level and resting

102' ^CHARACTERISTICS OF EXISTING GLACIERS

upon the surface of the plateau. Of this fjeld broadly
lobate extensions lie upon its margin separated by deep
valleys or fjord heads. Much narrower extensions of the
central carapace often descend the steep slopes at the upper
end of these valleys and may continue down the valley floor.
Their narrowness is largely explained by their more rapid
motion upon the steeper slope and by the radiated heat from
the rock walls on either side (see Fig. 55 and plate 20) . 4
Near the margins of the ice carapace, the subjacent
terrane sometimes makes its appearance as rocky islets or
nunataks, as, for example, in the Hardangarjokull near
Finse in Southern Norway (see plate 17 A).

Ice-caps of Iceland. In Iceland are to be seen some of
the finest examples of ice-caps that are known, and, fortu-

FIG. 56. Maps of the Hofs Jokull and the Lang Jokull (after Thoroddsen).

nately, these have been carefully studied by Thoroddsen. 5
These ice-caps form gently domed crests or undulating ice-
fields situated upon the highest plateaus which rise above
the general table-land of the country. Projecting mountain
peaks appear with few exceptions only near the thinnest

PLATE 20.

2000*1000 02 * e

* a

Portion of the new map of the Jostedalsbriien, which displays the characteristic plan of
the surface physiography. The glacier sends out lobes upon the flat parts of the
spurs between fjords, and elsewhere descends in long, narrow tongues into the
fjords themselves, here dimpled above the fjord heads through indraught of the

THE ARCTIC GLACIER TYPE

103

margins of the ice, where they form either comb-ridges or
sharp peaks (see Figs. 56 and 57). White and altogether
free from surface rock debris except in the vicinity of their
margins, these ice-caps offer in this respect additional con-
trast to mountain glaciers. The largest of the Iceland ice-
caps is the Vatna Jokull, which has an area of 8500 square

FIG. 57. Map of the Vatna Jokull (after Thoroddsen).

kilometers, while the surfaces of the Hofs Jokull, Lang Jokull,
and Myrsdals Jokull, each exceed a thousand square kilo-
meters. The shield-like boss of the Vatna Jokull is brought
out in the section of Fig. 58. 6

Those borders of this ice mass which lie upon the plateau,
the northern and western areas, are broadly lobate; but
upon the southern and eastern margins, where the ice mass
descends to lower levels and approaches the sea, its tongues
sometimes end a few metres only above sea-level. It is
noteworthy, however, that where deeply incised valleys in-
vade the plateau upon this margin, the lobes of ice push out
mainly upon the upland remnants between the valleys,

104 CHARACTERISTICS OF EXISTING GLACIERS

though they send narrow tongues down the valleys them-
selves. This, as we shall see, is a peculiarity which ice-caps
and the northern inland-ice as well, have in common to
distinguish them further from mountain glaciers. As was
found true of the Norwegian glaciers, so here the tongue
which follows the valley bottom and which partakes of many
of the properties of a mountain glacier, is much the narrower. 7
From comparison with the Antarctic ice masses, these
rapidly moving extensions of the central mass through rock
gateways maybe designated " outlets " (see Part III, p. 186).
This peculiarity of ice-caps is well displayed upon the General

-2000 i|-r j ca.i900m . _ , ^ooo -

-1500-1 ^^^^~ ~~ ~--^^^^^ i 1500 -

-1000 ris^"^ ^^""^-^ 1-1000-

-500 X^- j 300 -

Cm ^s^ m

FIG. 58. Cross section of the Vatna Jokull from north to south (after Thoroddsen

and Spethmann).

Staff Sectional Map of Iceland, on scale 1 : 50,000 now in
process of publication. The sections which include the
Icelandic ice-caps show, not only the contours of the ice
surface, but, further, the nature of the crevassing, and they
are probably the finest glacier maps which have thus far
been issued. Plate 21 reproduces on a reduced scale, a
portion of section Oraefajokull.

From the north or plateau margin of the Vatnajokull, flow
mighty but sluggish streams which, near the glacier, are
braided into constantly shifting channels within a broad
zone of quicksand. In this sand, horse and rider, if once
entangled, are quickly lost. Upon the south margin, on the
other hand, the streams from the melting of the ice flow as
series of fast rushing rivers, sometimes so broad as not to be
bridged, and in these cases setting up impassable barriers
between districts (see Fig. 48, p. 86).

Icelandic ice-caps differ from all well-known glaciers at

PLATE 21.

Map of the margin of an Icelandic ice-cap. The tongue-like streams of ice in val-
leys and the apron-like extensions on the plateau level are shown (from the Gen-
eral Staff map, section Oraefajokull, 1905). The north side is here at the bottom
of the map.

THE ARCTIC GLACIER TYPE 105

least in this, that nowhere else are large ice masses in such
direct association with so active volcanoes. The jokulhlaup,
which is the Icelandic name applied to one of the charac-
teristic catastrophies of the island, occurs whenever a volcano,
either visible in the neighborhood of the glacier or hidden
beneath it, breaks suddenly into eruption. The first inti-
mation that such an event is transpiring, is often the drying
up of a stream which flows from the affected region. Some-
times the people are permitted to see great masses of lava
and volcanic ash issue together from the glacier. All at
once, the stream which had first dried up comes rushing down
its valley as a foaming flood of water, spreading out for
miles and having a depth sometimes as great as 100 feet.
The entire plain is then spread with mud and sown with
great rocks and also with ice blocks, some of which are as
large as the native houses. These ice blocks are often buried
in the mud, and later, when they have melted, they leave
deep pits in the plain similar to, though smaller than, the
depressions in a " pitted plain " from the continental glaciers
of Pleistocene time. The " glacier run " of 1903 produced
pits (Solle) in the Skeithardr Sander. In 1904 the par-
tially melted blocks of ice were to be seen in the pits. 8 It
is not, of course, here assumed that the cause of the pits
of the Pleistocene sand plains are in any way connected
with volcanic action, but only with the burial of ice blocks
under rock debris. Tarr has described the formation of
such plains in front of the Hidden glacier of Alaska, where
the melting ice margin is becoming buried beneath its own
burden of rock debris and is locally opened up in pit lakes. 9
During a volcanic eruption, water is seen to shoot up from
the glacier in great jets, and it has sometimes happened
that the entire ice mass of the jokull has been shattered, and a
chaotic mass of ice miles in width has slipped resistlessly down
the slopes. With the conclusion of the disturbance, the as-

106 CHARACTERISTICS OF EXISTING GLACIERS

pect of the entire district is sometimes found to be utterly
changed. All vegetation has been destroyed, and ridges
which had lent to the landscape its character have vanished,
so that streams have lost their old channels and entered
upon wholly different courses. 10

Ice-covered Archipelago of Franz Josef Land. The is-
lands of Franz Josef Land in the high latitude of 80 and over,
with altitudes of 2000 to 4000 feet, and situated as they are
on the borders of an open sea, are the most Arctic in their
aspect of all the smaller northern land masses. As a conse-
quence, they are with unimportant exceptions completely
snow-capped, the snow-ice covering sloping regularly into the
sea upon all sides. The Jackson-Harmsworth 11 and Ziegler 12
expeditions, following those of Nordenskiold, Nansen, the
Duke of the Abruzzi, and others, have now supplied us with
fairly accurate maps of all islands in the archipelago. One
or two of the western islands alone show a narrow strip of low
shore land, but with these exceptions all are ice covered save
for small projecting peaks or plateau edges near the margins
(see Fig. 59). They present, therefore, a unique exception to
the law which otherwise obtains, that within the northern
hemisphere glacial caps are smaller than the land areas upon
which they rest. The appearance of the island covers is
here, however, that of neve of low density, rather than of
compact glacier ice.

Prince Rudolph Island, which was the winter station of the
Italian Polar Expedition, is no doubt typical of most islands
in the archipelago. This land is described by Due d'
Abruzzi 13 as "completely buried under one immense glacier,
which descends to the sea in every direction except at a
few points, such as Cape Germania, Cape Saulen, Cape
Fligely, Cape Brorok, Cape Habermann, and Cape Auk.
At some of these points . . . the coast is almost perpendic-
ular, which prevents the ice from descending to the sea.

THE ARCTIC GLACIER TYPE

107

FIG. 59. Map of the ice-capped islands in the eastern part of the Franz Josef
Archipelago (after Fiala).

108 CHARACTERISTICS OF EXISTING GLACIERS

At others . . . the ice, stopped by a hollow, falls into the
sea on each side of the headland, which thus remains un-
covered. Moreover, wherever the snow can rest, there are
glaciers which end at the*sea in an ice cliff, like that formed
by the main glacier, so that it can be said that the entire coast,
with the exception of a short extent of strand near Teplitz
bay, is formed by a vertical ice cliff " (see Fig. 60).

The movement of the ice is so slow that though a line of
posts was established for the purpose of measuring during a

FIG. 60. Typical ice cliff of the coast of Prince Rudolph Island, Franz Josef
Land (after the Duke of the Abruzzi).

period of four months, no movement could be detected.
Except near the outermost margin, there were few crevasses,
and these were covered by snow. In summer, on days when
the temperature was above the freezing point, the snow
thawed rapidly so that torrents of water flowed from the
glacier to the sea, hollowing out channels, many feet in
width.

During the stay of the " Polar Star " near the island, it
was noteworthy that thaw and evaporation upon the island
exceeded the precipitation. Doubtless because of the slow
movement of the ice, no icebergs were seen to form during
the entire stay.

THE ARCTIC GLACIER TYPE 109

The Smaller Areas of Inland-ice within the Arctic Regions.
The ice-cap of the Vatna Jokull in Iceland, which is the

FIG. 61. Map of Nova Zembla, showing the supposed area covered by inland-ice
(from Andree's " Handatlas ").

largest to which this name has been applied, covers an area
of 8500 square kilometers. Ice carapaces, which are better
described as inland-ice, since they cover considerable propor-

110 CHARACTERISTICS OF EXISTING GLACIERS

tions of the interiors of the land areas upon which they rest,
occur to the northward of the continent of Europe in Nova
Zembla and Spitzbergen, and in the lands to the west of
Baffin's Bay, known as Baffin, Ellesmere, and Grinnell
lands.

FIG. 62. Map of Spitzbergen, showing the supposed glacier areas (from Andree's

" Handatlas").

Nova Zembla is a long, narrow island, stretching between
70 and 84 of north latitude (see Fig. 61). It is, in reality,
two islands separated by a narrow strait near the latitude of
76. The northern island, which to the north is a plateau
attaining an altitude of 1800 feet, is supposed to be in large

THE ARCTIC GLACIER TYPE

111

part covered by inland-ice, though it has been as yet but little
explored. 14

The Inland-ice of Spitzbergen. The group of islands to
which the name Spitzbergen has been applied lies between
the parallels of 76 and 81 of north latitude. The surface
is generally mountainous, the highest peaks rising to an ele-
vation of about 5000 feet, though the greater number range
from 2000 to 4000 feet in altitude. The large northeastern
land mass is called North East Land and is covered with in-
land-ice which was crossed by Nordenskjold and Palander in

FIG. 63. Inland-ice of New Friesland as viewed from Hinloopen Strait (after

Conway).

1873 15 (see Fig. 62). New Friesland, or the northeastern por-
tion of the main island, is also covered by inland-ice 16 (see
Fig. 63). The southwestern margin of this inland-ice was
somewhat carefully mapped by Conway and Gregory in
1896, 17 and as this presents some interesting general features,
the map is reproduced in part in Fig. 64.

In addition to the lobes which push out upon the crest of
the plateau, there is here an expansion laterally beyond the
main cap and at lower levels in the form of an apron which is
called the Ivory Gate (compare the Frederikshaab Glacier in
Fig. 94, p. 171). Surrounding the inland-ice to the westward
are small ice-caps resembling the fjelds and braes of Norway,
and also true mountain glaciers whose cirques have shaped
the mountains into the sharp pinnacles of comb ridges. It
is to these sharp peaks that Spitzbergen owes its name.

112 CHARACTERISTICS OF EXISTING GLACIERS

In the year 1890 Gustav Nordenskjold made a journey be-
tween Horn Sound and Bell Sound on the west coast, and
found behind the sharp peaks bordering the coast an ice sur-
face almost without crevasses or nunataks. 18 Upon the north-

FIG. 64. Map of the southwestern margin of an extension of the inland-ice of
New Friesland (after Con way).

west coast no sharp peaks or comb ridges are found, but there
is a low plateau with deep, narrow valleys similar to the west
coast of Norway, where it reaches the sea near the North Cape.
All the rock surfaces are glaciated.

The inland-ice of North East Land reaches the sea upon the
southern and eastern coasts, but is surrounded by a hem of
land upon the north and west. Over the surface of this ice

THE ARCTIC GLACIER TYPE 113

Nordenskjold journeyed in the spring of 1873, finding it to
be probably from 2000 to 3000 feet in thickness. Where it
reaches the sea on the east coast is a steep and inaccessible
cliff of ice, one of the largest in the northern hemisphere. 19
On the northern margin, however, the ice moves out upon a
plain with its own upper surface of gentle slope, which except
for the crevasses, is not difficult of ascent. From near this
northern border good seeing conditions enabled Nordenskj old
to say that the ice mass stretched away to the south and west
without any interruption from nunataks, but rising with
great uniformity into the great flat dome of its central area.
Over this snow surface every puff of wind drove before
it a stream of fine snow dust, which insinuated itself into
everything and was as troublesome as the sand of a desert. 20

The upper layer of the glacier was not of ice, but consisted
of hard, white, compacted snow which had been smoothed
and polished by the abrasion of the wind-driven snow dust.
In a depth of four to six feet the surface layer of compact
snow passed over into ice, first through a layer of magnificent
ice crystals, next to a distinctly granular ice, and finally
into a hard, coherent ice mass in which only the numerous
cavities filled with compressed air gave evidence of the
manner of its formation. When the ice-wall about these
cavities is by melting made too weak to sustain the pressure
of the air compressed within them, it breaks up with a pecul-
iar crackling sound which in summer is continually to be
heard from the pieces of granular ice floating about in the
fjords.

" We wandered," says Nordenskj old, " over a kind of
'neve region, that is to say, over a part of the glacier where the
surface is occupied by a layer of snow which does not melt
away during summer, whereas in Greenland at the beginning
of the month of July the snow upon the surface of the glacier
was, on the contrary, already nearly completely melted. No

114 CHARACTERISTICS OF EXISTING GLACIERS

trace of the glacier lakes, the beautiful and abundant glacier
streams, the fine waterfalls and fountains, etc., which occur
everywhere on the Greenland inland-ice can be observed here,
and the configuration of the surface showed that such forms
never occur, or only to a very limited extent. The melting
of the snow clearly goes on in Spitzbergen on too inconsider-
able a scale for such phenomena to arise."

" The surface of the snow was, as has been already men-
tioned, quite level, generally hard packed by the storms.

FIG. 65. Camping place in one of the " canals " upon the surface of the inland-ice
of North East Land, Spitzbergen (after Nordenskjold).

and completely glazed and polished by the stream of snow
which even the gentlest breeze of wind carried forward along
the ground. This stream of snow, or more correctly of air
mixed with snow, had, however, in the absence of a downfall,
and provided the wind was not all too violent, only a depth of
a few feet. It threw fragile bridges of snow over the cre-
vasses, but did not fill them; formed where there were great
precipices, true snow cascades; and filled up in a few minutes

THE ARCTIC GLACIER TYPE

115

FIG. 66. Hypothetical cross section of a glacial
canal upon the inland-ice of North East Land
(after Nordenskjold).

all shallow holes and depressions. Thus, for instance, when

we emerged from our tent in the morning, all trace that the

snow had been tram-

pled down the evening

before had generally

disappeared, and the

sledges were concealed

in a large drift." 21
Of especial note were

the great crevasses

which ran generally in

straight lines for long

distances in parallel

series, sometimes two

intersecting series being observed. More remarkable than

these, however, were the so-called " canals/ ' which also for

the most part ran parallel to each other, and in some cases

were only 100 feet apart. These
canals, which were found in the
southeastern part of the area near
Cape Mohn, were in reality deep,
flat-bottomed troughs within the
ice, bounded on either side by
parallel and rectilinear ice cliffs,
and were in places partially filled
by the indrifted snow. Stretching
for long distances over the snow
plain, and set so deeply that they

COuld

FIG. eT.-Mapshowingthesup-

posed area of inland-ice upon itOUS drifting of the SnOW Supplied

St= an incline > th ^ were utilized for

camping places (see Fig. 65).

Nordenskjold has explained these canals as trough faults
within the ice, and has assumed that this deformation was

116 CHARACTERISTICS OF EXISTING GLACIERS

due to changes of volume incidental to extreme temperature
range (see Fig. 66). This explanation in temperature changes
would leave the absence of such structures in other places-
wholly unaccounted for, and we venture to believe that a
recent trough faulting within the rock basement below the
ice, and communicated upward through it, would furnish a
more reasonable explanation, particularly in view of our later
knowledge of dislocations connected with earthquake dis-
turbances.

Still deeper inbreaks of the ice were encountered within
the same region. These, though deeper, were generally of

FIG. 68. View of the " Chinese Wall " surrounding the Agassiz Mer de Glace on
Grinnell Land (after Greely).

less extent, and were designated by the sailors of the party
" docks " or " glacier docks."

The Inland-ice of Grinnell, Ellesmere, and Baffin Lands.
Something has been learned of the inland-ice of Grinnell
Land (see Fig. 67) from the report of Lieutenant Lockwood
upon his crossing of Grinnell Land in 1883. 22 Of especial inter-
est is his description of the ice front or face as it was observed
for long distances in the form of a perpendicular wall which
he described under the name " Chinese Wall." Over upland
and plain this wall extended with little apparent change in
its character. At one place by pacing and sextant angle its
height was estimated at 143 feet (see Fig. 68).

The inland-ice of Ellesmere Land (see Fig. 67) has been to

THE ARCTIC GLACIER TYPE

117

some extent explored along its borders by members of the
Sverdrup Expedition. 23 The maps of the margin in the vicin-
ity of Buchanan Bay display much the same characters as

FIG. 69. Map showing the supposed area of inland-ice upon Baffin Land (from
Andree's " Handatlas ").

may be observed along the margins of the better-known ice-
caps and inland-ice masses of the northern hemisphere.

Of the inland-ice of Baffin Land little is known (see Fig. 69).
There are some indications that a small ice-cap exists upon
the neighboring island of North Devon.

REFERENCES

J Wm. Herbert Hobbs, "The Cycle of Mountain Glaciation," Geogr.
Jour., vol. 36, 1910, pp. 147, 148.

2 W. M. Conway, "An Exploration in 1897 of some of the Glaciers of
Spitzbergen," Geogr. Jour., vol. 12, 1898, pp. 142-147.

118 CHARACTERISTICS OF EXISTING GLACIERS

8 It has not in most cases yet been determined to what extent the
present nourishment of these glaciers suffices to maintain them, or, per
contra, to what extent they are mere waning remnants of larger pre-existing
masses. It is, however, known that formerly they were much larger.

*H. Hess, "Die Gletscher," 1904, pp. 66, 90-92.

5 Th. Thoroddsen, "Island, Grundriss der Geographic und Geologie,"
Pet. Mitt. (Erganzungshefts 152, 153), 1906, V., " Die Gletscher Islands,"
pp. 163-208.

6 Hans Spethmann, "Der Nordrand des islandischen Inlandeises Vatna-
jokull," Zeitsch. f. Gletscherk., vol. 3, 1909, pp. 36-43.

7 Carl Sapper, "Bemerkungen iiber einige siidislandische Gletscher,"
Zeitsch. f. Gletsch., vol. 3, 1909, pp. 297-305, two maps and three figures.
See especially Fig. 3.

8 Max Ebeling, " Eine Reise durch das islandische Siidland," Zeit.
Gesellsch. f. Erdkunde, Berlin, 1910, pp. 361-382.

9 R. S. Tarr, " Some phenomena of the glacier margins in the Yakutat
Bay Region, Alaska," Zeit. f. Gletscherk., vol. 3, 1909, pp. 94-96, Fig. 6.

10 Otto Nordenskjold, "Die Polarwelt," 1909, pp. 42-43.

11 F. G. Jackson, "A Thousand Days in the Arctic," 1899, map 5.

12 Anthony Fiala, "The Ziegler Polar-Expedition of 1803-05," 1907,
map C.

13 " On the ' Polar Star ' in the Arctic Sea," vol. 1, pp. 116-118.

14 Professor Hanns Hofer, " Graf Welczeks Nordpolar-fahrt im Jahre
1872, III Ueber die Gletscher von Nova Zembla," Pet. Mitt., vol. 21, 1875,
pp. 53-56. See also Commandant Charles Benard, " Dans 1'ocean gla-
cial et en Nouvelle-Zemble," Paris, 1910, pp. 1-193.

16 A. E. Nordenskjold, "Gronland," map on p. 141.

16 W. Martin Conway, "An Exploration in 1897 of some of the Glaciers
of Spitzbergen," Geogr. Jour., vol. 12, 1898, pp. 137-158.

17 Sir Wm. Martin Conway, "The First Crossing of Spitzbergen,"
London, 1897, pp. 371, 2 maps.

18 O. Nordenskjold, I.e., p. 52.

19 See O. Nordenskjold, Die Polarwelt, p. 52.

20 A. E. Nordenskjold, "Die Schlittenfahrt der schwedischen Expedi-
tion im nordostlichen Theile von Spitzbergen, 24 April-15 Juni 1873," Pet.
Mitt., vol. 19, 1873, pp. 450-453.

21 Nordenskjold, I.e., pp. 255-257.

22 A. W. Greely, "Report on the Proceedings of the United States Ex-
pedition to Lady Franklin Bay, Grinnell Land," vol. 1, especially Ap-
pendix No. 86, pp. 274-279, pis. 1-4. See also Salisbury, Jour. Geol.,
vol. 3, p. 890.

23 Otto Sverdrup, "New Land," 2 vols., London, 1904, pp. 496-504.

CHAPTER VIII

PHYSIOGRAPHY OF THE CONTINENTAL GLACIER OF
GREENLAND

General Form and Outlines. The Inland-ice of Greenland,
we have now good reason to believe, has the form of a flat
dome, the highest surfaces of which lie somewhat to the
eastward of the medial line of the continent. 1 This ice dome
envelops all but a relatively narrow marginal rim. The mar-
ginal ribbon of land is usually from five to twenty-five miles
in width, may decrease to nothing, but in two nearly opposite
stretches of shore it widens to from sixty to one hundred
miles (see Fig. 70).

At the heads of many deep fjords long and narrow mar-
ginal tongues pushing out from the central mass reach to be-
low sea level ; and within three limited stretches of shore the
ice mantle overlaps the borders of the continent and reaches
the sea in a broad front. The longest of these begins near
the Devil's Thumb on the west coast at about latitude 74' 30",
and extends with some interruptions for about one hundred
and fifty miles along the coast of Melville Bay. 2 Where
crossed by Nansen near the parallel of 64, and hence near
the southern margin, and also where traversed by Peary
near its northwestern borders, the inland-ice has revealed
much the same features. The great central area has never
been entered, although Baron Nordenskjold and Commander

119

120 CHARACTERISTICS OF EXISTING GLACIERS

Peary have each passed somewhat within the margin near
the latitude of 68, and Jensen near latitude 63 3 (see
Fig. 70).
In 1893 Garde at the extreme southern end of the con-

FIG. 70. Map of Greenland, showing the outlines of the inland-ice (from
Andree's "Handatlas," but corrected for the northeast shore from data of the
Danish expedition of 1908). The routes of the various expeditions on the
inland-ice have also been added.

THE CONTINENTAL GLACIER OF GREENLAND 121

46 30

tinent (latitude 61-62) penetrated the area of the inland-
ice a distance of about sixty-five miles. 4 The route of
the expedition and the
contours of the surface
are given in Fig. 71. 5
The first partially suc-
cessful attack upon
the inland-ice was that
of Dr. I. I. Hayes,
Commander of the
United States Explor-
ing Expedition, which
spent the winter of
1860-1861 on Smith
Sound in northwestern
Greenland. Hayes ^
succeeded in
reaching a point
seventy miles
within the margin
of the ice at an

plpvatirm nf ahrmt FIG. 71. Route of Garde across the margin of the inland-

elevation 01 about ice o gouth Greenland in 1893 (after Garde)
5000 feet. 6

Recently (1907) Mylius Ericksen met his tragic death in
crossing the inland-ice in northeast Greenland, but his re-
sults, most fortunately recovered, through the heroism of
Bronlund, are not yet published. Yet such is the monotony
of the surface thus far revealed, and such the uniformity of
conditions encountered, that there is little reason to think
future explorations in the interior will disclose anything but
a desert of snow, with such small variations from a horizon-
tal surface as are not strikingly apparent to the traveller at
any one observing point.

Nansen has laid stress upon the close adherence of the

122 CHARACTERISTICS OF EXISTING GLACIERS

curve of his section to that of a circle, and has attempted to
apply this interpretation to the sections of both Norden-
skjold and Peary made near the latitude of Disco Bay. 7 If
the marginal portions of the sections be disregarded, this in-
terpretation is possible for Nansen's own profile, since it is

FIG. 72. Sketch of the east coast of Greenland near Cape Dan. Shows the
inland-ice and the work of marginal mountain glaciers (after Nansen).

in any case very flat; but inasmuch as the margins only
were traversed in the other sections, the conclusions drawn
from them are likely to be misleading when extended into
the unknown interior.

Hess, 8 correcting Nansen's data so as to take account of the
curvature of the earth, finds the radius of this circle of the
section to be approximately 3700 km. (instead of 10,380 km.,
as given by Nansen). This radial distance being consider-
ably less than the average for the earth's surface, the curva-
ture of the ice surface where crossed by Nansen is consider-
ably more convex than an average continental section.

FIG. 73. The section across the inland-ice of Greenland, near the 64th parallel of
latitude in natural proportions and with vertical scale ten times the horizontal
(after Nansen).

We are absolutely without knowledge concerning either the
thickness of the ice shield or the elevation of the rock base-
ment beneath it, though a height of the snow surface of ap-
proximately 9000 feet was reached by Nansen at a point
where it could hardly be expected to be a maximum. The
snow surface to the north of his section was everywhere ris-

THE CONTINENTAL GLACIER OF GREENLAND 123

ing, and it is likely that it attains an altitude to the north-
eastward well above 10,000 feet.

Though doubtless almost flat within its central portions,
and only gently sloping outward at distances of from sev-
enty-five to one hundred miles within its margin, the snow
surface falls away so abruptly where it approaches its bor-
ders as to be often quite difficult of ascent (see Fig. 73). 9 The
monotony of the flatly arched central portion of the isblink 10

FIG. 74. Comparison of the several profiles across the margin of the inland-ice:
(a) at latitude 69| on the west coast (Peary) ; (6) at latitude 68 on the west
coast (Nordenskjold) ; (c) at latitude 64 on the west coast (Nansen) ; and (d) at
latitude 64| on the east coast (Nansen).

gives place to wholly different characters as the margins are
approached. The ice descends in broad terraces or steps,
which have treads of gentle inclination but whose risers are
of greater steepness, and this steepness is rapidly accelerated
as the margin is neared. In Fig. 74 have been placed to-
gether for comparison the profiles of Peary, Nordenskjold,

124 CHARACTERISTICS OF EXISTING GLACIERS

and Nansen tin the different routes which they travelled
toward the interior from the coast.

The margins of the Greenland continent where uncovered
by the ice, are generally mountainous, with heights reaching
in many cases to between 5000 and 8000 feet on the east
shore n and between 5000 and 6000 feet on the west shore.
The bordering ice-caps within these areas are developed in
special perfection on the islands of the archipelago about
King Oscars fjord and Kaiser Franz Josef fjord on the east
coast near latitude 75 N., as these have been mapped by the
Swedish Greenland Expedition of 1899 (see Fig. 75). 12 The

FIG. 75. Map of the region about King Oscars and Kaiser Franz Josef fjords,
Eastern Greenland, showing the areas of the numerous ice-caps (after P. Dusen).

work of mountain glaciers about King Oscars fjord is clearly
displayed by Nathorst's photograph reproduced in plate 22
A. Essentially the same features are shown also to the right
in Fig. 72 (p. 122).

PLATE 22.

A. Fretted upland carved by mountain glaciers about King Oscar's Fjord, eastern
Greenland. The highest points are from 1360 to 1570 metres above the sea
(after Nathorst).

B. Front of the Bryant glacier tongue showing the vertical wall and stratification
of ice. It also shows the absence of rock debris from the upper layers (after
Chamberlin).

THE CONTINENTAL GLACIER OF GREENLAND 125

While we are without absolute knowledge of the relief of
the land beneath most of the inland-ice, we know that the
mountainous upland of the
coast extends well within
the ice margins, since the
peaks project through the
surface as ice-bounded rock
islands or nunataks. The
irregularities of this base-
ment and the submergence
and consequent drowning
of the valleys to form deep
fjords within the marginal
zones, largely account for
the markedly lobate out-
lines of the so-called isblink
or inland-ice, as well as for
the ice-caps and mountain
glaciers, which, originating
in the outlying plateaus
and mountains, form a
fringe about the central ice
mass.

It has been shown to be
characteristic of the ice-
caps and smaller inland-ice
areas of the Arctic region
outside of Greenland, that
their lobate margins are in
part accounted for by ex-
tensions Of the Cap Upon FIG. 76. Map of a glacier tongue, which

the plateau between inter- "^ ** * h , e inland -i ce down the

f Umanak fjord (after von Drygalski).

sectmg valleys or fjords, as

well as by extensions down these valleys. These latter

126 CHARACTERISTICS OF EXISTING GLACIERS

extensions of the ice-sheets are, however, much the nar-
rower. Identically the same features are found to char-
acterize the Greenland inland-ice as well. The manner in
which this occurs in Greenland has been well brought out
in a map and section by Helland 13 of the Kangerdlugsuak
fjord and glacier, but even better by recent maps of the
Petermann fjord by Peary (Fig. 81, p. 133) and the Umanak
fjord by von Drygalski 14 (see Fig. 76). The manner in which
the ice sometimes descends from the higher levels over the
steep walls of the fjords has been strikingly brought out in a
photograph of the Foetal glacier (see Fig. 77). 15

FIG. 77. Tongues of ice descending from the Fostal glacier, McCormick Bay

(after Peary) .

As already stated, within one limited stretch upon the
west coast the ice mantle overlaps the borders of the con-
tinent and reaches the sea in a broad front. This stretch
of coast begins near the Devil's Thumb at about latitude

THE CONTINENTAL GLACIER OF GREENLAND 127

North East
and.

74 30' and extends, with some interruption, for about 150
miles along the coast of Melville bay. 16 On the northeast
coast the recent explorations of the Danes indicate that
there are two
stretches of 20
and 60 miles,
respectively,
within which
the ice in like
manner reaches
the sea. These
occur on Jokull bay
and on the north
shore of the North East
Foreland (see Fig. 78). 17
The Ice Face or Front.
Concerning the form
of the front of the inland-
ice where it lies upon the
land, widely different
descriptions have
been furnished from
different districts.
It is necessary to
remember that the
continent of Green-
land stretches northward through nearly 24 of latitude, and
after due regard is had to this consideration, the differences
in configuration may, perhaps, be found to be but expres-
sions of climatic variation. Those who have studied the
land margin of the isblink in North Greenland, all call
attention to the precipitous and generally vertical wall
which forms the ice face (see plate 22 B). As a result of
shearing and overthrusting movements within the ice near

FIG. 78. Map of the Greenland shore in the
vicinity of the North East Foreland (after
Trolle).

128 CHARACTERISTICS OF EXISTING GLACIERS

its margin, as well as to the effect of greater melting about
the rock fragments imbedded in the lower layers of the ice,
the face sometimes even overhangs in a massive ice cornice
at the summit of the wall (see plate 23 A). 18

That to this remarkable steepness of the ice face as ob-
served north of Cape York there are exceptions, has been
mentioned by both Chamberlin and Salisbury, but Peary
has also emphasized the vertical face as a widely character-
istic feature of North Greenland. The recent Danish Expe-
dition to the northeast coast of Greenland has likewise fur-
nished examples of such vertical walls. An instance where
the ice face appears as a beautifully jointed surface some-
what resembling the rectangular joint walls in the quarry
faces of certain compact limestones, is reproduced from the
report of the expedition in plate 23 B. 19

Attention has already been called to the precipitous front,
the so-called " Chinese Wall," which Lieutenant Lockwood
found to form the land face of the inland-ice of Ellesmere
Land a face which was followed up and down over irregu-
larities of the land surface, and whose height in 'one place
was roughly measured as 143 feet (see Fig. 68, p. 116).

From central and southern Greenland, on the other hand,
we hear little of such ice cliffs as have been described, and
Tarr in studies about the margin of the Cornell extension of
the isblink 20 has shown that here the vertical face is the ex-
ception. 21 The normal sloping face as there seen is repre-
sented in plate 24 A. In following the ice face for fifteen
miles, its slopes were here found to be sufficiently moderate
to permit of frequent and easy ascent and descent. Inas-
much as these sloping forms are characteristic of the ice
front in the warmer zones, and further correspond to that
generally characteristic of mountain glaciers in lower lati-
tudes, it seems likely that its occurrence in Greenland is
limited to districts where surface ablation plays a larger role.

PLATE 23.

Portion of the southeast face of the Tuktoo glacier tongue showing the
projection of the upper layers apparently as a result of overthrust (after
Chamberlin).

B. Ice-face at eastern margin of the inland-ice of Greenland in latitude 77 30' N.

(after Trolle).

THE CONTINENTAL GLACIER OF GREENLAND 129

In Northeast Greenland (lat. 77-82), according to the
Danes, " the frontier of the inland-ice is in some places
quite steep, in other places you might have mounted the
inland-ice without knowing it."

Features within the Marginal Zone. --The larger terraces
upon the ice-slope, Nansen has ascribed to peculiarities of
the rock floor on which the ice rests. Where the slopes
become still more accelerated toward the margin of the ice,
deep crevasses appear upon these steps running parallel to
their extension, and hence parallel to the margins of the ice.
Nansen found, however, that such crevasses were restricted
to the outer seven or eight miles on the eastern side of his
section, and to the outer twenty-five miles on its western
margin. Peary in his reconnoissance across the ice border

FIG. 79. A series of parallel crevasses on the inland-ice of South
Greenland (after Garde).

in latitude 69^-, saw such crevasses while they were open-
ing and 'the surface snow was sinking into the cleft thus
formed. The visible opening of the cleft was accompanied
by peculiar muffled reports which rumbled away beneath
the crust in every direction. 22

In addition to the crevasses which develop transversely

130 CHARACTERISTICS OF EXISTING GLACIERS

to the main direction of ice movement, and which are with
much probability located over " steps " in the rock floor,
there are evidently others which fall in a somewhat different
category. The series of parallel crevasses resembling
ravines which are figured by Garde 23 and take their course
over the gently swelling surface of the ice (see Fig. 79)
bear more resemblance to the longitudinal crevasses which
one finds between the nunataks upon the surface of the
plateau glaciers of Norway, as, for example, the Hardanger-
jokull. Of very considerable interest also are the rec-
tangular networks of crevasses which are described by the
same author from near the margin of the ice (see Fig. 80) , 24

FIG. 80. Rectangular network of crevasses on the surface of the inland-
ice near its margin in South Greenland (after Garde).

This network recalls the rectangular system of crevasses
which was observed by the German Expedition on the in-
land-ice of Kaiser Wilhelm Land (see Fig. 129).

Of the terraced slope and its fading into the plateau above
Peary says :

The surface of the " ice-blink " near the margin is a succession of
rounded hummocks, steepest and highest on their landward sides,
which are sometimes precipitous. Farther in these hummocks

PLATE 24.

A. Normal slope of the inland-ice at the land margin near the Cornell tongue

(after Tarr).

B. Hummocky moraine on the margin of the Cornell glacier tongue (after Tarr).

THE CONTINENTAL GLACIER OF GREENLAND 131

merge into long, flat swells, which in turn decrease in height towards
the interior, until at last a flat gently rising plain is revealed which
doubtless becomes ultimately level. 25

In sketching the general form of the Greenland conti-
nental glacier, it has been stated that the highest portion of
the shield lies to the eastward of the medial line of the con-
tinent. This is shown by Nansen's section, and is empha-
sized by Peary, who says : -

That the crest of the Greenland continental ice divide is east of
the country's median line there can be no doubt. 26

By von Drygalski 27 this lack of symmetry of the ice
mass has been ascribed to excessive nourishment upon the
east, whereas the losses from melting and from the discharge
of bergs occur mainly upon the west. The mountains of
the east are, he states, completely surrounded by ice so that
peaks alone project, while the mountains of the west stand
isolated from the ice. In attempting to make the eccentric
position of the boss in the ice shield depend upon the con-
figuration of the underlying rock surface, von Drygalski
has been less convincing, for we know that the Scandinavian
continental glacier of Pleistocene times moved northwest-
ward from the highest surface of the ice-shield up the grade
of the rock floor, and pushed out through portals in the moun-
tain barrier which lies along the common boundary of
Sweden and Norway.

We shall see, moreover, that the nourishment of the
Greenland ice is by a different process than that which he
has assumed. Still there would appear to be a clear parallel
between the marginal terraces of the inland-ice with their
crevassed steep surfaces, and the plateaus and ice-falls which
alternate upon the slopes of every mountain glacier which
descends rapidly in its valley.

132 CHARACTERISTICS OF EXISTING GLACIERS

Superimposed upon the flats of the larger ice terraces r
there are undulations of a secondary order of magnitude,
and these Nansen ascribed to the drifting of snow by the
wind. To the important action of wind in moulding the
surface of the inland-ice we shall refer again. There are in
addition many other irregularities of the surface due to
differential melting, and while of very great interest, their
consideration may profitably be deferred until the meteoro-
logical conditions of the region have been discussed. There
are, however, other features which like the broader terraces
are clearly independent of meteorological conditions, and
which are, therefore, best considered in this connection.

Dimples or Basins of Exudation above the Marginal
Tongues. Seen from the sea in Melville bay on the north-
west coast, the inland-ice offers special advantages for observ-
ing its contours in sections parallel to its front, that is to say r
in front elevation. Here only upon the west coast the ice
extends beyond the borders of the land and is cut back by
the sea to form cliffs. These ice cliffs are interrupted by
rocky promontories which are surrounded on all sides but
the front by ice, and hence in reality the cliff furnishes us
with sections through nunataks and inland-ice alike. Says-
Chamberlin : 28

Only a few of the promontories of the coast rise high enough to
be projected across this sky-line and interrupt the otherwise con-
tinuous stretch of the glacial horizon. The ice does not meet the
sky in a simple straight line. It undulates gently, indicating some
notable departure of the upper surface of the ice tract from a plane.
As the ice-field slopes down from the interior to the border of the
bay, it takes on a still more pronounced undulatory surface. It is
not unlike some of our gracefully rolling prairies as they descend
from uplands to valleys, when near their middle-life development.

The two 1200-mile sledge journeys of Peary in the years-
1891-1892 and 1893-1895 across the northern margin of the-

THE CONTINENTAL GLACIER OF GREENLAND 133

" Great Ice " of Green-
land, have added much
to our knowledge of the
physiography of the
inland-ice. These jour-
neys were made on
nearly parallel lines at
different distances from
the ice border, and so ;
if studied in relation to
each other, they display
to advantage the con-
figuration of the ice
surface near its margin
(see Fig. 81). The
routes were for the most
part nearly straight and
ran at nearly uniform
elevations which ranged
from 5000 to 8000 feet
above the sea. 29 In the
sections nearest the
coast, however, the
route at first ascended
a gentle rise to a flatly
domed crest upon the
ice, only to descend sub-
sequently into a broad
swale of the surface, the
bottom of which might
be described as a plain,
and which was con-
tinued in the direc-
tion of the coast by a
tongue-like extension of

FIG. 81. Map showing routes of sledge jour-
neys in North Greenland in their relation to
the margin of the ice (after Peary).

134 CHARACTERISTICS OF EXISTING GLACIERS

the ice, such as the tongue in Petermann fjord between Hall
Land and Washington Land (Fig. 81). On the farther side
of this basin-like depression, the surface again rose until
another domed crest had been reached, after which a
descent began into a swale similar to the first. On the
return journey by keeping farther from the ice margin these
elongated dimples upon the ice surface were avoided. The
broad domed surfaces which separate the dimples clearly lie
over the land ridges between the valleys down which the
glacier tongues descend toward the sea.

Peary has referred to these dimples on the surface of the
inland-ice as " basins of exudation/' and has compared the
cross profile in its ups and downs to that of a railroad located
along the foot-hills of a mountain system. 30 In his earlier
reconnaissance of the isblink from near Disco Bay, Peary
describes such a dimple above the Jakobshavn ice tongue
" stretching eastward into the ' ice-blink/ like a great bay,"
as a feeder basin. 31 The exact form of such dimples upon the
ice surface is well brought out in von Drygalski's map
of the Asakak glacier tongue on the Umanak fjord (see
Fig. 76, p. 125). 32

We may easily account for the existence of these dimples
by drawing a parallel from the behavior of water as it is
being discharged from a lake through a narrow and steeply
inclined channel. Under these circumstances the surface
is depressed through the indrawing of the water on all sides
to supply the demands of the outflowing current. That
within the upper portions of the glacier tongues of the Green-
land isblink the ice flows with a quite extraordinary velocity
has long been known. Values as high as 100 feet per day
have been determined upon the Upernavik glacier. 33 By
more accurate methods, von Drygalski has obtained on one
of the ice tongues which descends to a fjord a rate of about
18 meters or 59 feet in twenty-four hours. 34 Upon the in-

THE CONTINENTAL GLACIER OF GREENLAND 135

land-ice some distance back from the head of the fjord, on
the other hand, a rate was measured of only one to two
centimeters per day.

Scape Colks and Surface Moraines. The velocities of
ice movement which obtain within and about the heads
of the glacial outlets are, there is thus every reason to be-
lieve, as different as possible from the ordinary general out-
ward movement of the inland-ice. Within this marginal
zone areas of exceptional velocity of the inland-ice are likely
to be found wherever its progress is interfered with by the
projecting nunataks. Just as jetties by constricting the
channels greatly accelerate the velocity of stream flow
within those channels, so here within the space between
neighboring nunataks a local high rate of flow in the ice is
developed. An inevitable and quite important consequence
of this constriction was long ago pointed out by Suess and
illustrated by the area between Dalager's nunataks near the
southwestern border of the isblink. 35 Here again the conduct
of water which is being discharged through narrow outlets
has supplied both the illustration and the explanation. In
the regulation of the flow of the Danube below Vienna, the
river was partially closed by a dam, the Neu-Haufen dike,
and the floor in the channel below the dike was paved with
heavy stone blocks. The effect of thus narrowing the chan-
nel of the river was to raise the level of the water above the
dike by almost a metre, and under this increased head
the current tore out the heavy stone paving of the floor of
the channel and dug a depression above as well as below
the outlet. This excavation by the current represented a
hole dug to a depth of about fifteen metres. The blocks
which had been torn out from the pavement were left in a
crescent-shaped border to the depression upon its down-
stream side (see Fig. 82 a).

The position of a surface moraine which stretches in a

136 CHARACTERISTICS OF EXISTING GLACIERS

sweeping arc from the lower edge of one of Dalager's nuna-
taks to a similar point upon its neighbor, indicates a com-
plete parallel between the motion of the ice and the water
at the Neu-Haufen dyke, the rock debris of the deeper ice

FIG. 82. a, Closure of the Neu-Haufen dyke, Schilttau in the regulation of the
Danube below Vienna (after Taussig) ; b, Scape colks near Dalager's Nunataks
(after Jensen and Kornerup).

layers being here brought up to the surface. Study of the
Scandinavian inland-ice of late Pleistocene times throws
additional light upon the nature of this process. Flowing
from a central boss near the head of the Gulf of Bothnia,
the ice pushed westward and escaped through narrow portals
in the escarpment which now follows the international boun-
dary of Sweden and Norway. This constriction of its current
has been appealed to by Suess to account for the interesting
glint lakes which to-day lie across this barrier and extend
both above and below the former outlets for the ice. 36 Lakes
which have this origin he has described under the term " scape
colks." Perhaps if examined more carefully, we should find
that the bringing up of the englacial debris to the surface
of the ice, is only partially due to the inertia of motion in the
ice. With the more rapid flow of the ice within the con-
stricted portion, the basic layers, shod as they are with
rock fragments, accomplish excessive abrasion upon the rock

THE CONTINENTAL GLACIER OF GREENLAND 137

bed. This is in accord with Penck's law of adjusted cross-
sections in glacial erosion. Where the ice channel broadens
below the nunataks, the abrasion again becomes normal so
that a wall develops at this place in the course of the stream.

Here, therefore, a new process comes into play due to the
peculiar properties of the plastic ice, a process which has been
illustrated in the formation of drumlins beneath former
continental glaciers, and has been given an experimental
verification. Case has shown that paraffin mixed with
proper proportions of refined petroleum, and maintained at
suitable temperatures, can be forced by means of plungers 37
through narrow boxes open at both ends. It was shown
in the experiments that an obstruction interposed at the
bottom and in the path of the moving paraffin, forced the
bottom layers upward, and this upward movement continued
beyond the position of the obstruction. The experiments
of Hess 38 give results which are consistent with those of Case.
Hess employed in his experiments parallel wax disks of alter-
nating red and white colors, and these were forced under
hydraulic pressure through a small opening. It was found
that the layers turn up to the surface in this " model glacier "
apparently as a result of the friction upon the bottom, and
at only moderate distances from the opening where the
energy of the active moving substance pressing from the rear
has to some extent been dissipated.

In Chamberlin's studies of certain Greenland glaciers, he
was permitted to observe the effect upon the motion of the
glacier of a low prominence in its bed. These observations
are confirmatory of the experiments described. 39

The swirl colks or eddies which Suess has suggested as
occurring below nunataks, in order to account for certain
lakes in Norway, seem to be much less clear, and it is a
little difficult to assume an eddy in the ice which is in any
way comparable to the eddies of water.

138 CHARACTERISTICS OF EXISTING GLACIERS

Marginal Moraines. Inasmuch as the rock appears above
the surface of the ice of the Greenland continental glacier
only in the vicinity of its margins, and here only as small
islands or nunataks, the rock debris carried by the Green-
land ice must be derived almost solely from its basement.
As described in detail by Chamberlin, it is the lower 100
feet of ice to which englacial debris is largely restricted. 40
Medial moraines, if the term may be properly applied to
those ridges of rock debris which upon the surface of the ice
go out from the. lower angles of nunataks, have been fre-
quently described by Nansen and others. They seem to
differ but little from certain of the medial moraines which
have been described in connection with the larger mountain
glaciers.

Nansen has mentioned heavy terminal moraines in the
Austmann Valley, where he came down from the inland-ice
after crossing the continent. The material of these moraines
consisted mainly of rounded and polished rock fragments,
and is obviously englacial material. 41 Along the land margin
of the Cornell ice tongue Tarr found a nearly continuous
morainic ridge parallel to the ice front. This ridge usually
rests at the base of the ice foot, and is sometimes a part of
this foot, wherever debris has accumulated and protected
the ice beneath from the warmth of the sun. Such an accu-
mulation causes this part of the glacier to rise as a ridge.
In other cases the ridge is, however, separated from the ice
margin, and sometimes there are several parallel ridges from
which the ice front has successively withdrawn 42 (see plate
24 B).

According to von Drygalski the marginal moraines of the
Greenland ice sheet, as regards their occurrence, form, and
composition, are in every way like those remaining in
Northern Europe from the time of the Pleistocene glaciation,
and this is true of those which run along the present border

THE CONTINENTAL GLACIER OF GREENLAND 139

of the inland-ice as well as of those still mightier ancient
moraines which follow at certain distances. 43 These moraines
are generally closely packed blocks with relatively slight
admixture of finer material. They are the largest where the

FIG. 83. Diagram to show the effect of a basal obstruction in the path of the ice
near its margin (after Chamberlin).

ice border enters the plains, or pushes out upon a gentle
slope, and they are smallest where the ice passes steep rocky
angles.

It is worthy of note that the marginal moraines of Green-
land become locally so compact and resistant that they

FIG. 84. Surface marginal moraine of the inland-ice of Greenland (after

Chamberlin).

oppose a firm obstruction to the ice movement. Then the
ice pushes out laterally into the marginal lakes which develop
there or pushes up upon the moraines. It thus comes to

140 CHARACTERISTICS OF EXISTING GLACIERS

arrange its layers parallel to the slope of the morainic sur-
face or, in other words, so that they dip toward the ice. 44

Another type of marginal moraine which was mentioned
by Mohn and Nansen from South Greenland, and later fully
described by Chamberlin from North Greenland, is explained
by the upturning effect of obstructions in the bed, and by
the shearing and overthrusting movements which are found
to exist in inland-ice near its margin 45 (see Figs. 83 and 84).
This process has much in common with that which we have
already described in connection with scape colks.

Fluvio-glacial Deposits. Where studied by Chamberlin
near Inglefield gulf, there appears to be little or no gush-
ing of water from beneath the inland-ice. Small streamlets
only appeared beneath the ice border, bringing gravel and
sand which they distributed among the coarser morainic
material. So far as land has been recently uncovered b^^iie
ice in North Greenland, and so far as differentiated from the
topography of the underlying rock, it was found to be nearly
plane. So far as known, no eskers have been observed
about the border of the inland-ice of Greenland, and only
a few irregular kames near Olrik's bay. 46

REFERENCES

1 F. Nansen, " The First Crossing of Greenland," vol. 2, p. 404 ; R. E.
Peary, "Journeys in North Greenland," Geogr. Jour., vol. II., 1898, p. 232.

2 T. C. Chamberlin, "Glacial Studies in Greenland," III., Jour. GeoL,
vol. 3, 1895, pp. 62-63.

3 J. A. D. Jensen, " Expeditionen till Syd-Gronland, 1878," Meddelel-
ser om Gronland, heft 1, pp. 17-76.

4 J. V. Garde, "Beskrivelse of Expeditionen til Sydvest Gronland, 1893,''
Meddelelser om Gronland, heft 16, 1895, pp. 1-72.

5 Garde, I.e., pi. 7.

6 1. I. Hayes, " The open polar sea," London, 1867, p. 72.

7 H. Mohn und Fridtjof Nansen, " Wissenschaf tlichen Ergebnisse von
Dr. F. Nansens Durchquerung von Gronland, 1888," Pet. Mitt., Erganz-
ungsh., vol. 105, 1892, pp. 1-111, 6 pis., 10 figs. Especially Plate 5.

8 " Die Gletscher," pp. 105-106.

9 R. E. Peary, "A Reconnoissance of the Greenland Inland-ice," Jour.
Am. Geogr. Soc., vol. 19, 1887, pp. 261-289.

10 "Iceblink," which has been suggested by some writers, is a term gen-

THE CONTINENTAL GLACIER OF GREENLAND 141

erally applied among navigators to describe the appearance of ice on the
horizon, and is contrasted with "land blink," which describes the peculiar
loom of the land. In order to apply the term to the inland-ice without
confusion, it is, therefore, better to retain the Danish form of the word.

11 Petermann Peak near Franz Josef fjord on the east coast, which,
according to Nansen, has an estimated height of 11,000-14,000 feet,
has recently been shown to be not more than half that height (A. G.
Nathorst, Pet. Mitt., vol. 46, 1899, p. 242).

12 A. G. Nathorst, "Den svenska expeditionen till nordostra Gronland,"
1899, Ymer, vol. 20, 1900, map 11.

13 A. Helland, "On the Ice Fjords of North Greenland and on the Forma-
tion of Fjords, Lakes, and Cirques in Norway and Greenland," Quart.
Jour. Geol. Soc., vol. 33, 1877, pp. 142-176.

14 E. von Drygalski, "Gronland-Expedition," vol. 1, 1897, Map 7.

15 R. E. Peary, "Journey in North Greenland," Geogr. Jour., vol. 11,
1898, pp. 213-240.

16 T. C. Chamberlin, "Glacial Studies in Greenland, III.," Jour. Geol.,
vol. 3, 1895, p. 61.

17 Lieut. A. Trolle, "The Danish Northeast Greenland Expedition,"
Scot. Geogr. Mag., vol. 25, 1909, pp. 57-70, map and illustrations.

18 Chamberlin, Jour. Geol., vol. 3, 1895, p. 566. Salisbury, ibid., vol. 4,
li, p. 778.

19 Gunnar Andersson, .' ' Danmarks expeditionen till Gronlands nordost-
kust," Ymer, vol. 28, 1908, pp. 225-239, maps and 7 figures.

20 To the south of the upper Nugsuak Peninsula in latitude 70 10' N.

21 R. S. Tarr, "The Margin of the Cornell Glacier," Am. Geologist, vol.
20, 1897, pp. 139-156, pis. 6-12.

22 Jour. Am. Geogr. Soc., vol. 19, 1887, p. 277.

23 Garde, I.e., pi. IV. See also J. A. D. Jensen, "Expeditionen till Syd-
Gronland, 1878," Meddelelser om Gronland, heft 1, pi. ii.

24 Garde, I.e., pi. V.

25 Geogr. Jour., vol. 11, 1898, pp. 217, 218.

26 Geogr. Jour., I.e., p. 232.

27 E. von Drygalski, "Die Eisbewegung, ihre physikalischen Ursachen
und ihre geographischen Wirkungen," Pet. Mitt., vol. 44, 1898, pp. 55-64.

28 "Glacial Studies in Greenland, III.," Jour. Geol., vol. 3, 1895, p. 63.

29 Geogr. Jour., vol. 11, 1898, p. 215. See also his map, Bull. Am. Geogr.
Soc., vol. 35, 1903, p. 496.

30 Geogr. Jour., vol. 11, p. 232.

31 Jour. Am. Geogr. Soc., vol. 19, 1887, p. 269.

32 " Gronland-Expedition," I.e., map 7.

33 Lieut. C. Ryder in 1886. Helland on a glacier of the Jakobshavnf jord
found a rate of 64 feet daily.

34 E. von Drygalski, "Die Bewegung des antarktischen Inlandeises,"
Zeitsch. f. Gletscherk, vol. 1, 1906-7, pp. 61-65.

35 Ed. Suess, "Face of the Earth," vol. 2, 1888 (translation, 1906), pp.
342-344.

142 CHARACTERISTICS OF EXISTING GLACIERS

36 Suess, I.e., pp. 337-346.

37 E. C. Case, "Experiments in Ice Motion," Jour. GeoL, vol. 3, 1895, pp.
918-934.

38 "Die Gletscher," 1904, p. 171, fig. 28.

39 "Recent Glacial Studies in Greenland." Annual address of the Presi-
dent of the Geological Society of America, Bull. Geol. Soc. Am., vol. 6,
1895, pp. 199-220, pis. 3-10.

40 Chamberlin, I.e., p. 205.

41 Mohn u. Nansen, " Wissenschaf tlichen Ergebnisse von Dr. F. Nan-
sen's Durchquerung von Gronland, 1888," Pet. Mitt., Erganzungsh., vol.
105, 1892, p. 91.

42 R. S. Tarr, "The Margin of the Cornell Glacier," Am. GeoL, vol. 20,
1897, p. 148.

43 "Gronland Expedition," I.e.

44 von Drygalski, I.e., p. 529.

45 Chamberlin, I.e., p. 92.

46 Salisbury, I.e., p. 809.

CHAPTER IX
NOURISHMENT OF THE GREENLAND INLAND-ICE

Few and Inexact Data. The problems involving the
gains and the losses of the inland-ice of Greenland require
for their satisfactory solution a much larger body of exact
data than we now possess. Barring a few scattered and not
always exact or reliable observations, we are practically
without knowledge of the amount or the variations of atmos-
pheric pressure, or of snowfall away from the coastal areas
of the continent. Even within these marginal zones, the
losses from ablation and through the calving of bergs, have
been estimated by crude methods only. Again, the great
height of the ice surface within the central plateau, and the
lack of any knowledge of the elevation of the land surface
in those regions, has raised questions concerning the con-
ditions of flow and of fusion upon the bottom which will
probably long remain subjects of controversy.

An international cooperative undertaking with one or
more stations established in the interior at points where
altitude has been determined by other than barometric
methods, and with coast stations maintained contempo-
raneously and for a period of at least a year, particularly
if they could be supplemented by balloon or kite observa-
tions, would yield results of the very greatest importance. 1
The Greenland ice having shrunk greatly since the Pleis-

143

144 CHARACTERISTICS OF EXISTING GLACIERS

tocene period, it is almost certain that its alimentation to-
day does not equal the losses which it suffers along its mar-
gins which in but slightly altered form applies to the
Antarctic continental glacier as well.

Snowfall in the Interior of Greenland. Almost the only
data upon this subject are derived from a rough section of
the surface layers of snow, as this was determined by Nansen
with the use of a staff near the highest point in his journey
across the inland-ice along the 64th parallel. At elevations
in excess of 2270 meters Nansen found the surface snow
" soft " and freshly fallen, but of dust like fineness. Be-
neath the surface layer, a few inches in thickness only, there
was a crust less than an inch in thickness which was ascribed
to the slight melting of the surface in midsummer, 2 and below
this crust other layers of the fine " frost snow " more and
more compact in the lower portions, but reaching a thickness
of fifteen inches or thereabouts before another crust and
layer was encountered. 3 Other sections made in like manner,
by pushing down a staff, revealed similar stratification of
the surface snow with individual layers never exceeding
in thickness a few feet. From these observations Nansen
has drawn the conclusion that the layers of his sections cor-
respond to seasonal snowfalls, the thin crust upon the sur-
face of each being due to surface melting in the few warm
days of midsummer. He cites Nordenskjold as believing
that the moist winds which reach the continent of Greenland
deposit most of their moisture near the margin. 4

The sky during almost the entire time of the journey is
described by Nansen as so nearly clear that the sun could
be seen, and there were few days in which the sky was com-
pletely overcast. Even when snow was falling, which often
happened, the falling snow was not thick enough to pre-
vent the sun showing through. This clearly indicates that
the snow falls from layers of air very near the snow surface

NOURISHMENT OF THE GREENLAND INLAND-ICE 145

below. The particles which fell were always fine, like frozen
mist what in certain parts of Norway is known as
" frost snow "; that is, snow which falls without the mois-
ture first passing through the cloud stage. 5

The air temperatures, even in August and September,
when the crossing was made, were on the highest levels
seldom much above the zero of the Fahrenheit scale, and at
night they sank by over 40 F. (in one case to 50 F.).

Peary, while on the inland-ice in North Greenland in the
month of March, 1894, registered on his thermograph a
temperature of 66 F., and several of his dogs were frozen
as they slept. 6 The high altitudes and the general absence
of thick clouds over the inland-ice permit rapid radiation,
so that cold snow wastes and hot sand deserts have in com-
mon the property of wide diurnal ranges of temperature.
The poverty of the air over the inland-ice in its content of
carbon dioxide, as shown by the analysis of samples collected
by Nansen, must greatly facilitate this daily temperature
change. 7

From studies in the Antarctic it is now known that most
of the snow falls there in the summer season, and that little,
if any, moisture can reach the interior from surface winds.
The same is probably true also of the interior of Greenland.

Though the absolute humidity of the air upon the ice
plateau of Greenland is always low, the relative humidity
is large, and never below 73 per cent of saturation in
the levels above 1000 metres. Evaporation occurs chiefly
when the sun is relatively high, and when the air is again
chilled, the abstracted moisture is returned to the surface in
the form of the almost daily snow mists or frost snow. The
observations went to show that only in the warmest days
of summer do the sun's rays succeed in melting a very thin
surface layer of the snow. Of the thirty days that Nan-
sen's party was at altitudes in excess of 1000 metres, on

146 CHARACTERISTICS OF EXISTING GLACIERS

only six is a definite snowfall reported. Within the in-
terior of Greenland it appears that no snow whatever is per-
manently lost from the surface by melting*

While the relative humidity of the air over the central
plateau is so high, the absolute humidity is extremely low,
being measured from 1.4 mm. to 4 mm., though generally
much below the maximum value. The average absolute
humidity was 2.5 mm., while the average relative humidity
was 92 per cent. 9

It has been claimed by von Drygalski that the eastern
portion of the Greenland ice sheet is a great nourishing
region, while the western slope, on the other hand, is the
locus of excessive melting and discharge. In support of
this view he adduces chiefly the admitted lack of symmetry
of the ice mass. 10 So far as alimentation is concerned, the
view does not seem to be as yet supported by any observa-
tions, and it can hardly be regarded as a tenable hypothesis.

The Circulation of Air over the Isblink. No exact data
upon atmospheric pressures are as yet available, except from
stations near the sea level, mainly along the western and
northern coasts. Until stations have been maintained
for a more or less protracted period within the interior of
Greenland, none can be expected. None the less, upon
the basis of the observed winds in those portions of Green-
land which have been traversed, it may be safely asserted
that a fixed area of high atmospheric pressure is centered
over the Greenland isblink, and that the cold surface of
this mass of ice is directly responsible for its location there.
Nansen, as early as 1890, announced this fact, having
observed " that the winds which prevail on the coasts
have an especial tendency to blow outwards at all points." 11
After many years of experience in different portions of
Greenland, Peary stated the law of air circulation above the
continent in clear and forceful language: 12

NOURISHMENT OF THE GREENLAND INLAND-ICE 147

Except during atmospheric disturbances of exceptional mag-
nitude, which cause storms to sweep across the country against all
ordinary rules, the direction of the wind of the " Great Ice" of
Greenland is invariably radial from the centre outward, normal
to the nearest part of the coast-land ribbon. So steady is this wind
and so closely does it adhere to this normal course, that I can liken
it only to the flow of a sheet of water descending the slopes from
the central interior to the coast. The direction of the nearest land
is always easily determinable in this way. The neighborhood of
great fjords is always indicated by a change in the wind's direction ;
and the crossing of a divide, by an area of calm or variable winds,
followed by winds in the opposite direction, independent of any
indications of the barometer.

Except for light sea breezes blowing on to the land in
February, the Danish Northeast Greenland expedition found
" the wind was constantly from the northwest, this being
the result of the high pressure of air which is found over
the inland ice." 13

FIG. 85. Diagram to illustrate the air circulation over the isblink of Greenland.

These conditions of circulation are schematically repre-
sented in Fig. 85. In March, 1894, Peary encountered on
the north slope of the inland-ice a series of blizzards before
unprecedented in Arctic work, one lasting for three days,
during which for a period of 34 hours the average wind ve-
locity, as recorded by anemometer, was 48 miles per hour.
Viewed in the light of the violent southerly blizzards which
Shackelton found to prevail upon the ice plateau in the
Antarctic, these winds must be considered as belonging to

148 CHARACTERISTICS OF EXISTING GLACIERS

the same Greenland or isblink system which has been de-
scribed as of such general prevalence.

After comparing the meteorological data from his journey
with contemporaneous observations on the shores of Baffin's
Bay, Nansen believed that he was able to make out faintly
the influence of general cyclonic movements. He says: 14

The plateau seems to be too high and the air too cold to allow
depressions or storm centres to pass across, though, nevertheless,
our observations show that in several instances the depressions of
Baffin's Bay, Davis' Strait, and Denmark's Strait, can make them-
selves felt in the interior.

This, it must be remembered, was in the narrow southern
extension of the continent and essentially marginal to the
main ice mass. Commenting upon Peary's conclusions
above quoted, Chamberlin 15 ascribes the wind which flows
downward and outward from the isblink to a notable
increase of its specific gravity through contact with and
consequent cooling by the snow surface. 16

It is perhaps well to here allude to the conditions under
which heat is added to or abstracted from fluid masses,
whether they be liquid or gaseous. Communication is by
contact, and distribution by the process known as convec-
tion adjustments of position due to changes in specific
gravity resulting from change of temperature. These con-
vection currents must be started either by rendering the
upper layers heavier, or the lower layers lighter, than they
were when in equilibrium. No distributing convection
currents can be set up by heating (making lighter) at the
top, or by cooling (making heavier) at the bottom; and
so long as confined, no motions of any kind can thus be
initiated. Water may be boiled in the upper layers within
a test-tube, or frozen in the lower layers, without disturb-
ing the conditions of equilibrium.

NOURISHMENT OF THE GREENLAND INLAND-ICE 149

Now it is at the bottom that the air above the isblink
is cooled by contact; and it is due to the peculiar shield-
like form of this ice mass that the heavier cooled bottom
layer is able to slide off radially as would a film of oil from
a model of similar form. The centrifugal nature of this
motion tends to produce a vacuum above the central area
of the ice mass, and air must be drawn down from the
upper layers of the atmosphere in order to supply the void.
It is here that is located the " eye " of the anticyclone.

Foehn Winds within the Coastal Belt. --The sliding
down of masses of heavy air upon the snow surface of the
Greenland ice must bring about adiabatic heating of the
air and a consequent elevation of the dew point. The in-
crease of temperature being about 1 C. for every 100 metres
of descent, a rise of temperature of as much as 20 C. or 36
F. will result in a descent from the summit of the plateau,
assuming this to have an elevation of 10,000 feet. Some
reduction in the amount of this change of temperature will,
of course, result from the contact of the air with the cold
snow surface during its descent, this modification being
obviously dependent upon the velocity of the current.
The warm, dry winds which in different districts have been
described under the names foehn and chinook are the inev-
itable consequence of such conditions, and are, moreover,
particularly characteristic of steep mountain slopes more
or less covered by glaciers. Such foehn winds have long
been recognized as especially characteristic of western
Greenland. Dr. Henry Rink, who was a pioneer in the
scientific study of Greenland, referring to these winds,
wrote in 1877 : 17 -

Among the prevailing winds in Greenland the warm land wind
is the most remarkable. Its direction varies according to locality
from true E.S.E. to E.N.E. always proceeding though warm from
the ice-covered interior, and generally following the direction of

150 CHARACTERISTICS OF EXISTING GLACIERS

the fjord. It blows as frequently and as violently in the north
as in the south, but more especially at the fjord heads, while at
the same time in certain localities it is scarcely perceptible. It
often turns into a sudden gale; the squalls in some fjords rushing
down between the high rocks, in certain spots often sweep the
surface of the water with the force of a hurricane, raising columns
of fog, while the surrounding surface of the sea remains smooth.

Nansen encountered one of these foehn winds on his de-
scent, and Peary mentions their occurrence in the north.
In Scoresby's Land on the east coast, a foehn wind in the
winter season has been known in a single hour to change
the temperature by 24 C. (or 43 F.), and the maximum
change during such a wind is far greater. It is not yet
known from observations to what distance above the ice
surface the winds of the Greenland system extend, or how
the broad cyclonic areas of the atmosphere are modified.
The anti-cyclone of the continent is, however, none the less
clear and constant and is centred over the high interior.
Nansen has remarked the calms over the divide of his
section. 18

There is some evidence that in adopting the important
modern laws of adiabatic cooling of the air, we have allowed
the pendulum to swing too far, and have given too little
weight to the effect of cooling through contact of air with
either rock or snow. The latest results of Antarctic ex-
peditions furnish the most striking proof of this, if other than
Greenland examples were needed, and the Antarctic studies
throw much light upon the conditions of snow distribution
which are observed in Greenland.

Wind Transportation of Snow over the Desert of Inland-
ice. Whymper and Nordenskjold each called Greenland
a " Northern Sahara." In different ways Nansen and Peary
have also instituted comparisons between the wastes of snow
in the interior of Greenland and the desert of sand of the

NOURISHMENT OF THE GREENLAND INLAND-ICE 151

Sahara. The Norwegian explorer has emphasized especially
the wide daily ranges of temperature, which because of
generally cloudless atmospheres, both deserts have in com-
mon. Of the monotonous and elemental simplicity of the
snow vistas back from the ice margin in North Greenland,
Peary says: 19

It is an Arctic Sahara, in comparison with which the African
Sahara is insignificant. For on this frozen Sahara of inner Green-
land occurs no form of life, animal or vegetable ; no fragment of
rock, no grain of sand is visible. The traveller across its frozen
wastes, travelling as I have week after week, sees outside himself
and his own party but three things in all the world, namely, the
infinite expanse of the frozen plain, the infinite dome of the cold
blue sky, and the cold white sun nothing but these (see Fig. 86).

FIG. 86. On the Sahara of snow (after Peary).

There is, however, yet another marked parallel between
the snow waste and the sand desert. It is the importance
of wind as a transporting agent. In his shorter acquaintance

152 CHARACTERISTICS OF EXISTING GLACIERS

with southern Greenland Nansen was less impressed with
this., but he has explained the secondary snow ridges upon
the marginal terraces of the inland-ice as wind accumula-
tions. 20 These long parallel ranges of snow drift thus cor-
respond to the similar ranges of sand dunes which some-
times throughout a width of many miles hem in the deserts
of lower latitudes. In northern Greenland Peary's observa-
tions have a special value. He says: 21

There is one thing of especial interest to the glacialist the
transportation of snow on the ice-cap by the wind. . . .

The opinion has been forced upon me that the wind, with its
transporting effect upon the loose snow of the ice-cap, must be
counted as one of the most potent factors in preventing the increase
in height of the ice-cap a factor equal perhaps to the combined
effects of evaporation, littoral and subglacial melting, and glacial
discharge. I have walked for days in an incessant sibilant drift
of flying snow, rising to the height of the knees, sometimes to the
height of the head. If the wind becomes a gale, the air will be
thick with the blinding drift to the height of 100 feet or more. I
have seen in the autumn storms in this region round an amphi-
theatre of some 15 miles, snow pouring down in a way that reminds
one of Niagara. 22 When it is remembered that this flow of the at-
mosphere from the cold heights of the interior ice-cap to the lower
land of the coast is going on throughout the year with greater or
less intensity, and that a fine sheet of snow is being thus carried
beyond the ice-cap, to the ice-free land at every foot of the periphery
of the ice-cap, it will perhaps be seen that the above assumption
is not excessive. I feel confident that an investigation of the actual
amount of this transfer of snow by the wind is well worth the
attention of all glacialists.

Fringing Glaciers Formed from Wind Drift. In the

vicinity of Inglefield Gulf in northwest Greenland, the inland-
ice ends in a steep, snowy slope rising to a height of about
100 feet, where is a terminal moraine, above which moraine
rises the great dome of the inland-ice. The whiteness and
freshness of a portion of the snow of the outer border, when

NOURISHMENT OF THE GREENLAND INLAND-ICE 153

examined by Chamberlin, 23 showed it to be wind drift of
recent accumulation. Locally, however, older and discol-
ored snow appeared beneath the whiter surface snow, and
in a few places stratified granular ice with some included
rock debris. This snow and ice becomes augmented from
year to year and is, in Chamberlin 's opinion, a species of
fringing glacier. Such fringes were from a few rods to a
half mile in breadth, and where a favorable depression ex-
isted, one was observed extending for a mile or more down
the valley. Commander Peary has found this a dominant
feature on the north Greenland coast. Fringing glaciers
of this type have also been described by Salisbury from
the vicinity of Melville Bay. Their movement was clearly
evinced by their structure and by the debris which they
carried. 24

Nature of the Surface Snow of the Inland-ice. The
surface snow from the marginal zones of the inland-ice
has the granular form characteristic of neves, as has
been shown with exceptional clearness in elaborate studies
by von Drygalski. 25 Such grains, grown by accretions
from a single crystal nucleus and at the expense of neigh-
boring crystals, must require either fusion from tem-
porary elevation of temperature, or from pressure. The
observations of von Drygalski were made on the ice of the
marginal tongues and on the blue layers of the inland-ice;
but as the samples taken farthest from the margins were
found at a height of only 500 meters, the results throw
little light upon the conditions of surface snow within the
interior, where melting does not take place. In view of
Nordenskj old's observations in Spitzbergen 26 and recent
studies in Antarctica it is unlikely that firn or neVe snow
will be found within the interior except at some depths below
the surface.

Nansen has described the fine " frost snow " which falls

154 CHARACTERISTICS OF EXISTING GLACIERS

almost daily from an air layer near the snow surface, from
which its moisture has been derived. Melting does not
occur there, as already stated, except, perhaps, for a few
days in the height of summer when a thin crust develops
upon the surface. 27 Peary has referred to the snow at
the highest altitudes which he reached in north Green-
land as "unchanging and incoherent. " This dry hard
snow chased by the wind, has the cutting effect of sand
in a blast, and thus is offered still another parallel with
deserts and their wind blown sand. Each new storm, we
are told by Stein, 28 piles up a snowbank on the lee sides
of nunataks, but the next storm, coming from a somewhat
different direction and laden with fine hard snow, cuts
away the earlier deposit as would a sand blast. Peary dis-
covered one of his earlier snow huts partly cut away by
this process.

Snow Drift Forms of Deposition and Erosion Sastrugi.
The minor inequalities of the snow surface as determined
by the wind blowing over the inland-ice, have been mentioned
more or less persistently by all Arctic travellers, since upon
the character of this surface has so largely depended the
celerity of movement in sledge journeys. It is unfortunate
that no one has discussed the subject from a scientific stand-
point, for it has great significance in connection with the
study of the strength and direction of the wind over the
snow surface. All minor hummocks and ridges of this
nature are included under the general term sastrugi (see Fig.
87).

The student may learn much concerning their form within
the Antarctic regions from examination of the many beau-
tiful photographs recently published by the Royal Society
in connection with the British Antarctic Expedition. 29 On
plate 92 of this collection, sastrugi are shown which were
originally laid down in " elongated domes " and " crescent

NOURISHMENT OF THE GREENLAND INLAND-ICE 155

hollows/ 7 but which on account of change in the wind di-
rection the drifting snow granules have cut away both on
the soft surface and in the harder deep layers. As a result
of this erosion cross flutings have been superimposed upon the
original forms.

Our best study of snow drift forms has been made by Dr.

FIG. 87. Sastrugi on the inland-ice of North Greenland (after Peary).

Vaughan Cornish, who, after a series of monographs dealing
with waves in other materials, has spent a winter in Canada
in order to study the phenomena connected with the drifting
of snow. 30 It is found that snow which falls at temperatures
near 32 F. is wet and sticky, and behaves quite differently
from that which falls near or below the zero of the same
scale; which, on the contrary, is dry and slippery. Sub-
sequent modifications of either of these forms of snow
depend chiefly upon pressure, temperature, radiation, and

156 CHARACTERISTICS OF EXISTING GLACIERS

wind. It is the cold, dry, and granular snow only which
makes so-called normal waves, and it must be this form
which plays the major role in producing the surface irregu-
larities of the inland-ice of Greenland.

Ripples and larger waves alike, when formed from granu-
lar snow and when shaped by wind accumulation, have the
steep side always to leeward, in which respect the snow
behaves like drifted sand. In order to produce waves or
ripples, the wind must have a velocity sufficient to be thrown
into undulations by the irregularities of the surface over
which it blows. The most perfectly moulded forms are
naturally produced upon a relatively plane surface, such as
is realized on the inland-ice of Greenland the " imperial
highway " of Commander Peary.

FIG. 88. Barchans in snow, a, of deposition ; 6, of erosion (after Cornish) .

Apparently the direction of the greatest extension of the
sastrugi will depend upon the strength of the wind and upon
the amount of snow which is being transported, much as
has been found to be the case with drifted sand. 31 Thus, with
small amounts of snow and moderate winds, the character-
istic form of sastrugi is a short, scalloped ridge lying across
the wind direction and in form not unlike an ox-yoke

NOURISHMENT OF THE GREENLAND INLAND-ICE 157

something intermediate between a barchan and a transverse
ridge. Barchans of snow almost identical in form with sand
barchans, are produced apparently under like conditions,
the chief differences being that lighter winds suffice to ac-
complish the result with the less ponderous snow, and that
the resulting forms set quicker in the snow (see Fig. 88 a).

Cornish has realized the full importance of snow-blast
erosion in modifying the form of snow drifts. His barchans
of erosion, in plan resemble the barchans of deposition from
which they are derived, but unlike the depositional forms
their broader surface is concave upward instead of convex,
and their steeper face is toward the wind (see Fig. 88 6).

Some facts of importance which concern the density of the
snow are emphasized by Cornish, and apply with especial
force to the surface snow of inland-ice. It was found that
crusts upon the surface of snow do not necessarily imply
melting, but are produced in temperatures below the fusion
point. When the air temperatures at Winnipeg ranged
from 25 to 28 F. the snow surface over the river set so
hard that the moccasined heel did not dent it. Pieces of this
snow broken off and held up to the sunlight showed a
" mosaic of small translucent icy blocks cemented firmly by
opaque ice." The effect upon snow density of the radiation
from the surface and of pressure from the wind, were strik-
ingly brought out by a number of observations. Newly
fallen snow in Canada has a density of about 0.1. Over
the level surface about Winnipeg in the month of January
and at a temperature of 10 F., the snow was found to have
a density within the upper two feet of 0.38; while in the
woods at the same time and at the same depth, here without
a crust, its density was 0.19. Thus it is seen that the snow
in the woods is about twice as heavy as newly fallen snow,
but only about half as heavy as that which has been chased
about by the wind. At Glacier House in the Selkirks, where

158 CHARACTERISTICS OF EXISTING GLACIERS

the snow is shielded from the wind within a narrow valley,
experiments showed a density of 0.106 at the surface, whereas
at a depth of one foot below the surface the density was
0.195, and at a depth of four feet, 1.354. The middle value
being that of the snow in the woods at Winnipeg, it is seen
that the weight of an additional three feet of snow is neces-
sary in order to pack snow as tightly as is done by the wind
blowing over the prairie. After a time, as a result of this
treatment by the wind, an eight-inch snowfall dwindles by
packing in the woods to four inches, and over the open plain
to a two-inch layer. According to Gourdon, a cubic metre
of Antarctic snow may exceed 700 kilogrammes in weight. 32

In eroding a drift, the wind first attacks the softer surface
layer. This removed, the snow of the blast adheres less
to the surface of the drift, and in consequence abrades it
more vigorously. Thus, notches in the ridges, instead of
being mended by the detritus, are increased by it, and trans-
verse ridges are presently cut through, and we pass by stages
from an arrangement of ridges transverse to the wind to that
of longitudinal structures having the greatest extension
parallel to the wind. 33 These longitudinal sastrugi appear
to be the dominant ones, and from them the direction of
prevailing winds may be determined as has been already
proven in the Antarctic. On the Siberian tundras the
sastrugi are often the only guides of direction which the
natives have. 34

Source of the Snow in Cirrus Clouds. What has been
learned of the circulation of air above the continental ice of
Greenland, makes it extremely unlikely that any such exces-
sive alimentation upon the eastern margin through ordinary
snow fall, as has been advocated by von Drygalski, can
occur. 35 Such moisture-laden air as can, under normal con-
ditions, reach the interior plateau must descend from higher
levels in the anti-cyclone above the central boss, and be

NOURISHMENT OF THE GREENLAND INLAND-ICE 159

distributed by the outward flowing surface currents. From
such altitudes the moisture would probably be congealed in
the form of fine ice needles, such as are believed to exist in
cirrus clouds. This ice, in descending to the plateau, would
be adiabatically heated with, as a consequence, the melting
and vaporization of the ice crystals, which on reaching the
cold air layer directly enveloping the ice surface would be
congealed without passing through the cloud stage, thus
yielding the characteristic frost snow. This process will be
more fully treated in part III, after Antarctic ice masses
have been considered. Of greatest interest in this connec-
tion, is the observation of Nansen that while the sky was,
during the time of his crossing, in the main clear, those
clouds which were present were generally either cirrus clouds
or some combination of cirrus with cumulus and stratus
clouds. No cumulus clouds whatever were observed. In
tabular form his results are as follows:

FORM OF CLOUDS

No. OP DAYS

PER CENT.

Cirrus

23 1

Cirro-stratus

17[ 51

Cirro-cumulus

Cumulo-stratus

j
22

Stratus

As already stated, such snow as reaches the central area
must, it would seem, be derived from the cirrus clouds which
at higher levels move in toward the anti-cyclone and de-
scend in its center to become outward flowing surface cur-
rents over the " Great Ice." This subject will be more fully
developed in connection with the Antarctic continental gla-
cier (Chapter XVI).

160 CHARACTERISTICS OF EXISTING GLACIERS

REFERENCES

1 Robert Stein, " Suggestion of a Scientific Expedition to the center of
Greenland," Congres Intern, pour 1'Etude des Regions Polaires, Brussels,
1906, pp. 1-4 (separate).

2 In the light of later studies this may as satisfactorily be explained
through hardening by the wind.

3 Mohn u. Nansen, 1. c., p. 86.

4 Nansen, L c., vol. 1, p. 495.

5 Nansen, L c., vol. 2, p. 56.

6 Geogr. Jour., vol. 11, 1898, p. 228.

7 Mohn u. Nansen, L c., pp. 109-111.

8 Nansen, L c., vol. 2, p. 491. See also Peary, Geogr. Journ., I. c., p. 214.

9 Mohn u. Nansen, I. c., pp. 44^-45.

10 E. v. Drygalski, "Die Eisbewegung, ihre physikalischen Ursachen und
ihre geographischen Wirkungen," Pet. Mitt., vol. 44, 1898, pp. 55-64. See
also by the same author, "Gronland-Expedition," etc., pp. 533-539.

11 Nansen, 1. c., vol. 2, p. 496. Also Mohn and Nansen, I. c., pp. 44-47.

12 "Journeys in North Greenland," Geogr. Jour., vol. 11, 1898, pp. 233-
234. See also "Northward over the 'Great Ice,'" vol. 1, pp. Ixix-lxx.

"Lieut. A. Trolle, R. D. N., "The Danish Northeast Greenland Ex-
pedition," Scot. Geogr. Mag., vol. 25, 1909, pp. 57-70 (map and illustra-
tions).

14 Nansen, 1. c., vol. 2, p. 496.

15 Jour. Geol, vol. 3, 1895, pp. 578-579.

16 Professor v. Drygalski has shown that on the Great Karajak glacier
near the coast in central western Greenland, the temperature of the snow
and ice down to a depth of 60 feet or more undergoes a fall of tempera-
ture in response to the severity of the winter's cold, but in time this fall
in temperature lags behind the period of maximum cold. Below that
depth, however, it approximates in temperature to the zero of the centi-
grade scale. Temperatures of the snow measured just below the surface,
varied from - 11 to - 26 C. (E. von Drygalski, "Gronland-Expedition
der Gesellschaft fur Erdkunde zu Berlin," 1891-1893, vol. 1, 1897, pp.
470-472.)

17 Henry Rink, "Danish Greenland, Its People and Its Products,'*
London, 1877, p. 468.

18 Nansen, I. c., vol. 2, pp. 487-488, 496.

19 Geogr. Journ., I. c., pp. 214, 215. .

20 Mohn u. Nansen, I. c., p. 78.

21 Geogr. Jour., L c., pp. 233-234.

22 See Nordenskjdld ante, p. 114.

23 T. C. Chamberlin, "Glacial Studies in Greenland, VI.," Jour. Geol.,
vol. 3, 1895, pp. 580-581.

24 Salisbury, Jour. Geol., vol. 3, p. 886.

25 "Gronland-Expedition," etc., vol. 1, 1897. See also C. H. Ryder,

NOURISHMENT OF THE GREENLAND INLAND-ICE 161

Undersogelse af Gronlands vestkyst fra 72 till 74 35', 1886-1887, Med-
delelser om Gronland, heft 8, pi. xvii.

26 A. E. Nordenskjold, " Die Schlittenfahrt der Schwedischen Expedition
im nordostlichen Theile von Spitzbergen," 24 April-15 Juni, Pet. Mitt.,
vol. 19, 1873, pp. 450-453.

27 "Thus it will be seen that at no great distance from the east coast
the surface of dry snow begins, on which the sun has no other effect than
to form a thin crust of ice. The whole of the surface of the interior is
entirely the same." (Nansen, 1. c., vol. 2, p. 478.)

28 Robt. Stein, Congres international pour 1'etude des regions polaires,
Brussels, 1906, pp. 1-4 (separate).

29 National Antarctic Expedition, 19014. Album of photographs and
sketches (with brief descriptions, Ed.), London, 1908.

^Vaughan Cornish, "On Snow-waves and Snow-drifts in Canada,"
Geogr. Jour., vol. 20, 1902, pp. 137-175.

31 P. N. Tschirwinsky, "Schneedunen und Schneebarchane in ihrer
Beziehungen zu aolischen Schneeablagerungen im Allgemeinen," Zeitsch.
f. Gletscherk., vol. 2, 1907, pp. 103-112.

32 E. Gourdon, Expedition Antarctique Francaise, 1903-1905, Glaciologie,
Paris, 1908, p. 75.

33 Cornish, I. c., pp. 159-160.

34 Tschirwinsky, 1. c., p. 107.

35 "The east is to be regarded as the region of origin of snow, the west
as the terminal region of the Greenlandic glaciation." (E. von Dry-
galski, "Die Eisbewegung, ihre physikalischen Ursachen und ihre geo-
graphischen Wirkungen," Pet. Mitt., vol. 44, 1898, pp. 55-64.)

CHAPTER X

DEPLETION OF THE GREENLAND ICE FROM SURFACE

MELTING

Eastern and Western Slopes Compared. --Though it is
probably not true, as has been claimed by von Drygalski,
that the eastern border of the continent is the locus of
nourishment for the ice, it is almost certain that the losses
are much greater along the western margins. For this there
are several reasons. In the first place, the eastern base is
apparently characterized by lower temperatures. The cold
ocean current, which carries ice bergs and floes from the
Arctic Ocean southward in Baffin's Bay, follows the western
shore, while a warmer counter current flows northward along
the eastern or Greenland coast at least in its southern
stretches. Tarr thinks this current may reach as far as
Melville Bay. 1

Again, ablation or surface melting is to a large extent
dependent upon the quantity of rock debris which is blown
onto the ice surface from its margins. 2 In southern Green-
land, at least, the wider ribbon of exposed shore land upon
the western coast conspires with the prevailing westerly winds
to make a more effective marginal attack upon the anti-
cyclone of the continent. Nansen reports that he found on
the east coast none of the rock dust first described by
Nordenskjold as " cryoconite," though it extended inward
from the western coast as much as 30 kilometres. 3

162

DEPLETION OF THE GREENLAND ICE 163

Still further it is to be remembered that the ice of the
west margin is intersected by many deep fjords, which
communicating with the open sea, remove an enormous
quantity of ice in the form of bergs. Upon the eastern
coast, the pack-ice prevents the removal of bergs except
from the southern latitudes.

Effect of the Warm Season Within the Marginal Zones of
the Inland-ice. In winter the entire surface of the ice, and
the border of the land as well, are covered with an un-
broken layer of fine, dry snow. The suddenness of the
change to summer within the land zone outside the ice
front has been emphasized by Trolle. The temperature of
the snow upon the land in northeast Greenland rose gradu-
ally with the arrival of summer until the melting-point was
reached, and then in one day all the snow melted. " The
rivers were rushing along, flowers were budding forth, and
in the air the butterflies were fluttering. 7 ' 4

The snow upon the surface of the inland-ice, where studied
by von Drygalski within the western marginal zone, was
found to have temperatures which in the winter season were
normally lowest just below the surface, and which approxi-
mated to the zero of the centigrade scale at depths of gener-
ally a few metres only. In October with a sub-surface
temperature of 11 C. the zone of zero temperature was
reached at a depth of a little more than 2 metres. The
sub-surface temperature steadily lowered from this time as
the colder months came on, and the depth of zero tempera-
ture descended to below the limit of the experiments, which
was only a little more than 2 meters. The form of the
temperature curves in dependence upon depth showed
clearly, however, that at very moderate depths equalization
occurred. Late in March the lowest temperatures were
reached with 26.3 C. for the immediate sub-surface
temperature, and 9 C. for the temperature at depth of

164 CHARACTERISTICS OF EXISTING GLACIERS

-12.0"

-98"

Feb

-26.3*

/.$

2.m

2 metres. Warm weather
at the surface resulted in
a warm wave which de-
% scended through the snow,
following the colder one,
and so resulted in a maxi-
mum temperature not
immediately below the
surface, but at increasing
distances from it depend-
ing upon the duration of
the warmer air tempera-
tures at the surface.
Thus, a ten day foehn in
January raised the tem-
perature at a depth of 2.2
metres, by half a degree.
It required over two days-
for this rise in temperature
to proceed to a depth of 1
meter, and ten days for it
to reach the depth of 2
meters. Similar effects
are produced with the
coming of the more pro-
longed warm weather of
the summer season (see
Fig. 89). 5

When the surface zone
of the snow has reached
the fusing-point of snow,
melting begins rapidly.
Peary has drawn a graphic
picture of the effect of the

DEPLETION OF THE GREENLAND ICE

165

warm season upon the margins of the Greenland ice. Late
in the spring the warmth of the sun at midday softens the
surface first along the outermost borders of the ice, and
this, freezing at night, forms a light crust. Gradually this
crust extends up in the direction of the interior, and as the
season advances the surface of the marginal rim becomes
saturated with water. This zone of slush follows behind
the crust towards the interior in a continually widening
zone as the summer advances. Within the outermost zone

FIG. 90. Map showing the superglacial streams within the marginal zone of
the inland-ice of Greenland (after Nordenskjold).

the ice is so decomposed that pools come to occupy depres-
sions upon the surface, and streams cut deep gullies into the
ice. At the same time, the ice shows a more dirty appear-
ance through the concentration of the rock debris due to the
melting of surface layers of ice. By the 'end of the season,
pebbles, boulders, and moraines have in places made their
appearance on the surface, and the streams have left a sur-
face of almost impassable roughness. 6

Differential Surface Melting of the Ice. In his ascent of
the western margin of the ice near the latitude of Disco Bay,

166 CHARACTERISTICS OF EXISTING GLACIERS

Peary encountered lakes surrounded by morasses of water
saturated with snow. The ice within this zone is crevassed,
and down the fissures some of the surface streams disappear,
at times in a large water-fall, and again in a " mill " of its
own shaping. Baron Nordenskjold earlier observed almost
identically the same phenomena along the line of his route.
The intricate ramifications of the superglacial rivers and the
occupation of almost the entire remaining surface of the ice
by shallow ice wells and basins along his route are shown in

I c-t s'ck^r s be^cw

FIG. 91. Diagrams to show the effects on differential melting on the ice surface :
a, dust wells ; 6, basins ; c, glacier stars ; d, bagnoires.

Figs. 90 and 93. 7 These ice wells are in no wise restricted to
inland-ice, but are found on mountain glaciers as well, and
represent but one of a series of allied phenomena dependent
upon differential melting due to the presence of rock frag-
ments upon the ice.

The influence of heat radiated from rock particles which lie
upon the surface of snow or ice, has never been properly
recognized. During the construction of the Bergen Railway,
which was completed across Norway in December, 1909, it
was necessary each summer to clear away great banks of
snow lying upon the right of way, before the work of the sea-

DEPLETION OF THE GREENLAND ICE

167

son could be said to be begun. The labors of an army of
shovellers which had at first made somewhat ineffectual at-
tacks upon the drifts, were later replaced by the sun, a layer
of earth or sand having been spread over the snow surface.
In this way it was learned that drifts which would otherwise
have been but little diminished in size sank as much as six
feet in the course of a month.

Ice wells and allied phenomena due to differential melting
about rock particles on the ice surface were described by

= Layer warmed by

FIG. 92. Fragments of rock of different sizes to show their effect upon melting

on the ice surface.

Agassiz in his " System Glaciare." The particles of rock if
not contiguous upon the ice surface absorb the sun's rays
and cause excessive melting of the ice about and beneath
them. They thus sink down into the ice and form dust wells
(Fig. 91, a). The thin walls which separate those wells which
are close together, being now attacked by the warm air on
their sides instead of on the top only, they in their turn melt
away to form a small basin, which soon either wholly or in
part fills with water (Fig. 91, 6). Where in contact with their
neighbors and where of such thickness of accumulation as
not to be heated through by the sun's rays, these rock parti-
cles behave in quite a different manner and protect the ice

168 CHARACTERISTICS OF EXISTING GLACIERS

beneath them from the sun (note margins of wells and basins
in Fig. 91, a and 6). The same effect is brought .about if the
fragments are too large, for the thickness of surface layer of
rock which can be sensibly warmed by the sun's rays is quite
independent of the size of the fragment. Thus the familiar
ice tables developed especially upon mountain glaciers are
formed. Fig. 92 brings this out by showing the relation of
the warmed surface layer to the whole fragment (a) in a
dust well, (6) in a pebble that sinks slightly into the ice until
it reaches equilibrium, (c) in a slab of such size as to neither
facilitate nor retard surface melting, and (d) in a large pro-
tective slab of rock.

The basins which result from the dust wells induce still
other interesting structures. At night the water within these

FIG. 93. Section of the so-called " cryoconite holes" upon the surface of an ice
hummock (after Nordenskjold).

basins freezes in the form of needles which everywhere pro-
ject inward from the steep walls of the basin. After repeated
freezings the basins are often entirely closed by these needles
and thus form " glacier stars " (see Fig. 91, c). Elongated
basins have been given the name bagnoires (see Fig. 91, d).
From studies of such phenomena resulting from differential
melting as developed upon the Great Aletsch Glacier, we have
found that the segregation of the rock debris upon the bot-
tom of the basins later protects those areas after melting of
the general surface has drained them of their water. Thus

DEPLETION OF THE GREENLAND ICE 169

the familiar debris-covered ice cones come into existence and
further increase the irregularities of the ice surface. The
dust wells and basins which were described by Nordenskjold
over large areas covered the sides of steep hummocks in the
ice as well as its more level surfaces (see Fig. 93).

On his return from his attack upon the inland-ice near
Disco Bay, Peary travelled for seven hours through half-
frozen morasses alternating with hard blue ice honeycombed
with water cavities. Then the character of the ice com-
pletely changed, the slush and the water cavities disappeared,
and the entire surface was granular snow-ice, scored in every
direction with furrows, one to four feet deep, and two to ten
feet in width, with a little stream at the bottom of each. 8

Moats Between Rock and Ice Masses. Wherever the ice
sends an outlet down a valley, the edges of this ice shrink
away from the warmer rocks on either side, thus leaving
lateral canyons walled with ice on the one hand and with rock
upon the other. Down these canyons are the courses of
glacial streams. 9 An excellent example of such a lateral
stream is furnished by the Benedict glacier (Plate 25, A). 10

In most cases where nunataks project through the ice sur-
face, the absorption of the sun's rays by the rock melts back
the ice so as to leave a deep trench surrounding the island and
much resembling the moat about an ancient castle. Snow
drifted by the wind often bridges or partially fills the moat.
Upon nunataks forty miles within the border of the ice in
northeast Greenland the Danes found water running in the
ravines and disappearing under the ice at the margin of the
nunatak where it " formed the most fantastic ice-grottoes,
where the light was broken into all colors through the crystal
icicles." u

Such moats have been mentioned by nearly all explorers
upon the ice. It has been claimed by von Drygalski that
this phenomenon is characteristic of the west coast margin

170 . CHARACTERISTICS OF EXISTING GLACIERS

only, more ample nourishment upon the eastern coast making
the snow rise about the rock like a water meniscus. Ryder 12
and Jensen 13 have each figured such moats from the extreme
south of Greenland. By Jensen's party these moats were
made use of for camping-places. Peary, however, has shown
that the moats upon the west coast are often largely filled
with snow. 14 Stein mentions this as a common feature after
snow storms, 15 and Chamberlin 16 asserts that wherever the
motion of the ice is considerable the trench does not appear,
but the ice impinges forcibly upon the base of the nunatak.

Englacial and Subglacial Drainage of the Inland-ice. In
addition to the superglacial streams which are so much in evi-
dence, others which are englacial run beneath the surface of
the ice, as has often been discerned by putting the ear close to
the ice surface . Nordensk j old reports one instance where water
spouted up from the surface mixed with a good deal of air and
spray. 17 Salisbury also has mentioned a huge spring upon
the surface of the ice in north Greenland that shot up to a
distance of not less than ten feet above the bottom of the
basin from which it issued. Owing to the fact that near the
margin of the ice its surface is much crevassed, comparatively
little water can continue to the border in surface streams.
Salisbury mentions an instance where an englacial stream
with a diameter of about five feet issued from the vertical face
which formed the ice front. Most of the water flowing upon
the surface descended, however, to the bottom, and issued
largely below the surface of the fluvio-glacial materials. It
is, he says, a rare exception to find a visible stream issuing
from beneath the ice at its margin. In most cases, the water
undoubtedly comes out in quantities, though beneath the sur-
face of the outwash apron, as could be detected by the ear. 18
Poary has observed that a greater abundance of water issues
from beneath the ice-cap in extreme northeastern than in
northwestern Greenland. 19

PLATE 25.

A. Lateral glacial stream flowing between ice and rock, Benedict glacier tongue

(after Peary).

B The ice-dammed lake Argentine in Patagonia (after Sir Martin Con way).

DEPLETION OF THE GREENLAND ICE

171

The Marginal Lakes. Wherever the ice has withdrawn
from the rock surface, and where ice drainage permits of it,
small lakes marginal to the inland-ice have come into exist-
ence. Special interest attaches, however, to those bodies
of water which are impounded by the ice itself along its
margin, because of the light which is thrown upon the ori-
gin of somewhat similar bodies of water about the great
continental glaciers of Europe and North America during
late Pleistocene times. Attention was called to such ice-

FIG. 94. Map showing the margin of the Frederikshaab ice apron extending from
the inland-ice of Greenland and showing the position of ice-dammed marginal
lakes (after Jensen).

dammed lakes situated upon the margin of the Frederiks-
haab tongue of the inland-ice by the Jensen, Kornerup, and
Groth expedition of 1878. A map of this region was pub-
lished by Jensen (see Fig. 94). 20 Here the lakes filled with
water from the melting of the glacier by which their outlets
are blocked, stand at different levels. The Tasersuak on the
south, standing at a level of 940 feet above the sea, is blocked
by ice at both ends and is covered by bergs which are calved

172 CHARACTERISTICS OF EXISTING GLACIERS

from the ice cliffs. This lake drains through a canal upon the
ice to a much smaller lake standing at a level of 640 feet, and
thence through a small river to the head of the Tiningerf jord.
To the northward of the apron of ice another long fjord is
blocked by a T-shaped extension of ice into its central por-
tion. Thus there result two fresh water lakes standing at
different levels, the lower one, like the Tasersuak, with ice cliffs
at both ends, and the other blocked at one end only by the
ice. A slight retreat of the inland-ice of this district would
retire the T-shaped extension of the glacier, and the two
smaller lakes would thus become united into one at the level
of the lower. A still further withdrawal of the Frederiks-
haab glacier tongue would open an outlet for this lake to the

Sea Level.

FIG. 95. Diagram showing arrangement of shore lines from marginal lakes to the
northward of the Frederikshaab ice tongue, if its front should retire past the
outlet of the lower lake.

sea at a still lower level. Souvenirs of these events would be
left in a series of parallel shore lines ascending in step-like suc-
cession to the head of the fjord (see Fig. 95). Suess has used
this illustration to explain the vexed problem of the seter,
the abandoned shore lines of Norway, which he claims have
this peculiarity of arrangement. 21

The famous "parallel roads " of glens Roy, Glaster, and
Spean in the Scottish highlands, which have in similar
manner vexed geologists, but which were finally given a satis-
factory explanation by Jamieson, 22 find here a living model.
Still later a nearly identical example from Pleistocene times
has been supplied from the Green Mountains to the eastward
of Lake Champlain. 23

About the Cornell tongue of the inland-ice of Greenland

DEPLETION OF THE GREENLAND ICE 173

are many marginal lakes situated where the border drainage
has been blocked by the glacier itself. These lakes have
been described by Tarr, who says: 24

In its passage down the valley, between the ice and the land,
the marginal stream finally enters the sea. During its passage it
now and then encounters tongues of ice, and for a distance flows
along them, and finally beneath them, where the glacier edge rests
against a moraine, or the rock of the land. Again it falls over a
rock ledge as a cascade, or even a grand waterfall ; and every here
and there it is dammed to form a marginal lake. Dozens of these,
great and small, were seen along the margin ; and they varied in
size from tiny pools to ponds half a mile in length, and 200 to 300
yards in width.

Since the water of the marginal streams is everywhere milky
with sediment, these lakes are receiving quantities of muddy de-
posits, and in them tiny deltas are being built. Where the lake
waters bathed the ice front little icebergs are coming off, in exactly
the same way as in the fjord at the glacier front, and these are
bearing out into the lake large rock fragments which are being
strewn over the bottom or 'on the shores. Also at the base of the
cliffs, as well as on some of the deltas formed by rapidly flowing
streams, pebbles and boulders are being mixed with the clay.

Nearly every lake shows signs of alteration in level resulting
from the change in outflow either to some point beneath the ice,
when the lake may be entirely drained, or to some lower outlet
for the lake opened by a change in the ice front, or by the down
cutting of the stream bed where it is eating its way through a
morainic dam. The different elevations are plainly evident from
the absence of lichens on the rocks, the clay clinging to the rocky
shores, and the beach terraces along the old shore lines. In one
case, at the western end of Mount Schurman, a lake of this type,
with a depth of at least 100 feet has recently been drained. Where
these extinct lake beds exist one sees revealed an expanse of muddy
bottom with scattered blocks of rock.

In plate 26, A and B are represented after Tarr, in the one
case, one of the marginal lakes, and in the other, the forma-
tion of a delta under the conditions described. From the

174 CHARACTERISTICS OF EXISTING GLACIERS

Karajak district on the northern side of the Upper Nugsuak
Peninsula, 25 von Drygalski has described in addition to the
usual rock basin lakes left by the withdrawal of the ice front,
a true ice-dammed lake which appears upon his map as the
Randsee. 2 *

No one of the marginal lakes thus far described furnishes
a parallel to the interesting Pleistocene glacial lakes of the
Laurentian basin of North America, since these developed for
the most part upon a surface of relatively mild relief, and the
shores not formed by the glacier itself were generally mo-
raines registering an earlier position during the retirement of
the ice front. Perhaps an existing example comes nearest to
being realized in connection with those glaciers which descend
the eastern slopes of the Andes and enter the great lakes im-
pounded behind moraines of an earlier extension of the same
ice tongues. 27 In these cases the ice fronts of the glaciers are
cut back into cliffs from which are derived the bergs that
float upon the surface. The ice cliff and some of the bergs
of Lake Argentine are shown in plate 25, B. According to
Moreno, Lake Tyndall is bounded on the west by true inland-
ice, 28 the remnant of the larger sheet of Pleistocene times.

Ice Dams in Extraglacial Drainage. In north Greenland
outside the ice front, the brooks sometimes offer a striking
example of ice obstructions forming by irrigation. This is
often the case where their beds are wide and are covered
with boulders. The water generally continues to run beneath
the stones for a great part of the winter. Later, however,
its outlets may freeze up, whereupon the water rises, inundat-
ing the stones and covering them with an ice crust. Through
successive obstruction, overflowing, and freezing of these
streams, the ice dam which results may attain to such a thick-
ness that it is still to be found at these places late in the
summer when the ice and snow have elsewhere disappeared
from the low land. 29 The significance of such dams as

PLATE 26.

A. Ice-dammed lakes on the margin of the Cornell tongue of the inland-ice

(after Tarr).

B. Delta in one of the marginal lakes to the Cornell glacier tongue (after Tarr).

DEPLETION OF THE GREENLAND ICE 175

obstructions during a readvance of the ice front may well
be considerable.

Submarine Wells in Fjord Heads. Rink states that the
sea flowing into the fjord in front of the glacier outlet which
ends below the water level, is kept in almost continual motion
by eddies not unlike those which are seen where springs issue
from the bottom of a shallow lake. Such areas upon the
surface of the fjord may generally be recognized by the flocks
of sea birds which circle above them and now and then dive
for food. 30 The existence of such fresh water streams as this
implies may also be inferred from the strong seaward current
that prevails in the fjords and which is so effective in clearing
them of bergs. Such a whirlpool of fresh water or " submar-
ine well " was observed by Rink in the Kvanersokfjord (lat.
62 N.) which was over 100 yards in diameter. The kitti-
wakes flocked over the spot, and the water was muddy,
although no brooks were observed along neighboring shores.
This well Rink believed, from reports furnished by the na-
tives, to be much smaller than the similar ones in some other
fjords.

According to Rink 31 the lateral lake which borders the in-
land-ice of Greenland in one of the branches of the Godthaab-
fjord-Kangersunek suffered changes of level just when the
submarine wells before the ice cliff in the fjord showed marked
changes in volume. Thus, whenever the water of the lake
suddenly subsides, the submarine wells from the bottom of
the fjord burst out with violence. On the other hand, when
the water in the lake is rising, the wells are relatively quiet.
These sudden discharges of the water from lateral lakes,
save only that their outlet is submarine, seem thus to be
in every way analogous to the spasmodic discharges of the
famous Marjelensee upon the margin of the Great Aletsch
Glacier in Switzerland. When, as occasionally happens,
this lake empties through the opening of a passage beneath

176 CHARACTERISTICS OF EXISTING GLACIERS

the glacier, the villages which are situated miles below in the
valley are suddenly inundated with water.

REFERENCES

1 R. S. Tarr, "Difference in the Climate of the Greenland and Ameri-
can Sides of Davis' and Baffin's Bay," Am. Jour. Sci., vol. 3, 1897, pp.
315-320.

2 An interesting practical illustration of the effectiveness of such debris
as a melting agent has been furnished during the construction of the
Bergen Railway in Norway, which was completed in December, 1909. A
prime factor in the work was a means of clearing the snow so as to pro-
long the summer season. For this purpose covering the snow surface
with fine dirt proved more effective than a corps of shovellers, the sun in
this case performing the work.

3 Mohn und Nansen, 1. c., p. 90.

4 Trolle, 1. c., p. 66.

6 E. von Drygalski, "Gronland-Expedition," L c., pp. 460-466.

6 Peary, Geogr. Jour., L c., p. 218. See also Nordenskjold, " Gronland"
(German ed.), pp. 125-138.

7 A. E. Nordenskjold, "Gronland," pp. 197-204, map 3.

8 Jour. Am. Geogr. Soc., I. c., p. 282.

9 Peary, Jour. Am. Geogr. Soc., I. c., p. 286.

10 R. E. Peary, "North Polar Exploration, Field Work of the Peary
Arctic Club," 1898-02, Ann. Rept. Board of Regents Smith. Inst. for 1903,
1904, p. 517. Cf . also the almost identical (if smaller scale) effects for Ice-
landic and Norwegian ice-caps. This volume, ante, p. 101.

11 Trolle, Scot. Geogr. Mag., vol. 25, 1909, pp. 65-66.

12 C. H. Ryder, " Undersogelse af Gronlands Vestkyst fra 72 till 74 35',
1886-1887," Med. om Gronland, Heft 8, pi. xiii.

13 J. A. D. Jensen, " Expeditionen till Syd. Gronland," 1878, ibid., Heft

1, pi. iv.

14 Peary, Geogr. Jour., I.e., p. 217.

15 Stein, I. c.

16 Chamberlin, Jour. Geol, vol. 3, 1895, pp. 567-568.

17 A. E. Nordenskjold, "Gronland," p. 137.

18 Salisbury, L c., pp. 806-7.

19 Peary, Geogr. Jour., I. c., p. 224.

20 Meddelelser om Gronland, Heft 1. This map has been many times
copied, best by Nordenskjold in his "Gronland" on p. 161.

21 Beside the Jakobshavn ice tongue, there is another lake confined in
like manner to the Tasersuak. (Ed. Suess, "The Face of the Earth," vol.

2, pp. 346-363.)

22 Thomas T. Jamieson, On the parallel roads of Glen Roy and their
place in the history of the glacial period. Quart. Jour. Geol. Soc., vol. 19,
1863, pp. 235-259.

23 H. E. Merwin, " Some late Wisconsin and Post- Wisconsin Shore-lines

DEPLETION OF THE GREENLAND ICE 177

of Northwestern Vermont," Rept. Vermont State Geologist, 1907-08, pp.
113-137, pi. 21.

24 R. S. Tarr, "The Margin of the Cornell Glacier," Am. GeoL, vol. 20,
1897, pp. 150-151.

25 Tn e Cornell tongue is situated upon the southern side of the same
Peninsula."

26 E. von Drygalski, "Gronland-Expedition," vol. 1, pp. 61-63, map 2.

27 Francisco P. Moreno, "Explorations in Patagonia," Geogr. Jour., vol.
14, 1899, pp. 253-256. Also Hans Steffen, "The Patagonian Cordillera
and its Main Rivers between 41 and 48 South latitude," ibid., vol. 16,
1900, pp. 203-206. Also Sir Martin Conway, "Aconcagua and Tierra del
Fuego," London, 1902, pp. 134-135.

28 See also P. D. Quensel, "On the Influence of the Ice Age on the Con-
tinental watershed of Patagonia," Bull. GeoL Inst. Upsala, vol. 9, 1910,
pp. 60-92, 2 maps. See also, R. Hauthal, "Der Bismarck-Gletscher, ein
vorriickender Gletscher in der patagonischen Cordillere," Zeit. f. Gletscherk.,
vol. 5, 1910, pp. 133-143, figs. 1-7.

29 Henry Rink, "Danish Greenland, Its People and its Products," Lon-
don, 1877, p. 366.

30 Rink, /. c., pp. 50, 360-363.

31 Rink, I. c.

CHAPTER XI

DISCHARGE OF BERGS FROM THE ICE FRONT

The Ice Cliff at Fjord Heads. Wherever the inland-ice
reaches the sea in the fjord heads, and where it comes directly
to the sea in broad fronts, as it does near Melville Bay, at
Jokull Bay, and on the north side of Northeast Foreland, it is
here attacked directly by the waves and is further under-
mined through melting in the water. The crevassing of its-
surface over the generally steep descents to the fjords, in a
large measure facilitates the attack of the water upon the
ice by offering planes of weakness similar to the joint planes
in rock cliffs attacked by the sea on headlands. The fjords,
though often quite narrow, are generally of great depth, so
that, although the ice cliff often rises to a height of several
hundred feet, its base probably rests upon the bottom of
the fjord. To this a possible exception has been noted for
the great Karajak glacier, of which a relatively flat front sec-
tion may be assumed to be the surface of a floating portion
(see Fig. 96). 1 To this interesting example of a floating gla-
cier outlet in connection with the inland-ice of the northern
hemisphere, we may recall the probably floating front of the
Turner glacier, a dendritic glacier of the tide-water type in
Alaska. For its type this example is apparently unique. 2

Manner of Birth of Bergs from Studies in Alaska. The
birth of bergs from the parent glacier has been often described

178

DISCHARGE OF BERGS FROM THE ICE FRONT 179

by travellers, and the superlatives of the language have been
drawn upon to express the grandeur and beauty of the ob-
served phenomena. Simple as the process may appear to the
casual tourist who makes the usual summer excursion to
Alaska, it is not free from serious difficulties, and has given

inn

Kilometers.

FIG. 96. Sections from the inland-ice through the Great and Little Karajak out-
lets to the Karajak fjord (after Von Drygalski).

rise to conflicting views among specialists. The water in
front of the ice cliff is generally so muddy, and the danger of
approaching the ice front so great, that exact data are neces-
sarily difficult to obtain. The smaller bergs composed of
white ice, which are seen to fall into the water from the cliffs
at almost all hours, offer no difficulties of explanation, but
they are likewise without great significance as concerns the
manner of formation of those great floating masses of ice
which are carried far to sea and scattered over wide areas of
the ocean before their final dissolution in the warmer south-
ern waters. It is, however, interesting to find that the
overriding of the lower layers of the ice by the upper
greatly facilitates the separation of this type of small ice-
berg. Engell, who has studied the Greenland icebergs,
shows that while this is true of icebergs formed in fjord
heads, it plays no part with those calved from the sides of
glacier outlets where ice dammed lakes (see below) make
iceberg formation possible. 3
The larger bergs, instead of falling from the cliffs, suddenly

180 CHARACTERISTICS OF EXISTING GLACIERS

rise out of the water as ice-islands, often several hundred feet
in front of the ice cliff. A wholly satisfactory solution of the
problem of their birth involves a nice quantitative adjust-
ment of several factors, all of which are undoubtedly more or
less concerned. On the one hand, there is wave action which
is effective especially near the water level and has a direct
range of action extending from a distance below the surface
equal to the length of a storm wave in the fjord, and to a
height above the quiet level equal to the height of the wave's
dash. If there were no melting in the water, and if the lower
layers of the glacier moved forward as rapidly as the upper,
the tendency would undoubtedly be to develop an erosion
profile in every way like that of a rock-cut terrace upon the
sea shore. With emphasis upon this element in the problem
Russell has assumed that the ice cliff in the fjord is prolonged
outward beneath the water as an ice foot which thins gradu-
ally toward the toe. Upon this hypothesis the bergs which
rise from the water are born from the foot where the increas-
ing buoyancy of the outer portion overcomes the cohesive
strength of the material at the surface where rupture occurs.
This view accounts particularly well for those bergs which
rise from the water far in advance of the cliff (see Fig. 97). 4

FIG. 97. Origin of bergs as a result especially of wave erosion (after Russell).

Laying stress rather upon melting in the water and upon
the rapid forward movement of the upper layers of ice near
the glacier margin, Reid has arrived at a wholly different
conclusion concerning the origin of larger bergs: 5

DISCHARGE OF BERGS FROM THE ICE FRONT 181

The more rapid motion of the upper part would result in its
projection beyond the lower part, and this would become greater
and greater until its weight was sufficient in itself to break it off.
The extent of the projection before a break would occur, depends
evidently upon the strength of the ice. . . . That the ice for
several hundred feet below the surface does not in general project
farther than that above is evident from the fact that I have fre-
quently seen large masses, extending to the very top of the ice
front, shear off and sink vertically into the water, disappear for
some seconds, and then rise again almost to their original height
before turning over. If there were any projection within 300 feet
of the surface, this mass would have struck it and been overturned
so that it could not have arisen vertically out of the water.

Reid thinks there are three ways in which bergs come into
existence at the end of a glacier : -

(a) A piece may break off and fall over this is the usual
way with small pinnacles ; (6) a piece may shear off and sink into
the water this is the usual way with the larger masses ; or, again,
(c) ice may become detached under water and rise to the surface.

The supposed successive forms of the ice front, according
to Reid, are shown in Fig. 98.

It is easy to see that Russell's and Reid's views might
each apply in
special cases
dependent: (a)
upon the nar-
rowness or the
sinuosities of
the fjord,
which would

, . , FIG. 98. Supposed successive forms of a tide-water glacier

determine the front (after Reid)>

reach of the

waves; (b) upon the steepness of the slope back of the ice

cliff, which would regulate the different velocities of surface

182 CHARACTERISTICS OF EXISTING GLACIERS

and bottom layers of ice, and determine the measure of
crevassing; (c) upon the irregularities in the floor of the ice
tongue, which would largely fix the amount of shearing and
overthrusting; (d) upon the presence or absence of warm
ocean currents, which would regulate the rate of melting of
the ice by the fjord water; and (e) upon the freezing of the
water surface, 6 which must put a bar upon the action of the
waves during the colder period.

Studies of Bergs Born of the Inland-ice of Greenland.
Though ice bergs are discharged from the inland-ice through-
out practically the entire extent of the coast line of Green-
land wherever inland-ice reaches the sea, yet the great
bergs which push out into the broad Altantic arise either

FIG. 99. A large berg floating in Melville Bay and surrounded by sea-ice.

on the west coast between Disco Bay and Smith Sound, or
on the east coast south of the parallel of 68. To the north
of this latitude the bergs are firmly held in the heavy pack-
ice, while the bergs of southwest Greenland form for the most
part in such narrow fjords that they are too small to travel
far before their final dissolution.

DISCHARGE OF BERGS FROM THE ICE FRONT 183

The size of the Greenland ice bergs has probably been much
1 overestimated. Of 87 measurements made by von Drygal-
ski on the large bergs calved in the Great Karajak fjord, the
highest reached 137 metres above the water, or about 445
feet. This mass of ice was, however, against the glacier
front, and probably rested on the bottom. None of the others
measured were much above 100 metres high or about 325
feet. 7 The berg shown in Fig. 99, photographed by an
earlier explorer in Melville Bay, measured 250 feet in height.

During fourteen months spent in the immediate vicinity
of the steep front of the Great Karajak ice tongue, von
Drygalski carried out extensive studies upon the calving
of bergs, and has distinguished three classes. Those of the
third class form almost constantly, and consist of larger or
smaller fragments which separate along the crevasses and
fall into the sea. Only twice were calvings of the second
class observed, namely, in late October and in early Novem-
ber. Of one of these von Drygalski says :

I heard a thundering noise, but at first neither I nor the Green-
landers who were with me saw anything. Suddenly a great dis-
tance away from the margin of the glacier, an iceberg emerged from
the sea, rose out of the water, though not to the height of the cliff,
and then moved away accompanied by a continuous loud tumult
and by a rise in the level of the water, through the agency of which
it moved away from the cliff quite rapidly. It did not come from
the cliff, but certainly emerged from below. The Greenlanders,
whom I afterwards questioned about it, gave me the same impres-
sion. . . . The margin of the glacier was unchanged.

Here it was noticed that the berg was long, though not as
high as the ice cliff which terminated the glacier. It is the
opinion of von Drygalski that bergs of this class come from
the lowest layers of the glacier. Because of the sea-ice
which in winter forms in front of the glacier, the ice cliff
is at that time not cut away so fast, and it was, in fact,

184 CHARACTERISTICS OF EXISTING GLACIERS

observed in the winter farther out than during the summer.
This explanation in the main is in agreement with that of
Russell.

Bergs of von Drygalski's first class, which are the most
massive of all, separate from the entire thickness of the ice
front. Two such bergs were observed in process of calving
by von Drygalski and other members of his party. The
same loud sound which had been heard at the birth of bergs
of the second class accompanied the birth of those in the first
class, but the movement of separation from the glacier was
visible at the same instant. A portion of the cliff front was
seen to separate from the cliff, being thereby thrown some-
what out of equilibrium and started in a pendular vibration
which produced great waves in the fjord and increased
slightly its distance from the newly formed ice cliff. It
was here observed that the main pinnacle of the berg slightly
exceeded in height the highest pinnacles of the new glacier
rim. This, it will be remembered, is in contrast with the
bergs of the second class which did not reach to the height
of the cliff. Bergs of the first class usually regain their
equilibrium after rhythmic oscillations, and float away in an
upright position. The bergs of the second class often turn
over, displaying the beautiful blue color of the lower layers.
Studies confirmatory of those of von Drygalski have been
recently made by Engell, 8 and Salisbury's two types of
Greenland icebergs seem to correspond with von Drygalski's
bergs of the first and second classes. 9

The water waves which are sent out to the shores at the
birth of a great iceberg extend 50 kilometres or more within
the fjord, driving the smaller floating bergs together and
thus assisting in their fragmentation and consequent dis-
solution. The calving of bergs of the first class von Dry-
galski believes occurs where the depth of the fjord has so far
increased that the ice begins to leave the bottom and assume

DISCHARGE OF BERGS FROM THE ICE FRONT 185

a swimming attitude. The buoyancy of the water is, he
believes, thus the true cause of the separation of the bergs.

Depths which are four to five times as great as the thickness
of the inland-ice above the sea level, are not measured in Green-
land in front of attached ice masses, because the latter become
in that case broken up into ice bergs. 10

This view gains strength from Salisbury's studies of the
glaciers ending in Melville Bay and apparently floated for
a very short distance back from their fronts and generally
in the middle only. 11

REFERENCES

1 E. von Drygalski, "Gronland-Expedition," vol. 1, pi. 43. See also
R. D. Salisbury, "The Greenland Expedition of 1895," Jour. Geol, vol. 3,
1895, p. 885.

2 R. S. Tarr and B. S. Butler, "The Yakutat Bay Region, Alaska,
physiography and Glacial Geology," Prof. Pap. No. 64, U. S. Geol. Sur.,
1909, pp. 39-40, pi. 10-a.

3 M. C. Engell, " Ueber die Entstehung der Eisberge," Zeit. f. Gletscherk.,
vol. 5, 1910, pp. 122-132.

4 1. C. Russell, "An Expedition to Mt. St. Elias, Alaska," Nat. Geogr.
Mag., vol. 3, 1891, pp. 101-102, fig. 1.

5 H. F. Reid, "Studies of Muir Glacier, Alaska," ibid., vol. 4, 1892,
pp. 47-48.

6 R. S. Tarr, "The Arctic Ice as a Geological Agent," Am. Jour. Sci.,
vol. 3, 1897, p. 224.

7 E. von Drygalski, "Gronland-Expedition," etc., I. c., pp. 367-404.

8 "Ueber die Entstehung der Eisberge," Zeit. f. Gletscherk., vol. 6, 1910,
pp. 122-132.

9 Jour. Geol., vol. 3, pp. 892-897.

10 E. von Drygalski, I. c., p. 404.

11 Salisbury, Jour. Geol, vol. 3, 1895, pp. 885-886.

PART III

ANTARCTIC GLACIERS

CHAPTER XII
THE ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE

General Uniformity of Conditions in Contrast with the
North Polar Region. The essentially reciprocal physio-
graphical developments about the earth's two polar regions
are responsible for a striking contrast in their physical,
and especially in their glacial conditions. In the north a
deep polar sea is largely encircled by a rim of land masses,
interrupted, however, in one rather broad area by the north-
ern Atlantic ocean. Nearly opposite this interruption the
Pacific pushes a great bay so far to the northward as just
to pierce the land girdle. The ridge of the Aleutian arc
farther to the south imposes a bar to the movement of ocean
currents, and makes the break at this point a less important
one than it would at first appear.

The widely different specific heats of land and water,
the irregularities of the land surfaces, and especially the
large transfer of heat through the medium of northwardly
and southwardly directed ocean currents, together bring
about a great diversity of climatic conditions within the
northern polar regions. Along the same parallel of latitude

186

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 187

the widest differences of temperature and precipitation are
to be encountered.

Within the south polar region, on the contrary, the great
continental plateau, centred as it is so nearly over the pole
and having its borders for long distances so nearly in corre-
spondence with the Antarctic Circle, the surrounding ocean
permits of a relatively free circulation of oceanic waters
and of air currents. The result is a greater uniformity and
a symmetry in distribution of the principal climatic constants
with regard to the south pole as a centre. Here the iso-
therms more nearly follow the parallels of latitude, and, there
being a much smaller transfer of heat by ocean currents from
tropical regions, the climate is far more severe than within
the Arctic regions. For this reason the surface of the sea
freezes in considerably lower latitudes, so that the Antarctic
continent is encircled by a broad zone of pack ice which offers
the most serious bar in the way of those who would explore
it.

The uniformity of climatic conditions within the Antarctic
we express when we say that its climate is oceanic. To
fully understand the severity of this climate it is necessary
to emphasize a vital difference between the glaciation of
southern and of northern polar regions. Throughout the
Arctic regions, with a single noteworthy exception of the
Franz Josef archipelago, the snow-ice masses are all smaller
than the land areas upon which they lie. 1 Within the Ant-
arctic, on the contrary, the reverse is the case, and the
snow-ice masses quite generally cover all the land surface and
push out also upon the sea. Barring the peninsula of West
Antarctica, which sends a narrow tongue northward two
degrees or more beyond the Antarctic Circle, land has been
seen exposed only in the high Admiralty and Royal Society
ranges in Victoria Land, and in a few isolated volcanic peaks,
such as Erebus and Terror on the Ross Sea, and the Gauss-

188 CHARACTERISTICS OF EXISTING GLACIERS

berg of Kaiser Wilhelm II Land. Elsewhere the white snow
surface, variously moulded near its margin, is all that meets
the eye at the border of the continent.

An oceanic climate is possessed by bodies of land which
are surrounded by the sea and are so small that climatic
conditions are dominated by the sea rather than by the land.
Yet however small the land surface may be, since it is ex-
posed to the sun's rays, it is warmed and cooled more rapidly
than is the sea, and in consequence exerts a counteracting
influence upon the climate in the direction of a greater
changeability. Within the Antarctic, however, where the
surface is almost entirely snow-covered, the earth is shielded
from solar radiation, and no such influence is exerted. This
is an important cause of the difference in climate between
the Arctic and Antarctic regions.

Antarctic Temperatures. Nowhere is the uniformity of
conditions within the Antarctic region more strikingly
exemplified than in the temperatures which prevail and in
the small range which separates the winter from the summer
temperatures. Thus we find on the margin of the continent
at or near the level of the sea and in latitudes near the
Antarctic Circle an average summer 2 temperature which is
colder than Nansen encountered in the Arctic ice pack
within five degrees of the North Pole. Both the average
and the extreme winter 3 temperatures are on the other
hand as surprising by reason of their moderate values.
As illustrating the oceanic climate of Antarctica, we have
only to state that the extremes of cold encountered in the
Antarctic regions are equalled or exceeded by the tempera-
tures registered at stations south of the 50th parallel in
North America. The recent Antarctic expeditions have at
last supplied us with reliable data at several widely separated
points and for periods of a year or more. These data we
have collected and set forth in the following table:

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 189

ANTARCTIC TEMPERATURES IN FAHRENHEIT DEGREES

t |

Station

*o
o

<D^f

Authority

3 O

c3 S "

^ Q) "

a &

i-^l a2

jT-s

i-l 02

|||

III

Snow Hill Island,

West Antarctica .

643(X

57 W.

28.13

-4

10.76

-42.3

O. Nordenskiold *

"Gauss" in the ice

pack 50 miles off

Kaiser Wilhelm II

Land

90 E.

-29.5

v. Drygalski s

"Belgica" in the ice-

pack off West Ant-

arctica ....

70-71 36'

85-103 W.

29.3

1.8

9.6

-45.6

Arctowski 8

Cape Adare, Victoria

Land

71 \/ Q

170 E.

30.4

-11.3

7.05

-43.1

Bernacchi 7

Cape Armitage, Vic-

toria Land .

778/4

167 E.

-15.15

-21

Shackleton 8

Cape Armitage, Vic-

toria Land .

77%

167 E.

-13.17

-17

Shackleton 8

To the S.E. of White

Island on ice bar-

rier near Cape Ar-

mitage ....

-67.

Scott

(in September)

Thus we see that the average summer temperature upon
the borders of the Antarctic continent is from two to four
degrees below the freezing-point of water, or about the same
as the winter temperature of Southern Scandinavia. Pass-
ing northward from the Antarctic Circle and beyond the
margins of the continent, the rise in temperature is rapid.
Thus Bruce 's temperature record, taken in the Weddell Sea,
gave 7 F. as the lowest temperature reached, while the aver-
age summer temperature was near 31.4 F. and the average
winter temperature 13.7 F. 10 In the South Orkney Islands,
only 3j farther north than the Snow Hill station of West
Antarctica, the average winter temperature is higher by
14 F. 11

The above described Antarctic temperatures measured

190 CHARACTERISTICS OF EXISTING GLACIERS

near sea level and on the margin of the continent are, how-
ever, quite different from those which are encountered upon
the high ice plateau. Data from these levels are naturally
only available for brief periods during the summer season.
On his trip westward from Cape Armitage over the snow
plateau, Scott found that for fifty days the temperature
fell almost nightly to 40 F. and seldom rose during the
day above 25 F. 12 Shackleton's southern party even in
the height of summer nowhere upon the plateau encountered
a temperature above F., and temperatures between 35
and 40 F. were often registered. 13

Geographical Results of Exploration. --The wide zone of
sea-ice which surrounds the Antarctic continent has been
an effectual barrier to navigation of Antarctic waters. If
to this we add the remoteness of the region from centres of
civilization, and the lack of any lively commercial interest,
such as the supposed northwest and northeast passages
to the orient in the northern hemisphere, the tardiness of
exploration in the south polar regions finds a sufficient
explanation. The first important voyage of discovery
in that region, that of Captain James Cook in the years
1770 to 1774, was undertaken to solve the problem of the
supposed southern continent, the Terra Australis Incognita.
Cook largely circumnavigated the globe in the latitude of
50 south or more, and for a distance of 115 of longitude kept
south of 60 latitude. Three times he crossed the Antarctic
Circle, and at one point attained the high latitude of 71 10',
but without discovering the supposed continent. 14

It was the importance of the sealing industry in the South
seas which some sixty years after Cook's voyages furnished
the impetus to the second great period of Antarctic dis-
covery, that of 1838-1841. Three expensive expeditions
fitted out in the United States, France, and England were
commanded by Wilkes, 15 D'Urville, 16 and Ross 17 respectively.

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 191

It is the best commentary upon the difficulties of South
Polar exploration that while all these expeditions discovered
the Antarctic continent, no one of them succeeded in effect-
ing a landing. With the revival of interest in Antarctica
which came after another sixty years had elapsed, during
which time strong steam vessels had replaced the earlier
sailing ships, Borchgrevink in 1898 wintered at Cape
Adare in Victoria Land, and could claim that he was the first
to set foot upon the Antarctic continent. 18

The recent period of Antarctic exploration began, however,
with the " Belgica " expedition of 1897-1899, which was
assisted by the Belgian government and was commanded
by Lieutenant Adrian de Gerlache, 19 though it was soon
followed in 1898-1900 by the British Antarctic expedition
under Borchgrevink. These two expeditions greatly stim-
ulated an interest in South Polar exploration, and in 1902
three national expeditions wintered in the Antarctic an
English under the command of Captain R. F. Scott, 20 a Ger-
man under command of Professor Erich v. Drygalski, 21 and
a Swedish under Dr. Otto Nordenskiold. 22 A Scotch explor-
ing expedition commanded by Bruce, 23 and a French explor-
ing vessel under the command of Charcot 24 soon followed.
The altogether exceptional importance of the results obtained
by the British expedition under Scott led Lieutenant, now
Sir Ernest Shackleton, to fit out at his own expense the expe-
dition which for scientific results as well as for an exhibition
of fortitude in the face of exceptional difficulties, takes the
first rank in Antarctic discovery. 25

It should not, however, be overlooked that the " Chal-
lenger " exploring expedition, while undertaken primarily
for other than Antarctic exploration, entered Antarctic
waters in 1874, crossed the Antarctic Circle, and has fur-
nished especially valuable data upon the pack and berg ice
of that region. 26

192 CHARACTERISTICS OF EXISTING GLACIERS

Before the exploring expeditions of 1838 to 1841 had been
undertaken, much had been learned from the reports of
enterprising seal hunters in the Antarctic, such, for example,
as the Englishmen Biscoe, Kemp, and Weddell, and the
Americans Palmer, Pendleton, and others. It is certain
that as early as 1812 an American sealing station was main-
tained in West Antarctica. 27 Kemp and Enderby Lands
situated in longitude 80-90 W., and upon the Antarctic
Circle were discovered in this way and are no doubt continued
westward to the point earlier reached by Cook (see Fig. 100).

The expeditions of 1838-1841 were the first which really
discovered with certainty the Antarctic continent, though it
is now clear that Cook was in 1774 upon its borders. What
has been designated Wilkes Land was skirted by Captain
Wilkes from about 110 to 145 East in close correspondence
with the Antarctic Circle. Here either an undulating high
snow surface which Wilkes interpreted as a buried mountain
range, or a high ice cliff rising abruptly from the sea, was all
that could be made out, and in the fierce storms and almost
continual fogs which he encountered, seeing conditions were
most unsatisfactory. Ross, Borchgrevink, 28 and Scott 29
have all in turn sailed over the eastern portion of Wilkes
Land, and we now know that the coast does not here extend
along the Antarctic Circle as was supposed by Wilkes. The
Balleny Islands being near the coast as traced by Wilkes,
it is altogether probable that in the bad weather which he
encountered, these islands were mistaken for the continuation
of the Antarctic continent. 30 Von Drygalski has shown how
easy it was for Wilkes to have been mistaken in the loom
of the continent under the conditions which he encountered. 31
It is now probable that the coast line extends from near
Wilkes' Cape Carr in a more or less direct course to Cape
North in Victoria Land.

The same coast which Wilkes imperfectly and at great

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 193

risk to his vessels charted for 1500 miles, was seen also by
D'Urville commanding the French expedition; and on the
supposition that the land was seen by him a day earlier
than by Wilkes, Scott and Shackleton have each upon
their maps replaced the name Wilkes Land by Adelie and
Clarie Land, these being the names given by the French
commander. It has since been most conclusively shown
that owing to D'Urville's failure to drop a day from his
calendar when crossing the 180th meridian, his dates are in
error, and Wilkes' discovery was really made upon the same
day, but some hours earlier than D'Urville's, 32 so closely do
the two discoveries of the great Antarctic continent fall
together. Sir James Ross, experienced explorer as he was,
and in ships specially strengthened for the task in prospect,
achieved results of great importance, but was greatly cha-
grined that he had been anticipated by Wilkes in the discov-
ery of the Antarctic continent, and quite unjustly imputed
improper motives to the American commander. He sailed
along the coast which he named Victoria Land, with its
high range of bare peaks to which he gave the name Admi-
ralty Range, and for 500 miles he skirted the high ice cliff,
since generally known as the " Great Ross Barrier " (see
Fig. 100).

Since Ross's time three new land areas have been dis-
covered in the South Polar region, while Victoria Land and
West Antarctica have been much extended through explo-
ration. The three new land areas are King Edward VII
Land, described by the English Expedition under Scott,
Kaiser Wilhelm II Land, discovered by the German expe-
dition under von Drygalski, and Coats Land, which was
sighted by the Scotch exploring expedition under Bruce
(see Fig. 100). Coats Land was found upon Weddell Sea,
which, near longitude 20 West, forms a deep indentation in
the Antarctic continent nearly opposite the great indenta-

194 CHARACTERISTICS OF EXISTING GLACIERS

tion of Ross Sea. The question is still open whether these
seas may not eventually be connected across the Antarctic

FIG. 100. Map of Antarctica showing the principal points which have been
reached by exploring expeditions and their relation in position to the other con-
tinental masses.

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 195

barrier ice. Two expeditions, with a view to settle the ques-
tion, are to-day in prospect. 33

To summarize, land has now been definitely determined
to exist in King Edward VII Land (lat. 75 S., long. 150 W.),

FIG. 101. Map of the Antarctic regions showing the tracks of vessels. Based
on Murray's map of 1894, but brought up to date. The probable outline of the
continent has also been indicated, largely in accordance with Murray's view but
modified to express later discoveries.

Victoria Land (lat. 70-80 23' S., long. 165-170 E.), Wilkes
Land (lat. 66j S., long. 110-150 E.), Kaiser Wilhelm II
Land, Kemp, and Enderby Lands (all near the Antarctic
Circle and in longitudes 90, 60, and 50 E.), Coats Land
(lat. 74 S., long. 20 W.), and West Antarctica (lat. 65-70 S.,

196 CHARACTERISTICS OF EXISTING GLACIERS

long. 60-70 E). The tracks of vessels when charted to-
gether (see Fig. 101) have a value by showing where the
Antarctic land is proven not to extend. 34 Additional infor-
mation concerning the continental border may be obtained,
even where land has not been seen, through the observation
of true barrier ice. Whatever may be the origin of such ice,
in all cases where it has been explored, it has been found in
connection with land masses, and the supposition is strong
that all areas of true barrier ice indicate a connection with
land masses. Captain Cook in 1773, when near the Antarc-
tic Circle, and in longitude 40 E., saw a uniform and level
mass of ice extending along the horizon, which by imperfect
methods he estimated to be 15-20 feet high. This was cer-
tainly not pack ice. 35 Biscoe, in 1831, saw at a point some-
what east of Cook's position a perpendicular wall of ice
between 100 and 110 feet in height. Again, Kemp, in 1833,
at a point still farther to the eastward, discovered Enderby
Land, which, so far as he then knew, might be an island, but
connected with the discoveries of Cook and Biscoe, and
interpreted with our present knowledge, this barrier ice was
clearly part of the fringe surrounding the Antarctic continent.
The Submerged Continental Platform. Wilkes was care-
ful to confirm his discovery of the Antarctic continent by a
series of soundings which indicated the existence of a sub-
merged platform upon the margin of the continent. Ross
obtained off Victoria Land and also near the Ross barrier
depths of 100 to 500 fathoms, so that together the observa-
tions indicated the presence of a continental shelf, bordering
the Antarctic continent in these longitudes. 36 Arctowski, 37
from the soundings taken by the " Belgica " expedition to the
westward of West Antarctica, discovered a similar sub-
merged platform, which at its outer edge descended rapidly
from less than 300 to more than 1400 fathoms (see Fig. 102).
That the platform is at its margin more than twice the usual

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 197

depth of continental shelves, Arctowski interprets as evi-
dence of a general submergence of the region. If this inter-
pretation is correct, the amount of submergence is probably
even greater than the figures would indicate, for the sound-

1539,

950

FIG. 102. Soundings over the continental platform to the westward of West
Antarctica (after Arctowski).

ings of the " Challenger " farther to the westward showed
that the ocean bottom is there strewn with glacial debris,
due, doubtless, to the transporting action of drifting ice-
bergs. 38

The latest French Antarctic expedition has taken sound-
ings over a portion of the platform examined by the " Bel-
gica," and shown that it is characterized by rather remark-
able irregularities of surface. Farther to the westward along
the parallel of 70, and hence outside this shelf, a profound
fosse with depths in excess of 5000 metres was discovered,
though this decreased in depth to the westward near the
longitude of 126 W. 39

The Scotch Antarctic expedition, when off Coats Land,
found the bottom shelving very markedly to depths of 161
and 159 fathoms, which soundings were obtained some two

198 CHARACTERISTICS OF EXISTING GLACIERS

miles off the land. Then, within a distance of 50 sea miles,
the bottom dropped from 131 to 2370 fathoms, thus showing
that in this district also the slope bordering the submerged
Antarctic platform descends into great depths. 40

The winter station of the " Gauss " within the sea ice
north of Kaiser Wilhelm Land, was over a great submarine
plateau, the depths of which, as shown by soundings made
while approaching and retiring from the position, ranged
from 130 to 375 fathoms. Directly beneath the station
stretched a submarine ridge which may perhaps have repre-
sented a moraine. On the edge of this platform, the
" Gauss " determined the same abrupt descent to consider-
able depths which had before been found by the " Belgica "
farther to the eastward. At points quite near each other,
the " Gauss " determined depths of 241 and 2890 metres,
and at another place of 382 and 1103 metres. 41 Unlike
Arctowski, who assumed a subsidence of the platform to
account for its unusual depth, Philippi has interpreted this
as evidence that the platform was planed down by the
ice from the usual depth of about 100 fathoms when the
ice front extended to and beyond the border of the platform.
He points out that the general effect of the retirement of an
ice sheet is to induce elevation, rather than subsidence.

The Zone of Sea and Pack Ice. The sea ice of the South
Polar region encircles the Antarctic continent with outlines
roughly parallel to its borders. Sea ice is due to the freezing
of the surface of the sea during the winter months. Ross's
observations, but particularly those of the " Challenger "
expedition, 42 show that a wedge-shaped mass of cold water
extends in the sea through about 12 of latitude, the thin
northern edge terminating about latitude 53 S. Within
this wedge the temperature varies from 28 F. at the thick
southern end to 32.5 F. at the thin northern end. The
overlying warmer layer of water has temperatures varying

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 199

from 32 to 35 F., which represents also the range of tem-
perature of the water below the cold intermediate wedge.
During the winter, the warm surface layer is probably absent,
and in summer, as already indicated, this upper layer thins
toward the south so as to reach the surface at about latitude
65 S. 43 Over the continental platform to the westward of
West Antarctica, Racovitza found that the wedge-shaped
cold layer of water, here forming the surface, had a tempera-
ture of 2 C. (28j F.), and was thickest at the southern
end (lat. 31 1' S.), and that below this wedge the tempera-
tures increased gradually as far as the bottom, where they
ranged from to - 1 C. (32 to 33f F.). Above this con-
tinental plateau the cold water layer is thicker than the
warmer bottom layer. 44

At Cape Adare (lat. 71 15' S.) the water of the upper
layer remained constant at temperature 27.8 F. ( 1.5 C.)
whenever the surface was frozen. 45 At Wandel Island during
all the cold winter weather while the French expedition was
there, the sea water near the surface remained remarkably
constant at temperature 1.7 to 1.9C. 46

The term " field-ice " in the Antarctic regions applies to
the uniform sheet of frozen sea. During the formation of
this surface ice, some of the sea salts are squeezed upward
through capillary cracks to the surface and there freeze as
cryohydrates, which become the nuclei for further growth
from atmospheric water vapor. In this way, beautiful
rosette-like aggregates of crystals are produced. 47

The thickness of the sea ice becomes a matter of con-
siderable importance in the study of Antarctic barrier ice
soon to be considered, for it is probable that sea ice on
which snow accumulates may reach almost any thickness,
the ice being forced down below sea level by the weight of
the overlying snow. Under favorable conditions this ice is
later melted both from the bottom in the water, and from

200 CHARACTERISTICS OF EXISTING GLACIERS

the surface in the air. 48 This matter will be more fully dis-
cussed when the origin of the barrier, or shelf ice, of the
Antarctic regions is considered.

Where exposed to the wind, however, snow does not
generally accumulate upon smooth ice, in which case its
thickness is probably quite moderate. A thickness of
8J feet is the largest that was measured by the Scott expedi-
tion, 49 while 7 feet is the maximum reported by Shackleton. 50
These values were, however, obtained in exceptionally
high southern latitudes, and are much in excess of those
which have been measured outside of the Ross sea. Thus
the greatest thickness measured by Gourdon near Wandel
Island was 16 inches. 51

Not only is the thickness of the sea ice sometimes much
increased through snowfall, but when broken up in the
spring to form the pack ice, some layers are forced beneath
others and the whole is frozen into a compact mass of much
greater thickness. Thus blocks are built up to form slabs
(Schollen ice) which may be 25 > feet or more in height. 52

The cover of sea ice is subject to drift due to air currents
blowing over it, and this has special interest. Where
explored by the " Belgica " expedition, the coming of wind
was foretold by pressures within the ice. It was found that
during calm weather there was always a change in the pack
accompanied by cracks and open lanes, or leads. The
pressure is produced afterwards, though before the wind is
felt. Still later the wind arrives, and the pressure soon
thereafter ceases, when the ice pack is found to be drifting. 53
The pressure in the ice would, therefore, appear to be due to
the friction of the wind upon the upper surface near the wind-
ward side of the mass forcing that portion forward upon the
lee portion, or in some cases against a shore.

This discovery that the principal cause of the crushing
within the pack is the distant approach of wind has great

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 201

interest in showing that the inertia of rest possessed by such
a large body of ice the excess of the starting friction over
sliding friction induces a tremendous compressive stress
within this great ice raft along the wind direction. The wind
having exerted its frictional stress over a relatively small
area near the distant windward margin, the crushing condi-
tions are similar to those which exist in a long line of freight
cars suddenly struck and pushed
forward by a powerful engine at
the distant end.

The brittleness of ice at low
temperatures makes it pertinent to
consider these effects of compres-
sion within the pack in connection
with the well-known experiments
of Daubree and Tresca upon blocks
of moulder's wax. 54 A block of
this material in the form of a rec-
tangular prism was compressed
between the jaws of a testing
machine and found to yield by
rupture in a network of cracks
which were plane surfaces perpen-
dicular to the free surface of the
block and arranged within two sets
or series, which were approximately
perpendicular to each other, though
inclined 45 to the direction of com-
pression (see Fig. 103).

Theoretically in a body which
varies considerably from perfect
elasticity, there would be a larger angle than this with the
direction of compression, because of lateral yielding. Test
blocks of cement, for example, when similarly tested, show
much larger angles between the fracture planes.

FIG. 103. Cracks formed on
the free surface of a block of
moulder's wax when crushed
in a testing machine (after
Daubr6e and Tresca).

202 CHARACTERISTICS OF EXISTING GLACIERS

Arctowski, as a result of protracted studies upon the ice-
pack to the west of West Antarctica, found that the forms of
the /zigzagging open lanes of water, and the lakes which are
found within the pack, are both best explained by as-
suming the pack to

cz.

be made up of an
aggregation of simi-
lar quadrangular ele-
ments which compose
elements of similar
form, but of higher
order of magnitude.
When the pack is
subjected to traction,

^ j^y w j nc [g passing
' .

off its Surface, the
water j eadg n m

t r

parallel series, first in
one and later in another direction, though the larger rafts
continue to maintain a quite remarkable constancy of orien-
tation (see Fig. 104) , 55
That the neighboring lanes within the pack are essentially

FIG. 104. Open lane of water within the Antarctic
pack ice showing the minor elements of similar
form which are believed to be responsible for the
zigzagging courses of the water lanes (after Arc-
towski).

FIG. 105. Lozenge-shaped lakes within the pack arranged en echelon, and
believed to be due to separation and subsequent junction of the pack after a
differential shearing motion with reference to the line of rupture, the pack being
already divided into lozenge-shaped sections as a result of compression (after
Arctowski).

parallel was believed to be confirmed by the observation of
parallel ribbons of " water sky " in so many cases.

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 203

Moreover, the lakes of open water which are found within
the pack have quite generally a quadrilateral outline, and
could sometimes be seen to have their sides extended by
long rectilinear cracks (see Fig. 105).

This tendency of broadly extended ice plates to separate
into prismatic blocks is also common to the margin of the
shelf ice soon to be considered. A case where half submerged
glacier ice has been compressed against an obstructing island
and subsequently broken away so as to leave on open strait
between, reveals the same vertical fissures and the peculiar
zigzags as well. 56

The thickened sea ice of Posadowsky Bay near its eastern
margin has a zigzag outline, and there is found a series of
cracks parallel to the northern edge, to which cracks corre-
spond in direction parallel ridges of hummocks and ranks of
thickly packed icebergs. 57 This ice is stranded through the
icebergs on shallows of the bay, and v. Drygalski believes
that the wind friction upon the mass is the direct cause of the
phenomena. The outer edge, which is thickened sea ice,
changes its position from year to year, new ice being some-
times added to extend the mass, and at other times strips
being separated by traction of the wind to float northward
with the drifting pack.

While the drifts of the pack in which the " Belgica " was
frozen extended through 25 of longitude, its differential
motions appear to have been small. The pack and schollen
ice to the north of Kaiser Wilhelm Land where observed by
the German expedition, was notably stagnant, since the
" Gauss " maintained its position for many months. Under
such circumstances icebergs frozen into the pack maintained
their relative positions for long periods, and the snow chased
by the wind was arranged in long parallel sastrugi of great
perfection (see Fig. 106) , 58 Although in both these localities
unusually quiet, elsewhere sea ice has been seen to undergo

204 CHARACTERISTICS OF EXISTING GLACIERS

complex differential, so-called " screwing " movements,
which result in the well-known pressure ridges. Some of

these form from essen-

tially the same causes
as those on the surface
of small inland lakes of
the temperate regions.
During a fall of temper-
ature the ice contracts,
thus opening fissures
and leads, which in the
low temperatures are
quickly healed by the
formation of new ice.
The next rise of tem-
perature with the re-
sulting expansion of the ice, introduces powerful internal
stresses which, taking advantage of the relatively weak
" planks " of new ice within the fissures, buckles them up
into overfolds and overthrust faults. The more important
" hummocks " upon the
surface of the sea ice result,
however, from wind friction
upon its surface (see Figs.
107 and 108). Borchgre-
vink has thus described
this phenomenon : 59

FIG. 106. Sastrugi on Schollen ice as seen
from a balloon (after Von Drygalski).

^ s -jJt f^i^

FIG. 107. Pressure lines upon the sur-
face of sea ice (after Shackleton) .

In the evening, May 5th,
. . . we heard roaring and
crushing to the N.W. of our

peninsula, and when we came near the beach we witnessed a
scene of singular grandeur. The ice-fields were screwing and at
the beach the pressure must have been tremendous. Already a
broad wall some 30 feet high rose the whole length of the N.W.

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 205

beach, and coming nearer we saw that the whole of this barrier
was a moving mass of ice blocks, each several tons in weight. The
whole thing moved in undulations, and every minute this live
barrier grew in height and precipi-
tated large blocks on to the peninsula
where we watched the interesting
phenomena from the distance of a
few yards. The roar of the screwing
was appalling.

The "Antarctica/ 7 returning to
Snow Hill Island at the end of
the winter in order to take off
the Swedish Expedition, was
nipped in the pack ice, and its sides yielding to the pres-
sure, it sank so soon as a shift in the pack had released it
from its position. This pressure upon the ship was devel-
oped suddenly, " the ship began to tremble like an aspen
leaf, and a violent crash sent us all up on deck to see
what the matter was. The pressure was tremendous; the
vessel rose higher and higher, while the ice was crushed to
powder along her sides." 60 Later there was a second crash
and the ship's sides were crushed in (see Fio\ 109).

FIG. 108. Pressure ridge formed
on the shore of Victoria Land
(after Borchgrevink) .

FIG. 109. Sinking of the "Antarctica " (after Anderson).

The experiences of the crew of the " Antarctica " after the
abandonment of their vessel furnish us the best record of the
manner in which the pack is broken up in the spring into

206 CHARACTERISTICS OF EXISTING GLACIERS

great floes that are carried first in one direction and then in
another by the shifting winds. New leads suddenly open
at unexpected places, and a little later are as suddenly closed,
sending up pressure ridges and hummocks upon the ice sur-
face. 61

On Ross Sea the gales grow excessively violent towards the
end of September and in October (the Southern spring), and
by this time the sea-ice sheet has probably commenced to
weaken. The general break up of its surface was twice ob-
served by sledge parties connected with the Scott Antarc-
tic expedition. 62 On one day the sea would be seen com-
pletely covered with ice, and on the next appear as a clear
sheet of open water. Once freed, the ice drifts northward
and forms that heavy belt of pack ice which hems in the Ant-
arctic. Inasmuch as the small detached masses of ice move
faster than the main sheet, these float in advance upon the
north side like a line of outposts, whereas in the south and
rear there is a jam with loose pieces crowding hard upon the
pack. Amundsen has called attention to a striking con-
trast between the pack ice of the Antarctic regions and that
of the northern hemisphere, due to the more rapid currents
in the Arctic seas. While we find in the Arctic ice channels
and lakes several miles in length, no similar formations are
common in the South Polar region. Amundsen believes that
the " indolence" of the ice in which the " Belgica" was im-
prisoned is explained by the weak currents flowing beneath
it. 63

The manner of formation of the sea ice has been described
in some detail among others by Gourdon, who says : 64

With great cold the sea smokes like a furnace because of the
great difference between the water temperature and that of the
atmosphere. Its surface becomes glistening like that of oil. Mi-
nute needles of ice appear, which multiply rapidly and become
united into a firm network. The rate of accretion which I have

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 207

been able to follow by cutting little rectangles in the ice near the
boat is very rapid during the first hours of its formation. The
thickness often attains 6 to 7 centimeters (2J4 to 3 inches) in a few
hours. After that the increase goes on more and more slowly.
The first beds formed are an insolator which protects the subjacent
water, so that to attain a thickness of 12 to 15 centimeters (5 to 6
inches) requires several days. The ice of the first hours has a platy
structure, being formed of small lamellae which imprison between
them a little brine. The sea water in freezing throws off, in fact,
a large part of the salt which it contains ; these separated portions
having a saline concentration which lowers their freezing point.

This ice is pliable and plastic : a thickness of even a number
of centimeters undulates with the movements of the swell. If the
cold persists, it becomes compact, and with temperatures below
-20 C. (-4 F.) it is hard, brittle, and sonorous under the blows of
the pickaxe. Its transparency gives it a black appearance in con-
trast with the snow which covers it.

It is somewhat surprising to note what small thicknesses
of sea ice are formed in those cases where no snow has been de-
posited upon the surface. This has now been learned on the
basis of discoveries by Nansen, Peary, and others within the
Arctic regions. The greatest thickness measured by Gour-
don near Wandel Island was a little in excess of 16 inches.

Thus sea ice differs from lake ice in that it does not form in
vertical prisms, as does lake ice. According to Mawson 65 the
ice begins to form in scale-like crystals perhaps an inch in
diameter, which first float about within a few feet of the sur-
face. Through the motion of the water these scales soon
unite to form rosettes, and when they have become sufficiently
numerous, these in turn freeze together to form a complete
felt- work upon the surface. In this initial stage of the ice
cover, the ice is dark and partially transparent, as well as
peculiarly flexible. If there be a heavy swell, this cover is
broken into pieces a foot or more in diameter, depending

208 CHARACTERISTICS OF EXISTING GLACIERS

upon its thickness at the time, and these cakes by jos-
tling together become rounded and turned up at the edge
-the well-known " pancake ice." 66 Eventually, when the
cakes are again frozen together, a stronger cover is pro-
duced which increases in thickness through the growth of
vertical ice prisms upon its lower surface. These prisms
may be a half inch in diameter and many inches in length.
Snow falling upon the surface of ice increases the thickness of
the layer, and if continued through more than a single season,
the prisms of the lower layers grow upwards through re-
crystallization. The salt squeezed out of the water during
the formation of the prisms remains in white vertical tracts
between them.

Upon the landward side of sea or pack ice in contact with
the shore, there is generally to be found a fringe of thicker ice
known as shore ice or as the ice-foot. This foot usually rises
to a height of 6 to 10 feet above sea level and has the form of
a flat, narrow terrace 20 to 100 feet wide (see plate 27 B).
Sometimes, however, it shows a cliff 80 to 100 feet high with
a summit ascending inland in a more or less steep snow slope.
The ice-foot may have been formed either by the freezing of
the sea water which dashed as spray against the cliff, in
which case there are beautiful ice caves lined with stalactites,
but it is generally the result of a collection of snow in the form
of a drift under the lee of the cliff. 67 If formed either wholly
or in part from snow drift, the ice-foot is apt to have alter-
nate layers of compressed snow and of sand and gravel, both
alike the work of the fierce southerly blizzards. The sea ice,
which moves up and down with the tides, which on Ross Sea
have a range of from two to three feet, is usually separated
from the shore ice or ice-foot by one or 'more well-marked
" tide cracks."

The Ice Islands and Ice-foot Glaciers. The low islands in
high southern latitudes are always snow-covered, so that no

PLATE 27.

A. Fringing glaciers about Sturge Island, Balleny Group (after Scott).

B. Ice-foot with boat party landing (after Scott).

PLATE 28.

A. Ice-dome on Bouvet Island (after Chun).

B. Neve stratification in ice island (after Arctowski).

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 209

land is visible ^ the land is entirely enveloped in an ice-
cap (see Fig. 110). Such islands from a half mile to a mile
or more across are found to the northward of King Edward
Land. 69 Even in the low latitude of 54 the volcanic Bouvet
Island was sheathed in snow when visited by Krech in

FIG. 110. Ice island off King Edward Land (after Scott).

midsummer. 70 The clouds but partly conceal the perfect
shield form of the island in the view of plate 28 A.
Similar ice islands have been described by Arctowski. 71
Where these come down to the sea, the ice cliff shows the
characteristic neve stratification (see plate 28 B).

Where the islands are higher, the snow either wholly or in
part is blown by the wind from the higher surfaces into the
lee of the hills and thus forms a fringing zone of ice-foot.
Such a fringing ice-foot is illustrated by Sturge island of the
Balleny Group to the north of Victoria Land (see plate 27 A). 72
The ice-foot surrounding a land mass represents a type of
fringing glacier not unlike those described by Chamberlin
and Peary from Northern Greenland. 73 For long distances
these marginal bands of rather steeply sloping snow and ice
bound the elevated land and have in consequence been called
by Otto Nordenskjold ice-foot glaciers. They are obviously
built up from drift snow and have a definitely stratified struc-
ture. 74 These ice-foot glaciers are what Arctowski has de-
scribed as slope glaciers (Gehdngegletscher), and Gourdon
as "piedmont " glaciers. 76

Upon the larger islands of West Antarctica there are

210 CHARACTERISTICS OF EXISTING GLACIERS

found thin bodies of inland-ice through which the rock
peaks project as nunataks. This type of southern glacier
resembling as it does some of the ice-caps of Spitzbergen
has been designated by Otto Nordenskiold the Spitzbergen
type.

REFERENCES

1 Hobbs, "Characteristics of the Inland-ice of the Arctic Regions,"
Proc. Am. Phil. Soc., vol. 49, 1910, pp. 57-129, pis. xxvi-xxx.

2 December, January, and February.

3 June, July, and August.

4 Otto Nordenskiold and J. G. Andersson, " Antarctica, or Two Years
amongst the Ice of the South Pole," London, 1905, pp. 159-181. Also
" Die Polarwelt," Leipzig and Berlin, 1909, pp. 89-90.

6 E. v. Drygalski, " Zum Kontinent des eisigen Siidens," etc., Berlin,
1904, p. 387.

6 Henryk Arctowski. "The Antarctic Climate." Published as Ap-
pendix II of Cook's " Through the First Antarctic Night," New York,
1900, p. 427.

7 Louis Bernaechi, " Meteorology and Magnetism." Appendix in Borch-
grevink's " First on the Antarctic Continent," London, 1901, pp. 301-310.

8 E. H. Shackleton, " The Heart of the Antarctic," London, 1909, vol. 2,
pp. 386-389.

9 R. F. Scott, " The Voyage of the * Discovery,' " London, 1905, vol. 2,
pp. 208-211.

10 Robert C. Mossman, "Some Results of the Scottish National Ant-
arctic Expedition," Scot. Geog. Mag., vol. 21, 1905, p. 421.

11 O. Nordenskiold, "Die Polarwelt," p. 90.

12 Scott, " Voyage of the 'Discovery,' " vol. 2, p. 261.

13 Shackleton, "The Heart of the Antarctic," vol. 1, pp. 342-348.

14 For resumes of Antarctic exploration up to the revival of interest in
that region near the beginning of the twentieth century, see Karl Fricker,
" The Antarctic Regions," London, 1900, pp. xii and 292 ; also George
Murray and Sir Clements R. Markham (Editors), The Antarctic Manual
for the Use of the Expedition of 1901. Issued by the Royal Geographi-
cal Society, London, 1901, pp. 586. Also Georg v. Neumayer, "Auf
zum Siidpol, 45 Jahre Wirkens zur Forderung der Erforschung der Siid-
polarregion, 1855-1900," Berlin, 1901, pp. 1-483. Also E. S. Balch,
"Antarctica," Philadelphia, 1902, pp. 230. Also Hugh Robert Mill,
"The Siege of the South Pole," London, 1905, pp. 1-450. Later expedi-
tions have been treated by A. W. Greely in his " Handbook of Polar Dis-
coveries," 4th ed., 1909, pp. 1-336, in most respects an authoritative work,,
but marred by inclusion of the fictitious polar journey of the fakir Cook.

15 Charles Wilkes, "Narrative of the United States Exploring Expedi-
tion during the Years 1838-1842," especiaUy vol. 2, 1844, chaps. IX-XL
Also Atlas.

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 211

16 J. S. C. Dumont d'Urville, "Voyage au Pole Sud et dans 1'Oceanie."
1841-1854, vols. 2 and 8 and Atlas.

17 J. C. Ross, "Voyage of Discovery and Research to the Southern and
Antarctic Regions," 2 vols., 1846.

18 C. E. Borchgrevink, "First on the Antarctic Continent, being an
Account of the British Antarctic Expedition," 1898-1900, London, 1901,
pp. xv and 333.

19 Com. de Gerlache, "Quinze mois dans 1'antarctique," Paris, 1902, pp.
1-284. See also appendices in Frederick A. Cook, "Through the First
Antarctic Night, 1898-1899. A narrative of the voyage of the 'Bel-
gica' among newly discovered lands and over an unknown sea about the
South Pole," New York, 1900. Appendix I, on 'General Results,' by E.
Racovitza; Appendix II, 'Antarctic Climate/ by H. Arctowski; Appen-
dix III, ' Bathymetrical Conditions,' by H. Arctowski; and Appendix
IV, 'Navigation of Antarctic Pack Ice,' by R. Amundsen.

20 R. F. Scott, " The Voyage of the ' Discovery,' " 2 vols., London, 1905,
pp. xix, 556 and xii, 508.

21 E. v. Drygalski, "Zum Kontinent des eisigen Siidens, Deutsche Siid-
polarexpeditionen des 'Gauss,' 1901-1903," Berlin, 1904, pp. 668.

22 N. Otto Nordenskiold, and Joh. Gunnar Andersson, "Antarctica, or
Two Years amongst the Ice of the South Pole," London, 1905, pp. xviii
and 608.

23 Brown, et aL, "The Voyage of the ' Scotia,' being the record of a voyage
of exploration in Antarctic seas." By three of the staff. Edinburgh and
London, 1906, pp. xxiv and 375.

24 J. Charcot, "Le 'Francais' au pole sud," Flammarion, Paris, 1906.

25 E. H. Shackleton, "The Heart of the Antarctic, being the story of
the British Antarctic Expedition 1907-1909," 2 vols., London, 1910, pp.
xi, 371 and xv, 419.

26 " Report on the scientific results of the voyage of H. M. S. Challenger
during the years 1873-1876," London, 1885, "Narrative," vol. 1, pp. 39&-
434.

27 This small area of land, or some portion of it, has received so many
names that it seems well to avoid confusion by adopting the one general
term which is without international significance. TJie names Dirk Ger-
ritz Archipelago, Graham Land, Palmer Land, Danco Land, Alexander
I Land, and King Oscar II Land recall respectively Dutch, English,
American, Belgian, Russian and Swedish affiliations connected with
discovery.

28 Borchgrevink, 1. c., pp. 55-57, 2d map at end of volume.

29 Scott, "Voyage of the 'Discovery,'" vol. 2, pp. 390-393. Chart in
cover.

30 It is clear from the reading of Wilkes' narrative that the term "icy
barrier" which he repeatedly employs should not be interpreted in the
technical sense which it has since acquired. While in many cases it
clearly refers to true barrier ice, it is none the less evident from the lan-
guage used that in other cases pack ice only is referred to.

31 "Zum Kontinent des eisigen Siidens," etc., p. 389.

212 CHARACTERISTICS OF EXISTING GLACIERS

32 Rear Admiral John E. Pillsbury, U. S. N., " Wilkes' and D'Urville's
Discoveries in Wilkes Land," Nat. Geogr. Mag., vol. 21, 1910, pp. 171-173.

33 Wilhelm Filchner, A. Penek, et al., "Plan einer deutschen antark-
tischen Expedition," Zeitsch. Gesell. Erdkunde, Berlin, 1910, No. 3, pp. 1-6
(reprint). Also E. Bruckner, " Filchner' s deutsche antarktische Expedi-
tion," Zeit. f. Gletscherk., vol. 5, 1910, pp. 154-156, fig. Also W. S. Bruce,
" The New Scottish National Expedition, 1911," Scot. Geogr. Mag., vol. 26,
1910, pp. 192-195.

34 For a large-scale map showing tracks of vessels to 1905 see H. R. Mill,
" The siege of the South Pole," London, 1905, chart at end.

35 Fricker, " The Antarctic Regions," 1900, p. 225.

36 See John Murray and others, " Scientific advantages of an Antarctic
Expedition," Nature, vol. 57, No. 1479, 1898, reprinted in Smithsonian
Report for 1897, Washington, 1898, p. 419.

37 H. Arctowski, "The Bathymetrical Conditions of the Antarctic Re-
gions," Appendix III in Cook's "Through the First Antarctic Night,"
pp. 436-443.

38 John Murray, "The Renewal of Antarctic Exploration," Geogr. Jour.,
vol. 3, pp. 1-27. Reprinted in Smithsonian Report for 1893, Washing-
ton, 1894, p. 360. See also E. Philippi, " Ueber die Landeisbeobachtungen
der letzen fiinf Siidpolar-Expeditionen," Zeit. f. Gletscherk., vol. 2, 1907,
pp. 10-11.

39 Charcot, " Rapports preliminaires sur les travaux executes dans
1'antarctique de 1908-1910," Paris, 1910, pp. 101-102.

40 William S. Bruce, " Some results of the Scottish National Antarctic
expeditions," Scot. Geogr. Mag., vol. 21, 1905, p. 405, map plate opposite
p. 456.

41 E. Philippi, I.e., pp. 20-21.

42 Challenger Report, Narrative, vol. 1, pp. 417-428. See also Murray
and others, " Scientific advantages of an Antarctic expedition," Smithsonian
Report for 1897, Washington, 1898, pp. 418-419.

43 The warm lower stratum is probably due to waters heavy in saline
ingredients which come southward from the tropics, and, though diluted
by the Antarctic waters, have still a higher density because of their saline
content.

44 H. Arctowski and H. R. Mill, " Relations thermiques : rapport sur les
observations thermometriques faites aux stations de sondages," Exped.
Antarc. Beige, Antwerp, 1908, pp. 14-16, 20-24.

45 Bernacchi, I.e., p. 304.

46 Gourdon, I.e., 1908, p. 124.

47 Douglas Mawson, * Ice and Snow,' in Shackleton's " Heart of the
'Antarctic," vol. 2, p. 335.

48 Ferrar, in Scott's "Voyage of the 'Discovery,'" vol. 2, p. 459.
49 Scott, I.e., vol. 2, pp. 458-459. See also Racovitza, I.e., p. 417.
60 T. W. E. David, in App. II of Shackleton, I.e., vol. 2, p. 277.
51 Gourdon, I.e., 1908, p. 125.

62 Racovitza, I.e., p. 417.

63 Racovitza, I.e., p. 417.

ANTARCTIC CONTINENT AND ITS SEA-ICE GIRDLE 213

64 A. Daubree, " Etudes synthetiques de geologie experimental, " Paris,
1879, pp. 507-519, pi. II and figs. 93-94.

65 H. Arctowski, " Resultats du voyage du S. Y. Belgica en 1897-1898-
1899 sous le Commandement de A. de Gerlache de Gomery ; Oceano-
graphie, les glaces, glace de mer et banquises," Antwerp, 1908, pp. 39-44.

56 See the map of the Sefstrom glacier of Spitzbergen in De Geer, " Guide
de 1'excursion au Spitzberg," XI e Cong. Geol. Intern., Stockholm, 1910,
pi. 4.

57 E. v. Drygalski, "Das Schelfeis der Antarktis am Gaussberg," Sitz-
ungsber. k. bay. Akad. d. Wiss., Math.-phys. KL, 1910, pp. 12-15.

68 E. v. Drygalski, I.e.

69 Borchgrevink, I.e., pp. 120-121.

60 C. J. Skottsberg in " Antarctica," I.e., p. 524.

61 Skottsberg, I.e., pp. 537-543.

62 Scott, I.e., vol. 2, pp. 405-406.

63 Roald Amundsen, "The Navigation of the Antarctic Ice-pack,"
Appendix V in Cook's "Through the First Antarctic Night," pp. 450-451.

64 Gourdon, I.e., 1908, p. 124. See also Arctowski, I.e., 1908, p. 19.

65 Douglas Mawson in Shackle ton's "Heart of the Antarctic," vol. 2,
p. 337.

66 Scoresby, " An account of the Arctic Regions," p. 239.

67 H. F. Ferrar, I.e., pp. 459-460. Mawson, I.e., p. 338. T. W. E.
David, ibid., pp. 279-281.

68 O. Nordenskiold, " Einige Beobachtungen iiber Eisformen und Ver-
gletscherung der antarktischen Gebiete," Zeit. f. Gletscherk., vol. 3,
1908, p. 322.

69 Royal Society, National Antarctic Expedition 1901-1904, Album of
Photographs and Sketches, London, 1908.

70 Chun, et al., " Wissenschaftliche Ergebnisse der deutsch. Tiefsee-Ex-
pedition auf dem Dampfer Valdivia, 1898-1899," Jena, 1902.

71 Geogr. Jour., vol. 18, 1901, pp. 370.

72 Scott, I.e., vol. 1, p. 390 and plate opposite.

73 See Proc. Am. Phil. Soc., vol. 49, 1910, p. 104.
^ Nordenskiold, Zeit. f. Gletscherk., I.e.

75 H. Arctowski, " Die antarktischen Eisverhaltnisse ; Auszug ausmeinem
Tagebuch der Siidpolarreise der 'Belgica,' 1898-1899," Pet. Mitt. Erg.
144, 1903, pp. 15, 19, 21.

76 E. Gourdon, in Charcot, Expedition Antarctique Frangaise (1903-
1905), Glaciologie, Paris, 1908, p. 110, pi. 1, Fig. 4, and pi. x, Fig. 35.

CHAPTER XIII
THE MARGINAL SHELF-ICE

Its Nature and Distribution. The so-called "barrier" ice,
such as was without doubt seen by Cook in 1774, offers one
of the peculiarities in which the South Polar area is sharply
differentiated from its antipodal region. Nowhere within
the Arctic regions is there found to-day anything which in any
degree can be compared to the Antarctic barrier ice. Until
the British Antarctic Expedition of 1901-2, the origin of
this ice was a complete mystery, and even to-day widely
different interpretations have been offered. There is, how-
ever, every reason to believe that during the period of
Pleistocene glaciation, similar ice masses occupied the Gulf
of Maine in Northeastern North America as well as the
borders of the continent of Greenland and of Patagonia.
It is this fact, especially, which lends unusual interest and
importance to the study of the existing barrier ice of the
Antarctic regions.

At the outset it is well to point out that the term " bar-
rier ice " is in every way inappropriate for scientific use,
for it suggests merely that this form of ice opposes a barrier
to navigation. The term shelf-ice proposed by Norden-
skjold 1 is aptly descriptive and will be adopted here. The
term " piedmonts afloat " proposed by Ferrar for such
masses of barrier ice on the margin of the Ross Sea, has

214

THE MARGINAL SHELF-ICE

215

much to recommend it, but suggests somewhat too strongly
the identity in origin with land piedmonts. 2

As already pointed out, both Cook and Biscoe encountered
true shelf ice to the westward of Kernp and Enderby Lands
(see ante, p. 196), and though Wilkes uses the term " icy

FIG. 111. King Edward VII Land, with shelf ice in front (after Scott).

barrier " for obstructing ice of any kind, his descriptions leave
us in no doubt that the shelf ice was encountered near Cape
Carr in Wilkes Land. The following extracts from his nar-
rative set forth the aspect of this shelf ice as viewed from
the sea :

In some places we sailed for more than fifty miles together
along a straight and perpendicular wall from one hundred and
fifty to two hundred feet in height, with the land behind it. The
ice-bergs found along the coast were from a quarter of a mile to
five miles in length.

At 10 o'clock we were not more than three or four miles dis-
tant. It appeared prodigious. We saw a cliff with a uniform
height of 100 to 150 feet forming a long line westward. . . .

Discovered a high barrier of ice to the northward with ice
islands to the southward. . . .

The immense perpendicular barrier encountered yesterday was
now in sight trending as far as the eye could reach to the
westward. 3

A year later Ross sailed for a distance of 500 miles along
the front of the similar ice wall which has since been named

216 CHARACTERISTICS OF EXISTING GLACIERS

in his honor the " Great Ross Barrier." Within the last dec-
ade shelf-ice has been discovered by the Swedish expedi-
tion near King Oscar II Land, by the German expedition
near Kaiser Wilhelm II Land, by the English Expedition in
King Edward VII Land, and by the Scotch Expedition in
Coats Land. The first two districts having been examined
in some detail upon the spot, will be more fully discussed
below under separate headings. Of the Coats Land " bar-
rier "" it is stated that it formed the terminal face or sea
front of the great inland ice 4 (see Fig. 112), which is also

FIG. 112. The Scotia off Coats Land, the shelf ice showing to the right in the
middle distance and also in the distance (after Bruce).

true of the shelf-ice of King Edward VII Land. (See Fig.
111.)

The " Great Ross Barrier/' Victoria Land. In 1840
Sir James Ross skirted for a distance of 500 miles an ice
cliff which according to his estimates had an average
height of 165 feet. The next visit to this ice wall was
made in 1899 when Borchgrevink sailed for some distance
along its front, 5 and in 1902 Scott made a detailed survey
of its entire length (see Fig. 113). 6 The appearance of

PLATE 29.

A. The margin of the Great Ross Barrier (after Scott).

B. Near view of the Great Ross Barrier where highest, 280 feet (after Scott).

THE MARGINAL SHELF-ICE

217

this mighty ice cliff as seen from the sea is brought out
to advantage in Plate 28 A and B. The height when
observed from a short distance appears remarkably uni-

FIG. 113. Map of the Great Ross Barrier showing heights of the cliff in feet and
soundings of the sea in fathoms. Full line is track of " Discovery " (after Scott).

form, though on approaching nearer it is seen to vary
from 50 to 280 feet, and in places is even lower than the
minimum figure given. Scott mentions a locality where the
ice face is so low that one could step from the rail of the
" Discovery " directly on to the summit of the barrier. 7
(See Fig. 115 a.) A higher edge is represented in Fig. 115 6 ;
in which the " Discovery " is seen against the ice cliff within
a narrow cove of the ice margin.

FIG. 114. Section along the Ross Barrier edge based on Scott's figures and show-
ing the underlying water layer upon the assumption that the submerged and
emerged portions of the ice are in the ratio by volume of 5 to 1.

Examination of the perpendicular face of the Ross Bar-
rier shows clearly that its structure is quite different from

218 CHARACTERISTICS OF EXISTING GLACIERS

that of true glacier ice. It is an immensely thick formation
of snow horizontally stratified. Even at a great distance
its horizontal upper surface, its vertical fractures, and its
dazzling whiteness, all distinguish it from ordinary glacier
ice. Studied in detail at different levels, it is seen that
pressure has transformed the snow grains into neve snow, the
granules of which increase in size and are more intimately
interlocked toward the bottom of the cliff. In the upper
portions particularly the snow is porous, and hence im-
prisons a large quantity of air. A study of bergs derived
from the barrier which had floated into McMurdo Sound
where they were frozen into the sea-ice, showed that ex-
cept where spray had frozen over the surface they con-
tained no solid ice whatever in the levels above the sea
surface. Inasmuch as they were much tunnelled by sea
caves, it was possible to follow the study well into the
interior. Everywhere, however,, they showed only com-
pressed snow. 8

The specific gravity of the shelf ice must as a consequence
be much below that of true glacier ice, so that the barrier,
if afloat, should float relatively high. Scott estimates that
fully one-fifth of the mass must be above the water surface.
Even this proportion may not fairly represent the buoyancy
of shelf ice, for Captain Evans of the " Nirnrod " took
soundings around a typical tabular iceberg derived from
the barrier, and found that although its height was 80 feet,
it was aground in water of the same depth. 9 In this case,
therefore, half the mass projected above the water. Off the
Ross Barrier, Sir James Ross obtained soundings of 1360,
1800, and 2400 feet, 10 and more recently Scott has shown by
soundings that even if one-fifth only of the mass were above
the water, there would still be some hundreds of fathoms
between its bottom and the bottom of the sea (see Fig.
114). Moreover, since Scott's surveys show that much of

THE MARGINAL SHELF-ICE

219

the cliff is to-day twenty to thirty miles farther south than
when Ross visited it in 1840, these later soundings are well

FIG. 115 a. Margins of the Ross Barrier on Balloon Inlet, where so low that one
could embark directly from the ship's rail.

FIG. 115 6. Where relatively high.

within the border of the shelf ice of the earlier date (see
Fig. 113).

220 CHARACTERISTICS OF EXISTING GLACIERS

Further evidence that the barrier is afloat is derived from
the fact that for some distance back from its edge the ice
rises and falls with the tide and leaves behind a complex
system of vertical fractures as evidence.

FIG. 116. Outline map of the known portions of the Great Ross Barrier showing
the position of the outlets from the ice plateau (based on Shackleton's map).

Although the Ross Barrier has been crossed by Scott,
Royds, and Shackleton for long distances and in one
case for over three hundred miles (see Fig. 116), almost
the whole mass is believed to be afloat. Soundings not
being possible at points within the margin, the best evi-
dence is obtained from its almost perfectly level surface.
Scott took aneroid readings at every half degree of latitude
along the line of his southern journey, and corrected his
readings by comparison with the hypsometer and later
with simultaneous readings of the barometer made at the
winter quarters near Cape Royds. When thus corrected
it was found that the aneroid readings indicated no increase

THE MARGINAL SHELF-ICE 221

of elevation toward the South, but on the contrary, a slight
and gradual rise of barometer was noticeable such as might
be ascribed to the gradual advance toward a fixed area of
high atmospheric pressure. 11

Strong confirmatory evidence for the floating of the barrier
is derived also from measurements of temperatures within
fissures of the ice. Lieut. Royds found that whereas near
the visible land of White Island the serial temperatures in
fissures of the shelf ice fell to a mean level of 9 F., at dis-
tances of ten miles off the island such temperatures first
fell, but at greater depths rose, and at nineteen fathoms
(the limit of the test) showed F. This rise in the tempera-
ture with depth is best explained through the approach to a
water layer beneath the ice. 12

The surface of the Ross Barrier ice, as already stated, is
remarkably level. Within narrow limits this is well brought
out in plate 30 A and B, which represents photographs of the
surface, in one case from a captive balloon. The statement
requires modification for those portions only of the shelf
ice which approach the continent. In part the Ross Barrier
clearly derives its nourishment from the inland plateau ice
lying to the south and west. The outlets for this material
are great ice streams (one of them fifty miles in width), and so
unlike any other known type of glacier that they are deserv-
ing of a new and technical name. In the reports of the
British expeditions they have been referred to as " in-
lets " because they offer a possible ingress to the plateau.
The term outlet would better describe their function in the
ice economy, and they will hereafter be referred to by that
term. Off these great outlets from the inland plateau ice,
the surface of the shelf ice is found to be thrown into long
undulations which are recognizable for a distance of twenty
miles or more. 13 Elsewhere in the vicinity of the land a
similar but narrower zone of disturbance is noticed, which

222 CHARACTERISTICS OF EXISTING GLACIERS

may generally be followed out from the borders for a distance
of ten to fifteen miles. Within these marginal zones the
surface of the ice is much crevassed and in striking con-
trast with its otherwise smooth surface. Similar disturb-
ances accompanied by complex crevassing are observed also
about islands which project through the ice nearer to its
outer margins. Within a zone immediately adjacent to the
mountain borders on the south and west and within the
disturbed zone, there is a notably smooth ice surface, which
is a result of melting through radiation from the rock
surface. The ice surface is here in reality that of a
frozen lake. 14

A motion within the Ross Barrier was determined by
Scott's party from observations at " Depot A " near
Minna Bluff seventy-five miles or more from the cliff edge.
Here during a period of 13 J months the movement was
1824 feet in a direction a little to the east of north or
toward the barrier edge. This corresponds to an annual
rate of something over 1600 feet. The determination came
about through an accidental rediscovery of the station;
but even more important, the depot was again rediscovered
and relocated by the Shackelton party after another in-
terval, this time of over six years. The movement during
this interval amounted to 9600 feet, or about 1500 feet per
year, in a direction east-northeast.

This important verification of the earlier determination
that the shelf ice of the Ross Barrier moves at a rate of more
than four feet per day, or nearly four times as fast as the edge
of the inland ice of Kaiser Wilhelm Land, 15 must certainly
be accounted of the greatest importance. If its cause is
the contribution of plateau ice furnished through the out-
lets along its borders, the ice in these must either have
a very rapid movement or be exceptionally important
at points beyond where exploration has been carried to

PLATE 30.

Horizontal surface of the Ross Barrier, to the south of Minna Bluff, with sastrugi

(after Scott).

B. View of surface of Ross Barrier taken from a captive balloon, showing sastrugi.
black spots are men and the long dark lines their shadows (after Scott).

The

PLATE 31.

A. A new ice-face on the Ross Barrier (after Scott).

B. An old ice-face on the Ross Barrier (after Scott).

THE MARGINAL SHELF-ICE 223

the south of the Beardmore outlet. The possibility is
not excluded that the Ross Barrier is directly connected
with the shelf-ice at the head of the Weddell Sea on the
opposite side of the pole, and that drift sets in the direction
of the former.

Although the shelf-ice is unquestionably in part nourished
by the outlet glaciers leading down from the ice plateau to
the south and west, it is itself a vast neve, as has already
been shown from study of its structure, and account must,
therefore, be taken of alimentation from the snow falling
upon its surface.

The annual snow fall at Depot A, about seventy-five
miles from the barrier edge, is equivalent to 1\ inches of
rain. Though usually reckoned as the equivalent of as
many feet of snow, the snow is here so compact as to
possess less than twice the volume of the equivalent water
(or 13^ inches). 16 At Cape Royds, the winter station near
the barrier edge, the annual snow fall was estimated on
the basis of measurements as the equivalent of 9J inches of
rain. These figures, however, like those obtained at Depot
A, include drift snow, and there is no means of telling what
proportion of the total was locally derived and what was
brought from a distance by the winds. Although still heavier
falls are assumed for the Drygalski ice barrier tongue to the
northward, it should be noted that at Cape A dare where the
likelihood of collecting drift is comparatively small (lat.
71 15' S.), the snow collected by the gauges of the Borch-
grevink party during an entire year was equivalent to but
3 inches of rain. 17

Whether from drift or from local precipitation, the effect
of snow in nourishing the shelf ice is much the same, and it
is estimated that on the average about one foot of heavy
snow is each year added to the surface of the Ross Barrier.
If the contribution of the ice from the Beardmore outlet be

224 CHARACTERISTICS OF EXISTING GLACIERS

estimated to have moved toward the barrier edge at the
uniform rate of about one-third mile annually, before it
could have covered the 300 miles separating the outlet from
the present margin, some 900 years must have elapsed,
and during this time this glacier ice will have been buried
beneath some 900 feet of compact snow as measured
at surface density. 18 The true glacier ice derived from the
outlets is, therefore, not to be looked for in the shelf ice ex-
cept in the submerged portions where direct observation has
not yet gone. The upper and visible portion of the Ross
Barrier is hence in all probability throughout of local deri-
vation and is properly regarded as neve snow. 19 Some
confirmation of these conclusions is derived from the study
of the structure of Antarctic icebergs, which after partial
melting, or after overturning, bring the bottom layers to
the light of day (see below under Icebergs).

The " Higher " and " Lower " Ice Terraces off King Oscar
Land. The Swedish Antarctic expedition of 1902 en-
countered large areas of shelf-ice in most respects resembling
that of the Ross Barrier. This was met on the long sledge
journey of Nordenskiold and Sobral in a direction west-
southwest ward from the winter quarters at Snow Hill Island.
After eight days upon the sea-ice of Larsen Bay and when
near the Seal Islands (see Fig. 117), 20 a high ice wall suddenly
appeared across their path. This wall was ascended over
the sloping surface of a snow drift banked against it, and the
course was laid over " an even plateau destitute of fissures."
Once only, a faint depression was noted from which the land
could not be seen. Near the marginal cliff of this " lower "
terrace a few lava islands projected through its surface, and
here alone smooth ice cracks or pressure ridges were en-
countered. After travelling about 100 miles over the sur-
face of this " lower " terrace, the land was approached,
and for the first time, the surface appeared broken by numer-

THE MARGINAL SHELF-ICE

225

ous crevasses so deep and broad as effectually to block further
passage in that direction. 21 Here there rose a second terrace
of ice going out from the shore of King Oscar Land and
extending in a nearly straight line toward the east until it

FIG. 117. Map showing the " higher " and " lower " terraces of shelf-ice near King
Oscar Land (after Nordenskj old) .

was lost in the horizon. In contrast with the " lower "
terrace this " higher " terrace was broken into numberless
fissures. 22 Nordenskj old's belief is that the shelf-ice (the
" lower " terrace) is in the main nourished through the
precipitation and gradual accumulation of snow upon the
surface of sea-ice above a shallow sea. Over the ice of Larsen
Bay the sledging party had found in October a thick layer
of snow covered by a light crust, through which the ice-axe

226 CHARACTERISTICS OF EXISTING GLACIERS

could be driven to the depth of a metre. As the snow layer
upon the ice deepens, and the underlying sea-ice is by its
weight more and more depressed toward the shallow bottom,
the warming effect of the water would gradually decrease,
and the snow layer in consequence would increase in thick-
ness at an accelerated rate. 23

It is strongly emphasized by Nordenskjold that in this
accumulation of the snow the wind plays a larger role than
local precipitation. On the Snow Hill Island ice-foot the
surface was raised a few centimetres only during the winter,
whereas it increased fully thirty centimetres during the
summer. It should be borne in mind that the summer
months have air temperatures corresponding to those of
winter in lower latitudes (about 1 F. in the warmest month),
and more snow is precipitated during the summer months.
This snow is, moreover, softer, and adheres more readily to
the surface on which it is deposited. Still further, the winds
during the summer months are upon the average only about
half as strong as during the winter. Wherever protected
from the wind snow accumulates, so that small islands are
covered, and the ice-foot glacier pushes out from the margins
to be extended in the form of shelf-ice.

Valuable data bearing upon this point are also being
supplied from a different quarter. Mr. J. B. Tyrrell during
many winters spent about the fresh water lakes of Canada,
has found that if snow falls to a considerable depth soon after
the ice has first formed, this load will press the ice down
into the water. Young and flexible ice will bear up less
than its own thickness of the dense snow of the Canadian
wastes. With the greater thicknesses which are common,
the ice is bent down and water rises through fissures so as
to wet the lower snow layers. With severe weather this
wet snow is frozen and the ice thickened from the upper
surface. 24

THE MARGINAL SHELF-ICE 227

We have seen that the Scott, Shackelton, and Nordenskjold
expeditions are practically in agreement as to the importance
of local snow deposition in the alimentation of shelf ice
formations. Nordenskjold would ascribe both the origin
and growth of the shelf ice of West Antarctica to this cause,
whereas Scott regards the Ross Barrier as the relic of a much
larger area of ice shelf which once filled all of Ross Sea and
rested throughout upon its floor. This view of the former
extension of the Ross Barrier is, as we shall see, abundantly
supported by evidence. Of great importance is the com-
parison of the barrier margins of 1840 and of 1902 (see Fig.
113), since during a period of sixty years this wall has retired
in places from twenty to thirty miles.

The " West-ice " of Kaiser Wilhelm Land. To the west
of Posadowsky Bay and westward and northward from the
inland ice of Kaiser Wilhelm Land, lies a dead mass of ice
which the late Professor Philippi regarded as true shelf-ice,
and which in the main may be compared to that of the Ross
Barrier. 25 Owing, however, to the different opinions which
have been expressed concerning its origin, this area of stag-
nant ice has been given the colorless designation " West-
ice." Unlike the Ross Barrier, with which the West-ice
has been compared, it has a blue color, and as already men-
tioned, it appears to be stagnant, since no evidences of dis-
turbance have been found at either its sea or its inland-ice
margins. Unlike the Ross Barrier, also, it lacks the smooth
surface of that body, where it has been explored. For the
most part, its surface is very uneven, and might even be
described in places as chaotic or labyrinthine. In its
northeastern portion it is traversed by deep rift-like valleys,
which led von Drygalski to believe that it is constituted of a
group of closely crowded icebergs more or less welded to-
gether and with the intervening passages partially healed by
the indrifted snow. He has, however, referred to the West-

228 CHARACTERISTICS OF EXISTING GLACIERS

ice as similar to the shelf ice of Ross Sound and West Ant-
arctica. 26

Seen from the sea when the " Gauss " skirted its front, the
West-ice showed a high perpendicular wall in all respects

FIG. 118. West-ice seen from the "Gauss" off Kaiser Wilhelm Land (after von

Drygalski) .

resembling the cliff faces of the other bodies of Antarctic
shelf ice (see Fig. 118), and this wall was followed through
three degrees of longitude. The eastern portion of the mass,
which was examined by sledging parties, pushes its margin
out to the northward and ends in three great ice tongues
separated by bays and terminating in steep cliffs. These

FIG. 119. The junction of the West-ice and the sea ice (after von Drygalski).

cliffs at the different localities that were visited varied in
height from fifteen to sixty-five feet. Locally drifts of snow
formed sloping bridges down to the sea ice (see Fig. 119).
Both sea and shelf ice rose and fell together with the tides,
since no tide cracks were observed to separate them. This
indication that the West-ice is afloat was confirmed through

THE MARGINAL SHELF-ICE 229

the absence of any ice-foot, as well as by soundings, which
showed a depth of water along its borders of six hundred
metres. Deep disintegration of the West-ice through
melting upon its upper surface was everywhere apparent.
Old cracks running parallel to its margin were melted on
one side, so that steep cliffs faced northward and formed the
south wall of channels for surface streams. Broad trough-
like inbreaks led into the mass from its eastern margin,
and on one of these the floor had sunk unequally so as to
leave the north side high and the south side at a far lower
level.

In general, however, the surface of the West-ice is flat
with no apparent increase of elevation toward the west and
south, though far in the distance along these directions were
seen the rising slopes of the inland-ice. That there is to-day
no functional connection between the West-ice and the
inland-ice has been asserted by von Drygalski (see below,
p. 250).

According to von Drygalski, the West-ice is kept in place
because of the shallowness of Posadowsky Bay, icebergs being
first stranded on the shallows and the intervening lanes
being thereafter filled in by drifted snow. While this process
furnishes an explanation for the southern or older sections
of the mass, the northern or newer portions he believes to
have been formed by the thickening of sea ice which has
remained in place for a number of seasons. Three north
and south bands are made out whose order from east to
west appears to be significant in showing the manner of
formation of the greater portion of the West-ice mass.
The zone to the eastward and on the margin is pack-ice
(Scholkneis)', the middle zone is largely berg ice frozen
into a continuous sheet; while to the west the intervening
spaces which separate similar fleets of bergs have become
either partially or wholly filled in by drift snow, the product

230 CHARACTERISTICS OF EXISTING GLACIERS

being called " full-ice " (Votteis), or, in other words, the West-
ice proper 27 (see Fig. 120).

Ice

FIG. 120. Diagram showing manner of formation of West-ice. Eroded icebergs
crowded together, cemented by pack-ice and the intervening lanes partly filled
in with snow (after v. Drygalski).

The Shelf -ice Tongues of Victoria Land. Victoria Land
has furnished several examples of a new type of glacier ice

FIG. 121. Map of the glaciers and ice barrier tongues about the head of Robert-
son Bay, Victoria Land (after Borchgrevink) .

THE MARGINAL SHELF-ICE 231

which has interesting relationships to the shelf-ice of the
Antarctic regions. It is deserving of a distinct name, and
the term shelf-ice tongue (ice barrier tongue of the Shack-
leton expedition) seems on the whole the most characteristic
and descriptive. A related form of tongue was first described
by Borchgrevink in his map of the Sir John Murray glacier
on Robertson Bay, which lies behind Cape Adare in Victoria
Land (see Fig. 121). This glacier with the Dugdale glacier
descends to the sea below Geikie Land, where it is for some
distance wedged in between Duke of York Island and the
shore. It pushes out to sea in the form of a long dock,
which is 80 feet in height near its margins and rises into the
form of one of the dry deltas of an arid region. This form
of its surface is of special interest in showing clearly the
connection as regards nourishment between the ice of the
tongue and the glacier outlet above. From the peculiarities
of its surface it would appear to include no true shelf-ice
such as is found in the Ross Barrier.

Three large and well marked examples of shelf-ice tongue
or " piedmonts afloat " have been reported on by the
Shackleton expedition. These are Glacier Tongue, about
five miles long and located near the winter quarters on
McMurdo Sound, and the much larger Nordenskiold and
Drygalski shelf-ice tongues on the shore of Victoria Land
to the west of Ross Sea (see Fig. 122). Smaller tongues of
the same type, Harbor and Cheetham shelf-ice tongues, lie
similarly at the foot of other but smaller glaciers upon the
same shore. Both the Glacier Tongue on McMurdo Sound
and the Drygalski tongue to the northward were shown by
soundings near their outer margins to be afloat. The Dry-
galski tongue pushes out some thirty miles from the shore
and is more than twelve miles in width. On the basis of
soundings it has been thought to be afloat for at least three-
fourths of its length, but inasmuch as it rises toward the

E. Longitude

Fig. 122. Map showing the shelf-ice tongues on the west of Ross Sea with the
glacier outlets which descend to them from Victoria Land (after Shackleton).

232

THE MARGINAL SHELF-ICE 233

centre to heights much above its edge, this may be true for
the marginal portions only. 28

The Drygalski ice barrier tongue is clearly nourished from
the ice plateau through the great David Outlet, the ice of
which raises its shoreward end into a steep and irregular
ice apron; but farther out this is " levelled up with snow "
and passes into the true flat shelf-ice. At a point only
eighteen miles from the shore the marginal cliff was about
fifty feet above the water. In all essential respects this
tongue appears to resemble that portion of the Ross Barrier
which is just below the Beardmore Outlet (see Fig. 134, p.
258), with the exception that the broad extension of shelf-ice
is here reduced to a small marginal rim (in the tongue of the
Sir John Murray Glacier there is no rim whatever). As there
is every indication that the Drygalski tongue is in motion

FIG. 123. Ideal section through shelf-ice tongue showing the apron-like foot
of the outlet which feeds it, and the probable pedestal by which it is connected
with the bottom and maintained in position. The relation of its glacier ice to
the n6v6 of local derivation is also indicated.

and receiving abundant nourishment from the David Outlet,
though levelled up with snow near its outer edge, additional
light is thrown upon the origin of shelf-ice in general. The
probable section of a shelf -ice tongue is represented sche-

234 CHARACTERISTICS OF EXISTING GLACIERS

matically in Fig. 123. The nourishing glacier raises the
surface of the tongue into an apron, and in consequence
depresses the bottom of the submerged portion, and being
nearest the shore where the water is shallowest, must develop
a sort of ice pedestal whose effect will be to stiffen the
structure and prevent its being shifted in position. The
attenuated form which some of the tongues maintain it
would otherwise be difficult to explain.

The Nordenskjold ice barrier tongue is somewhat smaller
than the Drygalski tongue and appears to be no longer deriv-
ing nourishment from the plateau ice above. It is thus a
relic only of the once larger Ross Barrier, and has additional
interest because its southern edge is formed of ice probably
originally derived from the Mawson Outlet, whereas its
northern edge is of snow forty to fifty feet in thickness
brought by the southerly blizzards from the southern side.
This is bounded by vertical sea cliffs where slices have been
carried away with the sea-ice during the summer. It thus
emphasizes the important role which wind drift plays in the
formation of shelf-ice.

On a portion of the earth's surface where rain is unknown,
and where the air temperatures seldom rise above the freez-
ing-point, unfrozen water as a geological agent has an almost
negligible importance. In certain localities, however, where
the foehn winds are especially strong, such, for example, as
the David Outlet, its importance may be considerable.
During the weeks of December and January torrents of water
rush off the surface of the Drygalski tongue in the form of
englacial and subglacial streams. These either cut deep
open valleys upon the surface, or tunnel channels under the
hard snow and ice.

The Rectangular Table Berg of Antarctic Waters. The
normal iceberg of Antarctic seas is as different as possible
from the Arctic type, and for reasons which are now suffi-

THE MARGINAL SHELF-ICE

235

ciently obvious. In Greenland, true glacier ice descends to
the fjord heads, and there gives birth to bergs of blue ice
which are limited in size both by the size of the fjord and by
the crevasses upon the ice. In the Antarctic, so far as yet
known, glacier ice descends directly to the open sea at few
points only, but in its place appears the shelf-ice, and tabular
bergs separate along broad sea fronts which are measured
sometimes in the hundreds of miles (see Fig. 124). The
size of Antarctic bergs is in consequence many times greater,

FIG. 124. The Ross Barrier breaking away to form a tabular and rectangular
iceberg (after Shackleton).

and their form is tabular 29 like the ice-shelf from which they
have been born (see Fig. 125). 30

Most of the bergs which were seen in Ross Sea had been
derived from the Ross Barrier. They separate from it in
great rectangular blocks and leave a relatively smooth ver-
tical face, which later under the action of the waves becomes
undercut and more irregular through the separation of small
bergs on rectangular joint planes. It is thus easy to deter-
mine those parts of the barrier edge which are relatively
fresh, and those which have not for a considerable time
given birth to a tabular berg (see plate 31, A and B). 31

236

CHARACTERISTICS OF EXISTING GLACIERS

Such bergs often show in addition a distinctly terraced struc-
ture (see Fig. 126). The term tabular berg, which is in com-
mon use, is, however, particularly well chosen, because it

i'lo. 125. Rectangular and tabular iceberg of Antarctic waters (after Wyville

Thomson).

describes, not alone the smooth horizontal upper surface,
but the well-squared rectangular outlines in the plan. Too
little attention seems to have been directed to this impor-
tant fact, to which practically all photographs of Southern
icebergs bear witness. It indicates, as we believe, that the
shelf-ice at least near its margins is, particularly near the

FIG. 126. Tabular Antarctic iceberg showing perpendicular and rectangular joint-
ing (after Wyville Thomson).

top, generally intersected by vertical joints after the manner
of horizontal bedded and compact rocks (see Figs. 126 and
127). 32 Such joints appear indeed in many views and might
perhaps be explained in part by the torsional strains set up
by the tidal movements not unlike those described in the
well-known experiments of Daubree. 33 References to this
jointed structure are, however, seldom met with in the liter-

THE MARGINAL SHELF-ICE

237

ature, but those of the " Challenger " reports are sufficiently
clear: 34

FIG. 127. View of a tilted tabular iceberg showing the rectangular lines of the
plan (after Wyville Thomson).

Nearly all of the flat-topped bergs showed numerous crevasses
in their cliffs near their summits, and these were always widest
towards their summits, and were irregularly perpendicular in gen-
eral direction.

The stratified structure of the bergs is best seen in the case of
the flat-topped rectangular bergs, where an opportunity is afforded
of examining at a corner two vertical cliff faces meeting one an-
other at a right angle.

Cliff surfaces, where freshly fractured, showed an irregular
jointing and cleavage of the entire mass, very like that shown in
a cliff of compact limestone.

Gourdon of the late French expeditions to the Antarctic
refers to such bergs as " absolutely prismatic at their birth." 35

Descriptions of the structure of Southern icebergs have
much in common with those of the Ross Barrier, save only
that they reveal near the bottom especially the presence of
blue ice layers intercalated in the white.

238 CHARACTERISTICS OF EXISTING GLACIERS

According to Arctowski the tabular icebergs which he saw
to the west of West Antarctica are neve near the top, while
the alternate blue and white bands appear only near the
base. Both these latter have the granular structure of neve
ice. 36

Wilkes reported icebergs which were from fifty to two hun-
dred and fifty feet in height with definite strata, of which
thirty were counted in the smaller bergs and eighty in some
of the largest, the average thickness of the layers being about
two feet. 37 Wyville Thomson says of such bergs that " the
entire mass shows a well marked stratification, being com-
posed of alternate layers of white, opaque-looking, and blue,
more compact and transparent ice."

Towards the lower part of the cliffs, the strata are seen to be
extremely fine and closely pressed, whilst they are thicker, with
the blue lines wider apart in proportion as they are traced
toward the summits of the cliff. In the lower regions of the
cliffs the strata are remarkably even and horizontal, whilst
toward the summit, where not subjected to pressure, slight
curvings are to be seen in them corresponding to the inequalities
of the surface and the drifting of the snow. 38

This presence of blue layers was not, however, observed
in the icebergs near the great barrier itself. 39 This, as well
as a thorough study of the barrier edge, makes it probable
that the icebergs studied by Wilkes and Thomson outside
the Arctic Circle were derived from some other masses
of Antarctic shelf-ice, which on the basis of their observa-
tions must contain blue ice layers. The definite separation
of the bergs into thick white layers near the top with thin
intermediate blue layers only, and the concentration of the
latter toward the bottom, where pressure has removed the
air from the more porous white layers, gives the strongest
confirmation to the views of Reid and Hess 40 based upon

THE MARGINAL SHELF-ICE 239

observations on mountain glaciers, that the blue veins sepa-
rate the annual snow deposits of the neve.

Speaking of the stratification in Southern icebergs, von
Drygalski says:

Without doubt it is similar to original neve stratification, only
that this in the South occurs down to the sea level, because no
separation exists between regions of alimentation and removal.
The clear layers are those which for a long time (not necessarily
annual periods) have lain on the surface without new piling up of
the snow. They are either melted by the sun's rays, and thus
hardened, or subjected to pressure and rendered firmer by the
wind. Between them there are more porous layers which appear
as the white ones in the stratification and are characterized by a
greater content of air. 41

To test the different properties of the white and blue por-
tions of a berg, two twelve-pound shots were fired from the
" Challenger/ 7 one at the blue lower layers, and the other
at the white upper zone. The first splintered the relatively
hard and brittle blue ice, leaving conchoidal surfaces, while
the second buried itself in the white porous mass. 42 Frag-
ments of the white layer were taken aboard the " Challenger "
and being subjected to pressure, were found to be easily
deformed, whereas the blue ice, under similar treatment,
did not yield. 43

Southern icebergs of a different type are also formed where
the inland-ice comes directly to the sea, with no intermediate
barrier of shelf-ice, as is the case in Kaiser Wilhelm Land.
The Ross Barrier is not only much the largest well known
mass of shelf-ice, but its edge is more than ten degrees nearer
the pole than those other barriers which have been merely
sighted by navigators. It seems certain that the land of
Wilkes Land is relatively near the barrier edge, and this, as
well as the climatic differences, might perhaps account for
the differences between the icebergs examined by Wilkes and

240 CHARACTERISTICS OF EXISTING GLACIERS

Wyville Thomson, and those which were seen in Ross Sea
and examined by the recent British expeditions. With a
narrower barrier, the local neve of the shelf-ice would be
relatively thin, so that the glacier ice with its blue layers
should be nearer the surface. The studies of Hess appear to
indicate that a differential motion between successive layers
of neve may account for the development of the blue layers
on these planes. 44 There is much need of study of the ice
masses in Wilkes Land in order to clear up the relationships
of the bergs encountered in neighboring seas.

The drift of the bergs which are born of the Ross Barrier
is to the northward, and after passing Cape Adare, to the
westward. 45 The icebergs derived from the barriers of Wilkes
Land are borne to the westward and the northward. When
they have passed the parallel of 65 S. they enter the warm
surface layers of sea water and are, in consequence, more
rapidly melted in the water, at the same time that the
warmer air temperatures reduce their exposed surfaces, trans-
forming them into fantastic groups of towers and minarets. 46

The surpassing beauty of these partially melted icebergs
has been described in picturesque language by Gourdon. 47

Thus in place of the great regular and prismatic tabular bergs
are formed those bizarre and complicated monuments, which recall
the ice bergs of the North : towers, pyramids, bell-towers, cathe-
drals, or palaces, Gothic spires, or Roman porticoes, all styles meet,
all architectures touch elbows; for these are forms more strange
and unexpected than the most capricious imagination could have
dreamed. The whole gamut of blues and greens plays over the
walls of these edifices or within the channelings which course about
them, and the whiteness of the purest marble does not equal theirs.
The transparency of the water permits of following the fairy land
of their azure grottoes far below the surface of the sea. During
the summer, little cascades fall over their sides, mingling their waters
with the waves which break against their glistening flanks ; stalac-
tites hang from cornices and capitals.

THE MARGINAL SHELF-ICE 241

Under the rays of the sun the ice sparkles with the fire of
jewels ; their silhouettes take on life in an atmosphere of extraor-
dinary transparency ; the warmest colorations invade the sky and
are reflected upon the sea, and there are enchanting tableaux
which are offered to the eye.

When, however, the sun disappears from the scene, it is a land
of death which is presented by these mountains of ice. Soon
gathered in great numbers, they resemble the fantastic ruins of a
gigantic marble city ; in a little while and once isolated, they pass,
white phantoms, majestic and silent, into the mystery of the oolar
mist.

Often before this stage has been reached, they have been
deeply tunnelled in sea arches, have been melted unequally,
and have lost some of their stability so as to become tilted
(see Fig. 127, p. 237), or even overturned. 48 Sir John Murray,
who in the " Challenger " had such excellent opportunity
to study floating ice, has said of the melted bergs: 49

Waves dash against the vertical faces of the floating ice-
islands as against a rocky shore, so that at the sea level they are
first cut into ledges and gulleys, and then into caves and caverns
of the most heavenly blue from out of which there comes the re-
sounding roar of the ocean, and into which the snow-white and
other petrels may be seen to wing their way through guards of
soldierlike penguins stationed at the entrances. As these ice-
islands are slowly drifted by wind and current to the north, they
tilt, turn, and sometimes capsize, and then submerged prongs
and spits are thrown high into the air, producing irregular pin-
nacled bergs higher, possibly, than the original table-shaped mass.

Before reaching the 40th parallel of south latitude, the
bergs are entirely dissolved. The tilting and overturning
which they first undergo, permits of an examination of their
under surfaces, and it does not appear that any glacier worn
rock debris has been observed in them. The debris of this
nature observed in the bottom of the blue icebergs described
by von Drygalski and Philippi in Posadowsky Bay, which

242 CHARACTERISTICS OF EXISTING GLACIERS

are of different origin and derived from the true inland-ice,
will be discussed under another section. The fact of impor-
tance is that the white tabular bergs have not as yet revealed
such materials.

REFERENCES

1 O. Nordenskiold, " Einge Beobachtungen iiber Eisformen und Ver-
gletscherung der antarktischen Gebiete," Zeit. f. Gletscherk., vol. 3,
1909, p. 322.

2 H. T. Ferrar, in Scott's " Voyage of the ' Discovery,' " vol. 2, pp. 461-2.
3 Wilkes, "Narrative U. S. Exploring Expedition, 1838-1842," vol. 2,

pp. 350, 365.

4 Brown, et al., "The Voyage of the ' Scotia,' being the record of a voy-
age of exploration in Antarctic seas, by three of the staff." Edinburgh and
London, 1906, p. 236.

5 Borchgrevink, I.e., final map.

6 Scott, "Voyage of the 'Discovery,'" vol. 1, pp. 163-204, map at end
of volume.

7 The Royal Society, National Antarctic Expedition, 1901-1904, Album
of photographs and sketches, London, 1906.

8 T. W. E. David and R. E. Priestley, in App. II of Shackelton's-
"Heart of the Antarctic," vol. 2, p. 288.
9 David and Priestley, I.e.

10 Quoted by Murray, Smithsonian Report for 1893, 1894, p. 358 ; also
ibid., for 1897, 1898, p. 415.

11 Scott, I.e., vol. 2, p. 418.

12 Scott, I.e., vol. 2, p. 420.

13 Scott, I.e., vol. 2, p. 419. David and Priestley, I.e., p. 289.

14 Shackelton, I.e., vol. 2, pp. 12-13. Cf. the moats about nunataks
(ante p. 169 and post p. 257).

15 E. von Drygalski, " Die Bewegung des antarktisches Inlandeises,"
Zeit. f. Gletscherk., vol. 1, 1906-7, pp. 61-65.

16 David and Priestley, I.e., p. 287.

17 Bernacchi, I.e., p. 308.

18 It should be stated that Mr. Bernacchi, an officer of the Scott expedi-
tion, does not accept the view that the Ross Barrier is floating except in
the vicinity of its margin, and, moreover, regards it as fed in the usual
manner of glaciers by material which moves down from the higher
levels along the southern and western margin (Geographical Journal,
vol. 25, 1905, p. 384). Gannett, also, has taken strong exception to the
view of partial surface alimentation as above expressed and as advocated
by Scott, Shackelton, and David (Nat. Geogr. Mag., vol. 21, 1910, pp.
173-174).

19 David and Priestley, I.e., p. 287.

20 Otto Nordenskiold and J. Gunnar Andersson, "Antarctica, or Two
Years amongst the Ice of the South Pole." London, 1905, p. 208.

THE MARGINAL SHELF-ICE 243

21 Otto Nordenskiold, " Die Polarwelt und ihre Nachbarlander," 1909,
pp. 82-84.

22 Nordenskiold and Andersson, " Antarctica," p. 220, and map opposite
p. 316.

23 Otto Nordenskiold, " Einige Beobachtungen, iiber Eisformen und
Vergletscherung der Antarktischen Gebiete," Zeit. f. Gletscherk., vol. 3,
1909, pp. 326-329. See, however, E. Philippi, "Ueber die Landeis-Beo-
bachtungen dej letzen fiinf Siidpolar-Expeditionen," Zeit. f. Gletscherk.,
vol. 2, 1907, pp. 1-21.

24 J. B. Tyrrell, " Ice on Canadian Lakes," Trans. Can. Inst., vol. 9, 1910,
pp. 45 (reprint).

25 E. Philippi, " Ueber die Landeis-Beobachtungen der letzen fiinf Siid-
polar-Expeditionen," Zeit. f. Gletscherk., vol. 2, 1907-1908, pp. 9-11.

26 E. von Drygalski, " Zum Kontinent des eisigen Siidens, etc.," p. 439.
From this view Philippi has strongly dissented (Zeit. f. Gletscherk., I.e.,
p. 10).

27 E. v. Drygalski, "Das Schelfeis der Antarktis am Gaussberg," Sitz-
ungsber. k. bay. Akad. d. Wiss., Math.-phys. KL, 1910, pp. 1-44, pi.

28 David and Priestley, I.e., pp. 283-286.

29 H. Stille, " Geologische Charakterbilder," heft 1, 1910, plates 2-6.

30 Concerning the ice of Antarctic bergs Wilkes has stated that those
encountered along the coast of Wilkes Land varied from a quarter of a
mile to five miles in length (I.e., p. 350). Scott has made mention of a
berg five or six miles in length, and apparently about as wide, but he
states that he saw few which exceeded a mile in length or 150 feet in
height. The highest which he observed was measured as 240 feet (Geogr.
Jour., vol. 26, p. 356). Some of the accounts of bergs of exceptional size
may perhaps be explained by the assemblage of a number closely crowded
together and appearing as one. Such groupings might easily be mistaken
for shelf-ice, and no doubt in some cases have been.

31 Scott, vol. 2, pp. 408-409, pi. opposite p. 408.

32 Shackelton, vol. 2, plate opposite p. 22.

33 Geologie Experimental, 1879, pp. 506-515.

34 Wyville Thomson, "Challenger Report," Narrative, vol. 1, 1865, pt.
I, pp. 431-432, pis. B. C. D.

35 Gourdon, I.e., 1908, p. 133.

36 H. Arctowski, "The Antarctic Voyage of the 'Belgica' during the
years 1897, 1898, and 1899," Geogr. Jour., vol. 18, 1901, p. 374. See also
Pet. Mitt., Erganzungsh., 144, 1903, pp. 15, 19, 21.

87 Wilkes, I.e., p. 253.

38 Wyville Thomson, I.e., pp. 431-432.

39 David and Priestley, I.e., pp. 287-289.

40 H. F. Reid, "The Relations of the blue veins of glaciers to their
stratification," C. R. IX me Congres Geol. Intern., 1903, Vienna, pp. 703-706.
H. Hess, "Die Gletscher," Braunschweig, 1904, pp. 175-178.

41 E. von Drygalski, " Zum Kontinent, etc.," p. 455.

42 Murray, Smithsonian Rept. for 1897, 1898, p. 419.

244 CHARACTERISTICS OF EXISTING GLACIERS

Murray, Smithson. Rept., 1893, 1894, p. 363.
44 Hess, l.b., p. 177.

46 Wilkes, 1.0., pp. 352-353. Scott, I.e., vol. 2, p. 412. Ferrar, I.e., p. 463.

46 Wilkes, I.e., p. 351.

47 Gourdon, I.e., 1908, p. 134.

48 Scott, I.e., pis. opposite pp. 380, 382, 393, 410.

* 'urray, Geogr. Jour., vol. 3. Reprinted in Smithson. Report for
p. 363.

CHAPTER XIV

THE ANTARCTIC CONTINENTAL GLACIER WHERE
UNCONFINED

Inland-ice Margin on Kaiser Wilhelm Land. The Ant-
arctic continental glacier, the great body of ice which is
supposed to occupy the vast central plateau region of the
continent, has been studied in but two districts Victoria
Land and Kaiser Wilhelm Land. 1 Such ice has been more
or less indistinctly seen from the sea at a number of points,
most recently in Coats Land on Weddell Sea by the " Scotia "
expedition. This view is thus described : 2 -

The surface of this great inland ice, of which the barrier was
the terminal face or sea-front, seemed to rise up very gradually
in undulating slopes, and faded away in height and distance into
the sky, though in one place there appeared to be the outline of
distant hills: if so, they were entirely ice-covered, no naked rock
being visible.

The ice here reached the sea in a narrow barrier with cliff
one hundred to one hundred and fifty feet high, while off its
edge the sea was found to have a depth of 940 feet. 3

It is this type of inland-ice not confined by an encircling
mountain rampart which was studied within a very narrow
marginal zone by the German Antarctic expedition of
1901-1903. 4

As seen from the sea, " it was beyond a doubt that the ice

245

246 CHARACTERISTICS OF EXISTING GLACIERS

all lay upon land, for one could see dark fissures in its surface
arranged in different systems. Everywhere this inland-
ice ended at the sea in a steep edge 40 to 50 metres in height.
The surfaces behind it might rise to 300 metres, but soon
graded over into flat slopes so that one could not see the
end." 5 (See Fig. 128 and plate 32.)

Of all the Antarctic inland-ice areas studied, 6 this seems to
be the only one which furnishes a parallel to the continental
glaciers which in Pleistocene times existed in North America
and in Northern Europe. In all other cases a rampart of
mountains encloses and materially modifies the physiography
of the ice surface. It is, therefore, much to be regretted that
we have no profile across its surface.

The crests upon the horizon of the inland-ice of Kaiser
Wilhelm Land appeared not straight, but gently undulating.
It was, therefore, concluded that the land beneath possesses

a similarly undulat-
ing character.
Near the ice margin,
which was a cliff
130 to 165 feet
high, soundings
made through the
neighboring sea-ice
gave depths ranging
from 550 to 810
feet, the greater
depths lying to the westward. If only four-fifths of the
ice is below sea level, the inland-ice of Kaiser Wilhelm
Land must be aground nearly, if not quite to its edge.
This is proven by the existence of a tide crack, which runs
along the front and upon which the sea-ice moves up and
down.
The convexly curving surfaces of the marginal zone of the

FIG. 128. The inland-ice of Kaiser Wilhelm Land
(after von Drygalski).

PLATE 32.

CONTINENTAL GLACIER WHERE UNCONFINED 247

inland-ice are thus in sufficiently striking contrast with the
horizontal top so characteristic of shelf-ice; but a no less
noteworthy difference is found in the colors. Even from a
great distance, the beautiful blue color of the inland-ice is
noticeable, whereas the shelf-ice of the Ross Barrier is daz-
zling white. The blue color of the inland-ice shows that its
surface is in general free from snow, and this appears to be
characteristic of it during both winter and summer. Under
the strong easterly winds which prevail, the snow falling
upon its surface is able
to find a lodgment only
within the fissures and
in the lee of the Gauss-
berg.

That the inland-ice is
moving forward is suf-
ficiently clear from the
existence of great gap-
ing fissures observed
from considerable dis-
tances. These are par-

, , . . 100 200 joo

ticularly prominent in ***.,-*

the Step-like terraces Of FIG. 129. Intersecting series of fissures in the
tVP npnr mnro-in nnr surface of the inland-ice to the west of the

the near-margin por Gaussberg (after von Drygalski) .
tions, and the ice shows

bucklings in the rear of them. Crevasses very generally
appear upon the surface in parallel series, and sometimes
two such series intersect each other at right angles (see
Fig. 129).

Such fissures were sometimes seen as they opened to the
accompaniment of rumbling reverberations, 7 and, in general,
their directions seemed to correspond to local disturbances
above buried projections of the floor, or else to the strains set
up due to general movement. The effect of the obstruction

248 CHARACTERISTICS OF EXISTING GLACIERS

of the Gaussberg in the path of the moving ice, was visible
in local fissures developed within its neighborhood. 8

Measurements of the rate of movement within the ice
were made during a period of five months, and showed that
at its margin the inland-ice moved forward at the remark-
ably uniform rate of about a foot per day. At a distance
of two kilometres back from its margin, this rate had
fallen off by 1|- inches. 9 In spite of this, the aspect of the
ice front was, in general, one of rest. No evidence of
push was observed along its base. 10 In the vicinity of the
only exposed land, the Gaussberg, the ice surface is lowered
within a broad encircling zone due to the greater ablation in
consequence of heat radiation from the rock surfaces (see
plate 33 A). Here the stratification within the ice is made
apparent by lines upon the surface, though elsewhere the
only traces of banding are to be observed in fissures. On the
surface of the bands were found the indications of " cryaco-
nite " wells and water basins, no doubt from dust blown
from the slopes of the Gaussberg.

It has been stated that the strong easterly winds suffice
to keep the surface of the inland-ice swept of snow, with the
exception of specially protected places such as the lee of the
Gaussberg. Thus though the snow fall is heavy, the evi-
dence showed that instead of increasing its thickness, the
inland-ice surface is being constantly lowered, and thus
confirms from a new region the many indications that the
present is included in a receding hemicycle of glaciation.
During five winter months the ice surface was found to have
lowered through ablation by about 4 centimeters (If inches).

The Blue Icebergs of Antarctica. In front of the inland-
ice of Kaiser Wilhelm Land prodigious fragments of the
continental glacier were found ranged in series more or less
parallel and separated only by narrow lanes (Gassen).
Farther out from the margin the bergs became less numerous

CONTINENTAL GLACIER WHERE UNCONFINED 249

and eventually they were more scattered and more or less
promiscuously frozen into the surface of the sea-ice.

From the typical tabular bergs of the Antarctic seas, these
differ strikingly in their beautiful blue color as well as in
their rounded contours. In Posadowsky Bay where they
are frozen into the sea-ice, they could be studied to advan-
tage. Their surfaces were found to be intersected by broad
furrows which were steep on one side only, and smoothly
polished upon the other. 11 The rounding of the angles is a
result of filing off the surface by hard snow, driven by the
storm winds. 12

These blue bergs reveal, especially at their bases, bands of
rock debris which must be regarded as portions of the
ground moraine which have been raised upon a subglacial
obstruction, 13 as has been shown to be characteristic of the
margins of the Greenland continental glacier. The rock
debris is here generally found in layers more or less parallel
to the blue ice strata. The individual rock fragments are
sometimes angular with a single scratched " sole " cut upon
the surface. In other specimens there are several facets,
or the block may be entirely covered with such smoothed
and striated surfaces.

Professor von Drygalski, in his classification of Antarctic
icebergs, has expressed his belief that the blue bergs arise
from the common tabular bergs through the action of the
wind driven snow, aided by evaporation. 14 The tabular
bergs he believes, further, are derived from the margin of
the inland-ice. This relationship to the usual tabular bergs
it is especially difficult to accept, since the blue bergs are
found mainly in contact with the inland-ice and near the
shore, and are further characterized by the same colors and
structures, whereas the usual tabular bergs seem to have
more the properties of shelf-ice, though a portion only have
the absolute uniformity of texture found in the best known
example of shelf-ice, the Ross Barrier (see ante, p. 239).

250 CHARACTERISTICS OF EXISTING GLACIERS

Origin of the West-ice. The peculiar labyrinthine surface
of the West-ice, and its resemblance in places to a jam of
blue bergs, as has been pointed out by Drygalski, in the
writer's opinion, permits of an explanation of this mass of
shelf-ice which is quite in harmony with the views of the
British and Swedish explorers concerning the origin of shelf-
ice in general. As von Drygalski has stated, the inland-ice
surface of Kaiser Wilhelm Land, is swept clear of snow by the
easterly storm winds, the sweepings finding lodgment only
in fissures and protected places. A crowded fleet of blue ice-
bergs massed upon the western or lee shore of Posadowsky
Bay would have furnished the narrow lanes within which
the snow could find lodgment. Still further to the west,
the intervening spaces would have been levelled up with the
tops, and thus a relatively even surface would result.

If cumulative loading of sea-ice by snow is to be assigned
as at least one cause of the formation of shelf-ice, as seems
now quite generally believed, it is evident that this process
cannot go on where sea-ice is annually broken up and car-
ried northward with the ice pack. The essential condition
for its formation is, therefore, an area within which the sea-ice
either attains a greater thickness, or is so protected by the
shores, that snow accumulates upon it from year to year. Now
it is worthy of note that the three great areas where shelf-ice
has thus far been studied have all this character in common.
The Ross Barrier is firmly wedged in Ross Sea between
Victoria and Edward VII Land. The " terrace " of West
Antarctica is held by the southeasterly storms against the
west shore of a great gulf, and has crowded against the
hook-like peninsula of West Antarctica. The West-ice of
Kaiser Wilhelm Land is similarly developed upon the western
or lee side of Posadowsky Bay, and its growth has been appar-
ently facilitated by the assembling of a fleet of icebergs to
collect the snow swept from the vast surface of the inland-

CONTINENTAL GLACIER WHERE UNCONFINED 251

ice to the south and east. The British expeditions to Vic-
toria Land have shown that vast quantities of snow blow off
the barrier into the sea, and the collection of snow upon the
northern side of the Nordenskiold shelf-ice tongue is most
illuminating in this connection (see ante, p. 234).

But additional evidence of this essential condition for the
formation of shelf-ice, has been furnished by the recent
French explorations. Gourdon has shown that near West
Antarctica the normal winter's thickness of field ice is only
about 16 inches, whereas in the sheltered upper end of
Flanders Bay it had reached a thickness of between four
and five metres (13 to 16 feet). Soft snow here lay upon the
surface, with stratified neve below and compact ice at the
bottom. At the margin of this terrace, which rose to a
height of about a metre above the sea, immense rafts resem-
bling in form, tabular icebergs were from time to time (in
February) detached on long rectilinear cracks intersecting
the terrace. 15

In connection with the latest French expedition to the
Antarctic, the great newly discovered bight which has been
named Marguerite Bay, and which is sheltered behind
Alexander Island, was found to have a similarly heavy cover
of field ice reaching a thickness in this instance, of from 2
to 3 metres. The separation of ice blocks or overgrown
floes from the margin, took place with fragmentation and
apparently quite resembled the calving of tabular icebergs.
Thus from the meagre reports of these late expeditions which
have been published, it would appear that the intermediate
stages in the transformation of field ice to shelf ice by accre-
tion of surface snow are fast being supplied. 16

REFERENCES

1 The small ice-cap on Louis Philippi Land in Northern West Ant-
arctica was in 1902 and 1903 crossed by Andersson and Duse from Hope
Bay to Erebus and Terror Gulf, a distance of about 22 miles (Norden-

252 CHARACTERISTICS OF EXISTING GLACIERS

skiold and Andersson, "Antarctica," 1905). The "upper terrace" which
was just reached by Nordenskiold near King Oscar Land probably rep-
resents inland ice.

2 Brown, et al. t "The voyage of the 'Scotia/ etc.," 1906, p. 236.

3 E. Philippi, I.e., p. 11.

4 E. von Drygalski, " Zum Kontinent des eisigen Siidens, Deutsche
Siidpolarexpedition, Fahrten und Forschungen des 'Gauss,' 1901-1903,"
Berlin, 1904, pp. 668, 21 pis. and maps and 400 cuts. E. Philippi, " Ueber
die Landeis-Beobachtungen der letzen fiinf. Siidpolar-Expeditionen,"
Zeit. f. Gletscherk., vol. 2, 1907, pp. 6-8.

6 E. von Drygalski, I.e., p. 241.

6 The "upper terrace" off King Oscar Land may prove to be another
instance.

7 Cf. Peary, ante p. 129.

8 E. von Drygalski, " Deutsche Siidpolar-Expedition, 1901-1903," vol. 2,
heft I, legends of plates.

9 E. von Drygalski, " Die Bewegung des Antarktischen Inlandeises,"
Zeit. f. Gletscherk., vol. 1, 1906-1907, pp. 61-65.

10 E. von Drygalski, " Zum Kontinent, etc.," p. 305.

11 This, it will be remembered, was a characteristic structure upon the
surface of the west-ice.

12 E. von Drygalski, " Zum Kontinent, etc.," p. 305. Also Philippi, I.e.,
p. 10.

13 Philippi, I.e., p. 11.

14 E. v. Drygalski, Sitzungsber. bay. Akad. Wiss., Math.-Phys. KL, 1910,
pp. 10-13 (reprint).

16 Gourdon, Exp. Ant. Franc., 1903-1905, 1908, p. 125.
16 Charcot, " Rapports preliminaires, etc.," 1910, p. 51.

CHAPTER XV

THE ANTARCTIC CONTINENTAL GLACIER WHERE BEHIND
A MOUNTAIN RAMPART

Inland-ice in Victoria Land. The inland-ice of Victoria
Land unlike that in Kaiser Wilhelm Land is held within an
encircling rampart of high mountains the Admiralty,
Prince Albert, Queen Alexandra, and smaller ranges joined
one to the other in a long chain. Its physiography is, there-
fore, in important respects different from that just described.
Held back by the ranges, as by a gigantic retaining wall, the
inland-ice now finds an outlet through somewhat widely
separated and relatively narrow portals. These gateways
have already been referred to as outlets, and to the south-
ward they are, so far as known, the Skelton, Mulock, Barne,
and Shackleton Inlets and the Beardmore Glacier (see
Fig. 134, p. 258) ; while farther to the north, are the Reeves,

GEOGRAPHICAL -MILES

FIG. 130. Section across the margin of the inland-ice of Victoria Land in a direc-
tion westward from McMurdo Sound (from data of the Scott expedition) .

David, and Ferrar Glaciers. Some of the latter are, how-
ever, apparently no longer in service as outlets for the ice.
Marginal Cross Sections of the Inland-ice along the Out-
lets. Thanks to the plucky efforts of British explorers, we

253

254 CHARACTERISTICS OF EXISTING GLACIERS

are fortunate in having no less than three sections across the
margins of the inland-ice. These are on the lines of the
Beardmore and Ferrar Glaciers, and between the David and
Reeves Glaciers the Backstairs Passage. The earliest of
these was made by Scott on an east and west line westward
from McMurdo Sound and up the Ferrar outlet. Later,
Shackleton made his section more nearly upon a north and
south line up the Beardmore outlet and toward the South

Great Ice Barrier

83 82

tat.72-25'i long. 155* 16'

iwrticai Scale of Feet

Sta Level

FIG. 131 (a and 6). Section across the Great Ross Barrier and up the Beardmore
outlet to and upon the ice plateau of Central Antarctica (after Shackleton, but
with barrier portion added). The part b should be joined to the right of a.

(c) Section across the Drygalski ice barrier tongue and up the Backstairs Pas-
sage on to the inland-ice in a direction toward the south magnetic pole (after
David).

Pole; while David of the Shackleton party travelled in a
direction nearly northwest from a point in latitude 75 S.
up toward the south magnetic pole upon the plateau (see
Fig. 130 and Fig. 131, a, b, and c). 1

While not large when compared to the Ross Barrier to
which it contributes its ice, the great Beardmore outlet com-
pared with other streams of ice is by far the greatest known.
On the map of Fig. 134, p. 258, we have added the Great

CONTINENTAL GLACIER BEHIND A RAMPART 255

Aletsch Glacier of Switzerland drawn to the same scale in

order to bring out this

contrast. The area of
the Beardmore outlet
is in excess of 5000
square miles, its width
is in places about 50
miles, and its fall
about 6000 feet with
an average of 60 feet
to the mile through-
out its entire length.
Its surface showed
every variety from
soft snow to cracked
blue ice. Crevasses
were everywhere, some
of them descending to
hundreds and perhaps
even a thousand feet.
An ancient medial
moraine was largely
buried beneath its
surface.

All the sections up
the outlets show in
common a steep slope

near the margin and FlG> 132 . _ Comparison of sections across the

gentler grades above;
as was found to be
characteristic also of
the margins of the
Greenland continental
glacier an ice body

margins of the continental glaciers of Green-
land and Antarctica. (a) West Greenland
(Peary) ; (6) West Greenland (Nordenskiold) ;
(c) Southwest Greenland (Nansen) ; (d) South-
east Greenland (Nansen) ; (e) South Greenland
(Garde) ; (/) Victoria Land west of McMurdo
Sound (Scott) ; (g) South Victoria Land
(Shackelton) ; (h) North Victoria Land, Ant-
arctica (David).

256 CHARACTERISTICS OF EXISTING GLACIERS

similarly held in by a wall of mountains (see Fig. 132).
The Antarctic sections are like them also in the step-like
alternation of steeper and more gradual slopes in the vicin-
ity of the margins, the steeper slopes being deeply crevassed
in a transverse direction.

Once the plateau surface has finally been reached, all
sections are alike in the monotony of the surface, no irregu-
larities in excess of a hundred feet being encountered. All
irregularities of surface are here due to sastrugi deposited and
later cut out by the fierce blizzards from the surface of the
plateau. The plateau is, however, in no case reached when
the mountain rampart has been passed, though the surface
slopes continue to become more gradual. Within its retain-
ing mountain wall the surface of the inland-ice is, therefore,
flatly domed or shieldlike. It is the same in general form,
though on a vastly grander scale, as the domed ice islands
encountered off the coast (Fig. 110 and pi. 29 A). Above the
great Beardmore outlet the surface continues to rise in ter-
races with crevasses upon the steeper slope, and these very
properly led Shackleton to the belief that the ice here rests
in moderate thickness upon a steeply sloping floor. The

FIG. 133. View from above the Ferrar outlet looking from the inland-ice toward
the outlet and showing the dip of the surface produced, by the indraught of the
ice (after Scott).

ridges rise abruptly with a great crevasse at the top of each
and a descent upon the other side of perhaps fifty feet on a
grade of one in three. Then smaller ridges and new waves
of pressure ice are encountered, the undulations of the first

CONTINENTAL GLACIER BEHIND A RAMPART 257

order of magnitude being separated by an interval of about
seven miles.

Dimples upon the Ice Surface above the Outlets. In

Greenland it was found that the tongues of ice which push
out through gaps in the mountain wall, the outlets, show
above them a dimple in the surface caused by the indraught
of the ice from the near-lying portions of the plateau. 2 The
same characters pertain to the ice of Victoria Land. 3 A
photograph by Scott looking back on the line of his route
toward the outlet up which he had come, shows this very
clearly. In Fig. 133, drawn from his view, we have added
a dashed line to bring out the dip in the ice surface. Farther
down the outlet this concavity of the ice surface obtains, 4
but in the lower reaches it changes to a convexly moulded
form. Upon the Beardmore outlet this concavity of the
surface with tendency to form an amphitheatre above is
brought out upon Shackleton's map (see fig. 134).

Above the Ferrar Outlet Scott found the transition from
the neve surface of the plateau to the outlet ice not a gradual
one, but abrupt, the outlet having a corrugated surface of
massive blue ice. The surface of the plateau ice is every-
where carved by the wind drift to form sastrugi of erosion
which will be discussed in connection with the winds of the
plateau.

Ice Aprons below Outlets. At the foot of each active
outlet, the ice is discharged upon the shelf-ice in an ice apron
which spreads out laterally as well as in front (see fig. 134).
In front of the Beardmore outlet this apron rises to a height
near its medial line of between 400 and 500 feet above the
general level of the barrier surface.

Moats about Rock Masses. About the continental glacier
of Greenland wherever the rock projects through its surface,
local melting results from heat radiation from the rock sur-
face (see p. 169). Except when filled in with drifted snow

258 CHARACTERISTICS OF EXISTING GLACIERS

FIG. 134. Map of the Beardmore Outlet from the inland-ice of Antarctica to the
Ross Barrier. Note the dimple at head and ice apron at foot (after Shackleton).
In the same scale the Great Aletsch glacier (A) is added.

PLATE 33.

A. The Gaussberg of Kaiser Wilhelm Land with ice surface depressed about it (after v.

Drygalski).

B. Moat surrounding projecting rock mass. Inland-ice of Victoria Land (after Scott),

CONTINENTAL GLACIER BEHIND A RAMPART 259

or when ice pressure has closed this gap, a deep and narrow
depression surrounds the rock mass. This has been
designated a moat from its resemblance to the moat sur-
rounding a castle. 5 The same holds true of Victoria Land,
where the wall of mountain ranges offers essentially the
same conditions (see Plate 33 B).

Mountain Glaciers on Outer Slope of the Retaining Ranges.
- Many of the well-known types of mountain glaciers 6 are
found in numbers on the outer slope of the mountain wall
which hems in the continental glacier of Victoria Land.
The scale of these is large, like everything in the region, but
the internal movement in most cases is slight, and the larger
number are in a relatively stagnant condition. The Blue
Glacier, which starts in the Royal Society Range, is cited by
Ferrar as an example of the Norwegian ice cap type, 7 and ends
at tide water in a cliff between 70 and 80 feet high on Mc-
Murdo Sound. Glaciers of the Alpine type occur according
to the same authority in great profusion in the Royal Society
Range. With them are associated horse-shoe or corrie
glaciers, and these are especially well represented at the foot
of the Inland Forts. The presence of the cirques here as
everywhere calls attention to the peculiar eroding process
which distinguishes mountain glaciers from their mightier
neighbors of the continental type.

There is but little superglacial material upon the surface of
the mountain glaciers, the lateral and medial moraines of the
Ferrar glacier being merely long lines of large stones with
very little finer material. About the margins, however, mo-
raines are found near the bottom intercalated with blue ice,
and at one point englacial rock debris was seen pushed up to
form a surface moraine where two glaciers meet coming from
opposite directions.

Ice Slabs. Glaciers of an essentially new type, which
Ferrar has referred to as ice slabs, are in the cases which he

260 CHARACTERISTICS OF EXISTING GLACIERS

cites masses of ice about 50 feet in thickness and from 4 to 6
square miles in area. These appear to be the dead aprons of
true piedmont glaciers, from which the feeders have disap-
peared. They offer the most striking illustration of the ne-
cessity for some modification of our views concerning the
nourishment and ablation of glaciers in polar latitudes, where,
as we shall see, wind may be a far more important factor
than temperature, and where melting occurs only under
special conditions. These ice slabs seem clearly to be the
relics of piedmonts retained during a receding hemicycle of
glaciation. Were the changes in their form and size due ta
a general rise of the air temperature only, since the slabs are
in the lower levels, they should first disappear; but where
there is practically no melting at all save in the vicinity of
exposed rock surfaces, or in connection with strong local
foehn winds, the relatively narrow and tributary ice streams,,
surrounded as they are on all sides by rock, would probably
be the first to disappear. The mountain valleys, moreover,
are the natural channels of the hot and dry foehn winds. 8 If
formed below true outlets from the inland ice, the recession
of the parent ice mass would alone suffice to explain them,,
since this might cut off their nourishment.

REFERENCES

1 R. F. Scott, " The Voyage of the ' Discovery,' " 2 vols., 1905. E. H.
Shackleton, " The Heart of the Antarctic," 2 vols., 1910.

2 Hobbs, Proc. Am. Phil. Soc., vol. 49, 1910, pp. 87-90.

3 Scott, I.e., voL 2, pi. opp. p. 240, lower view.

4 Scott, I.e., pi. opp. p. 224, upper view.

5 Proc. Am. Phil. Soc., vol. 49, p. 117.

6 Geogr. Jour., vol. 35, 1910, pp. 147-148.

7 Ferrar, I.e., pp. 462-463.

8 David, I.e., p. 151.

CHAPTER XVI

CLIMATIC CONDITIONS WHICH AFFECT THE NOURISH-
MENT OF ANTARCTIC ICE MASSES

The Greenland Ice in its Relation to the Antarctic Conti-
nental Glacier. The conditions of climate which determine
the alimentation of ice masses within polar regions are not
identical with those which have been worked out from study
of the local mountain glaciers which now exist in low lati-
tudes. This has already been pointed out for the great con-
tinental glacier of Greenland. 1 It was found that the vast
surface of the ice dome of Greenland controls the air circula-
tion above it ; and we see that in proportion to their dimen-
sions great continental glaciers must exert larger and larger
effects, and even, it may be, eventually put a limit upon their
own extension.

The continent of Greenland is not located in proximity to
the pole, and hence we are able the better to assert that the
conditions which we there find are not explained by the plan-
etary system of the winds. This is fortunate for our study of
the Antarctic region, since there an elevated plateau is more
nearly centred above the earth's axis, and more or less un-
known and mysterious causes might otherwise be invoked
to explain a system of circulation and a climatic condition
which in most respects are identical with those worked out
with some care for the northern continental glacier. For

261

262 CHARACTERISTICS OF EXISTING GLACIERS

these reasons the northern ice masses should be studied first,
because of the light which they throw upon the problems of
Antarctic glaciation.

Air Temperatures, Humidity, and Insolation. The Antarc-
tic, as already pointed out in an earlier section, is in contrast
to Arctic regions, characterized by greater severity of climate,
by lower average annual temperatures, and by less tempera-
ture range between winter and summer (ante, p. 188). Over
the Greenland ice the relative humidity of the air is extremely
high, while its absolute humidity, because of low temperature,
is very low. Even on the margins of the Antarctic continent
this was early proved to hold true. On between thirty and
forty per cent of the days that he was south of the parallel of
60 S., Ross found the air completely saturated with moisture.
Racovitza reports from the " Belgica " expedition that the
air was almost constantly saturated with water vapor. 2

Insolation, or solar radiation, on the borders of Antarctica
is during the summer months considerable. On the
" Belgica " expedition at the end of December, Racovitza
found that a black-bulb thermometer when exposed in the
sun registered 113.2 F., when the air temperature was only
31.6 F., although this effect was hardly felt upon the ice
pack. 3 Bernacchi with a black-bulb thermometer frequently
obtained at Cape Adare temperatures above 80 F., while
temperatures in the shade were below the freezing point. 4

Nature of the Snow Precipitated in Antarctica. Rain is
unknown on the Antarctic continent and most of the snow is
precipitated in spring and summer 5 and at relatively low tem-
peratures. Observations made in Victoria Land have shown
that if precipitated near the freezing point, the snow is of one
or the other of two types. With a sudden fall of temperature
the sago or tapioca snow was precipitated in the form of felted
spheres one-tenth of an inch or more in diameter. Other-
wise large six-rayed feathery flakes similar to those formed in

NOURISHMENT OF ANTARCTIC ICE MASSES 263

warmer climates resulted. In colder temperatures the air
became filled with minute ice crystals which were only one
one-hundredth of an inch in diameter and descended from a
cloudless sky. 6 This form of snow thus resembles the " frost
snow " described by Nansen as characteristic of the ice
plateau of Greenland.

When the softer snow falls in summer time, if the weather
becomes colder, the snow compacts itself and becomes hard.
Such superficial hardening yields a " pie-crust " surface and
the snow below is soon firmly bound together so as to yield
the usual " smooth-sledging type of winter snow-ice."

Upon the plateau a surface of somewhat different character
is produced when solar radiation on quiet summer days has
melted a thin superficial layer of the snow. Under such con-
ditions large and beautiful reconstructed ice crystals develop
which are about one-half inch across and one-sixteenth of an
inch in thickness. These develop throughout a layer extend-
ing about one-half an inch below the surface. Covered with
these sheets of brightly reflecting ice crystals, the snow sur-
face glitters " like a sea of diamonds." 7 The heavy sledge
runners rustle as they crush the crystals by the thousand.
With the first strong wind these crystals are picked up and
drifted away, the sastrugi in consequence exhibiting such
scaly crystals on their lee sides, whereas the windward
surfaces are much eroded and furrowed by the wind. 8

Off Kaiser Wilhelm Land it was noticed that late in August
when the actinometer had risen for the first time above freez-
ing, traces of fusion began to appear upon the snow surface.
These consisted in a smoothening and hardening of the sur-
face, and in the development of sublimed crystals beneath
the crust. This last feature was, moreover, not found in
those crusts which were formed by wind pressure and were
observed alike upon the north and south sides of drifts. 9 The
amount of the annual snow fall above the Great Barrier has

264 CHARACTERISTICS OF EXISTING GLACIERS

already been discussed (see ante, p. 223). Upon the plateau
snow is carried in the air and was observed to within 110
miles of the pole. It is significant that the snow comes
mainly in the summer time and invariably from the south
or southwest in connection with the peculiar blizzards. On
several occasions it was observed that whereas in the earlier
part of the blizzard the snow was largely redistributed snow
in the form of drift, a new fallen snow appeared near the end
accompanied by a rise of temperature (see below, p. 269). 10

Winds upon the Continental Margins. Throughout the
margins of the Antarctic regions the general direction of
strong surface winds seems to be within the quadrant be-
tween south and east. In this Wilkes, Ross, Wyville Thom-
son, and other navigators of the far southern seas are in
agreement. The zone of prevailing westerlies travelling
southward with 'the sun may, however, at points near the Ant-
arctic Circle sometimes bring about a partial seasonal reversal
of this wind direction. Thus Arctowski reported high air
pressure at the solstices and low pressure at the equinoxes to
the westward of West Antarctica, with easterly winds pre-
dominating over westerly, but with a relative high frequency
of the latter in the winter months. 11

At Cape Adare in Victoria Land the prevailing winds were
found by^Bernacchi to be from the east-southeast and south-
east in a very marked degree. Measured in observation
hours the calms were, however, even more important, there
being 1033 hours of calms, 973 hours of winds from the south-
easterly quadrant, and only 275 hours of winds from all
westerly points whatsoever. 12

At Cape Royds on McMurdo Sound, where Scott 's expedi-
tion wintered, the winds were much modified by local condi-
tions, but were generally from the east or southeast. No
winds, but only light airs, came from the west or northwest.
Blizzards invariably came from the south or southwest. 13

NOURISHMENT OF ANTARCTIC ICE MASSES 265

The Shackleton party in almost the same locality reports
either gentle northerly winds whose velocity seldom exceeded
twelve miles an hour, or winds from the south-southeast or
southwest, the latter ranging from gentle breezes to fierce
blizzards. A northwesterly wind was rare. 14

In Kaiser Wilhelm Land an absolute rule of easterly winds
was observed, and gales from the southeast kept the surface
of the inland-ice swept clear of snow. 15

The Antarctic Continental (Glacial) Anticyclone. The
prevalence of southeasterly winds about the borders of the
Antarctic continent finds its only explanation in the exist-
ence of an area of high atmospheric pressure above the
continent. Sir James Ross, as long ago as 1840, obtained
increasingly high atmospheric pressures in his cruise south-
ward in the Ross Sea. A South Polar anticyclone was as
early as 1893 declared to exist by Sir John Murray in a paper
read before the Royal Geographical Society and printed in the
Geographical Journal. 16 Unfortunately the theory of polar
eddies promulgated by Ferrel 17 and adopted by Davis in his
in many respects excellent treatise 18 is responsible for a
general prevalence of incorrect views concerning the winds
of both the earth's polar regions. As pointed out by
Buchan in 1898, the low pressures required by this theory
do not exist, and in place of the supposed northwesterly
winds blowing homeward toward the poles, as required by
the theory, we find in the Antarctic southeasterly and east-
erly ones. 19 If Ferrel's theory were correct, Antarctica
would be a land of rain and fog instead of what it is known
to be.

Bernacchi in 1901, as a result of his very important mete-
orological studies in connection with the " Belgica " expedi-
tion, set forth the evidence for the Antarctic anticyclone in
a most convincing manner. Speaking of the prevailing
southeasterly winds of the region, he says:

266 CHARACTERISTICS OF EXISTING GLACIERS

Their frequency and force, the persistency with which they
blow from the same direction, the invariable high rise in the tem-
perature, their dryness, the motion of the upper clouds from the
N.W., and, finally, the gradual rise in the mean height of the
barometer to the south of about latitude 73 S., seem to indicate
that the Antarctic lands are covered by what may be regarded
practically as a great permanent anticyclone, with a higher press-
ure than prevails over the open ocean to the northward. 20

The complete verification of the existence of an Antarctic
anticyclone has, however, been furnished by Shackleton,
who in his journey across the ice plateau to within one hun-
dred and ten miles of the earth's southern pole has brought
back the knowledge that throughout the entire distance the
winds blew strongly nearly all the time from the south or
southeast with an occasional change to the southwest, and
that all sastrugi pointed to the southward. 21

Wind Direction determined by Snow-ice Slope. It is
the author's belief that over the Antarctic continent this
anticyclonic circulation of the air is not determined in any
sense by latitudes, but is a consequence of air refrigeration
through contact with the elevated snow-ice dome, thus
causing air to slide off in all directions along the steepest
gradients. For the continent of Greenland this has now been
fully demonstrated through the work of several observers,
but especially of Commander Peary; 22 and there is every
reason to think that the conditions in Antarctica are essen-
tially the same. Upon this supposition the prevalent winds
and the strongly marked sastrugi which were observed by
Scott, Shackleton, and David upon the ice plateau, find a
simple explanation. The strikingly local character of the
winds about the margins of the great Ross Barrier are on
this assumption likewise accounted for.

As already stated, Shackleton encountered on his journey
of about two hundred miles across the ice plateau strong

NOURISHMENT OF ANTARCTIC ICE MASSES 267

winds blowing only from the southerly quarter, and the
sastrugi showed this direction only. Scott on his plateau
journey westward for about two hundred miles from
McMurdo Sound, in a latitude eight to ten degrees lower,
likewise encountered winds of constant direction here from
the west-southwest, and a single set of sastrugi with direc-
tions varying only between west by south and southwest
by west. Eight to ten degrees farther north, and upon what
now appears to be a relatively narrow peninsula of the con-

539

^ c

FIG. 135. 'Sketch map showing directions of the sastrugi along the line of David's
course to the south magnetic pole. The direction of the arrows indicates the
direction of the wind as evidenced by the sastrugi (based on Shackleton's map
and David's narrative).

tinent, David for the first time found variable wind direc-
tions and several sets of sastrugi. A more careful examina-

268 CHARACTERISTICS OF EXISTING GLACIERS

tion of his data confirms the view that the air currents upon
the peninsula are determined wholly by the direction of
snow slope upon the plateau, as is apparent from Fig. 135.
Off the coast the sastrugi betrayed the evidence of the strong
southerly blizzards and also of winds which blew down from
the plateau through the portals of the outlets. Until the
highest point of the plateau had been reached, winds and
sastrugi alike indicated a sliding down on the slopes toward
the coast. On January llth, when there was " no appreciable
general up-grade now/' it was noticed that the sastrugi
" had now changed direction, and instead of trending from
nearly west or north of west, eastwards, now came more from
the southeast directed towards the northwest.' 7 To the
west of the summit as shown by the map, the sastrugi point
southeastward, indicating that the shore line doubtless
continues its direction to the westward from Cape North.
Returning from the South magnetic pole toward the crest
of the dome, David states, " We had seen from the evidence
of the large sastrugi that blizzards of great violence must
occasionally blow in these quarters, and from the direction of
the sastrugi during our last few days' march, it was clear that
the dominant direction of the blizzard would be exactly in
our teeth." 23

The Foehn Blizzard of the Ice Plateau. Next to the obser-
vation that the prevailing winds blow outward from the
interior of the continent, the nature of the winds themselves
is most characteristic of an anticyclone developed above an
ice plateau as we have become familiar with it in Greenland.
In Victoria Land these winds sometimes blow with a violence
of seventy to eighty-five or more miles per hour, and are prob-
ably the most violent that are anywhere known. The
summer blizzard lasting for three days, which Shackleton
encountered near his farthest south and at an elevation in
excess of 10,000 feet may be cited as an example. 24

NOURISHMENT OF ANTARCTIC ICE MASSES 269

Despite the fact that these blizzards in summer at least
appear to bring snow, the wind may be described as dry.
Though at first cold, and, in fact, having its origin, it would
appear, in a general lowering of the temperature during a
period of calm, in a later stage the temperature rapidly rises,
due to the foehn effect. In Victoria Land an increase of
as much as 45 F. has been observed to take place within
twenty-four hours.

The sequence of events during a blizzard begins with gentle
northerly winds which continue for a day or two, during
which temperatures are low. David has suggested that dur-
ing this time air is flowing south to take the place of air
whose volume has been reduced as a result of the heat ab-
stracted from it on the ice surface. Then there follow two
or three days of absolute calm, during which the temperature
continues to fall. Still further cooled upon the ice surface,
the air, a week or more after the calm begins, starts to move
outward in all directions and so develops (on the edge of the
barrier) a southeasterly blizzard. Simultaneously with this
movement the steam cap over the volcano of Erebus, which
normally indicates an upper current from the northwest,
swings round to the north and takes on an accelerated move-
ment, as though it were being drawn from that direction to
supply air to the void resulting from the violent surface cur-
rent toward that direction. Corresponding to the increased
velocity, the normal foehn effect near the pole must be much
increased, as it is also on the descent of the surface current
from the plateau. As soon as the warming of the polar
air from this cause has become general, the high air pressure
of the central area is automatically reduced, and thus the
blizzard gradually brings about its own extinction. To the
warming effect of the descending air current there is rather
suddenly added the latent heat of condensation of the
moisture when it is precipitated in the form of fine ice

270 CHARACTERISTICS OF EXISTING GLACIERS

crystals within the air layers just above the snow-ice sur-
face. The rather sudden termination of the blizzard may
be thus in part explained. David has suggested that a
" hydraulic ram effect " may be induced in the air of the
upper currents, since the steam clouds over Erebus, nor-
mally the antitrades, are temporarily reversed in direction
at the termination of a blizzard, and for a short interval
blow northward. 26

Foehn winds were experienced by the German expedition
off Kaiser Wilhelm Land, and von Drygalski has remarked
upon the fact that the air above the inland-ice is more trans-
parent than that over the neighboring sea-ice, this arising
from its greater dryness due to its dropping down from the
heights in the interior. Frozen sleeping bags exposed in the
day-time on the slopes of the Gaussberg became soft and
dry within a surprisingly brief time, and particularly during
the storms. An ice wall built up about the tent was so
sucked up by the dry wind as to present an indented and
ragged surface. 26

The local effect of the foehn is naturally accentuated within
the steep and relatively narrow outlets from the interior
plateau. When ascending to the plateau from the Drygalski
tongue, David encountered a hot foehn which thawed the
snow, and upon the glacier tongue below the effects of earlier
foehns were found in the channelled surface and the buried
water tunnels. 27

At the winter station of the Swedish Antarctic Expedition
on Snow Hill Island, West Antarctica, even in the most
severe winter weather, sudden rises of temperature occurred
which lasted for a few minutes only, but which carried the
mercury in the thermometer up to 9j C. (49 F.), a point
higher than is reached even in the summer season. Such
remarkably abrupt changes Nordenskiold believes can only
be explained by very sudden falls of air, which in consequence
become heated adiabatically. 28

NOURISHMENT OF ANTARCTIC ICE MASSES 271

The discovery of the origin of both the Greenlandic and
Antarctic warm winds in a refrigeration of surface air layers
by contact with snow-ice masses raises the question whether
the so-called foehn winds of mountain regions have not a
similar cause in contact refrigeration. It is thus of special
interest to learn from studies of foehn winds where no such
extended snow-ice surfaces are to be found, that this expla-
nation has been offered, and that they are now believed to be
due to refrigeration by contact with elevated mountain sur-
faces.

In the Bavarian highlands the foehn winds are found to
be preceded by anticyclonic conditions and a very stable
stratification of the atmosphere. The foehn sets in earliest
at the high stations, and during its descent to lower levels
reaches stations having the same altitude simultaneously,
even though these be located in different valleys. Stagnant
air cooled by contact with the mountain sides always starts
the descending currents. The cold air drains away to lower
levels, and sometimes brings about a reversal of normal
conditions so that the higher air temperatures are at the
higher levels. When the currents have become established,
they flow down the valleys like rivers and the curve of in-
crease of temperature is found to correspond to the dry
adiabatic curve for air. 29

Recent studies of the warm outflowing currents of the
Rocky Mountain regions have shown that these flow east-
ward off the range as a broad sheet which has been followed
for many hundreds of miles in a north and south direction. 30

Wind Transportation of Snow. For the Greenland con-
tinental glacier it has been shown that the strong winds
are probably a far more potent factor in the transportation
of the snow than are all the other influences combined. The
same would appear to be no less true of the Antarctic con-
tinent. 31 During strong southerly gales the snow upon the

272 CHARACTERISTICS OF EXISTING GLACIERS

surface is picked up by the wind and the air is filled to a
height dependent upon the wind velocity. This in the case
of moderate winds may be only a foot or two, so that dogs
would be submerged in it, though ponies would still have
their heads clear. During fierce blizzards, however, the air
is loaded with snow to a much greater height. 32

The strong winds of the region, as we have seen, always
blow down off the plateau in the direction of the steepest
gradients. The southerly blizzards encountered by Shack-
leton near the end of his southern journey entirely swept
away all surface snow of recent deposit, leaving for the
return a hard and white snow surface resembling Carrara
marble. Descending the Beardmore outlet, the surface for
the first one hundred miles he likewise found swept clean,
but the lower forty miles was buried deep under drift snow. 33
Above the Backstairs Passage David found the ice surface
similarly hard and marble-like.

Over the Nordenskiold shelf-ice tongue, snow is carried
from the southern to the northern margin, where it builds a
great ice-foot which the sea-ice pulls away in sections to
form a high cliff. Within a zone surrounding the rocky
islands off the coast of Antarctica, ice-foot or fringing glaciers
are developed from the same cause. The Ross Barrier and
other bodies of shelf-ice are built higher, and, as already
explained, doubtless owe their origin to local snow deposit,
probably in large part borne from long distances by the wind.
The inland-ice of Kaiser Wilhelm Land is swept clear of snow
by the southeast storms, and the snow removed is probably
lodged farther to the west, where it forms the West-ice.

As upon the Greenland continental glacier, so here in
Antarctica, the resemblance to Sahara conditions is most
striking, the fine, hard snow grains driven by the wind be-
having as does the sand of the desert. Says Gourdon: 34

NOURISHMENT OF ANTARCTIC ICE MASSES 273

Such snow has received the name poudrin; in accumulating
upon the surface it has no consistency the grains remain with-
out cohesion. The foot has the sensation of sinking in fine sand.
Certain marches reminded me of nothing so much as of those which
at another time I had made in Southern Tunis.

Picked up by the wind this snowy powder is the chasse neige,
veritable blinding clouds which at times acquire a formidable
violence ; one sees them rise in whirls above the crests, sweep the
white plains and rob the glaciers of all the movable portions. A
great part of this snow is borne to the sea ; another part accumu-
lates in long undulations called sastrugi ; another, finally, pro-
tected behind some obstacle, ice or rock, is piled up in the form of
dunes. In short, this snow behaves like the sand of the desert,
and, further, like it, though of less intensity, it has eolian move-
ments.

FIG. 136. The lee side of a sand dune on the coast of northern California. Note
the resemblance of the curve of profile to that of continental glaciers (after a
photograph by Fairbanks).

Recognition of the importance of wind transportation in
connection with continental glaciers raises the question as

274 CHARACTERISTICS OF EXISTING GLACIERS

to how far their peculiar marginal sections are controlled
by this factor. There is evident in the sections across the
margin of the inland-ice an approximation to a regular curve
(see Fig. 132). The resemblance of this curve to the curve
of profile on the lee side of sand dunes (see Fig. 136) is most
striking. Fringing glaciers of both the Arctic and Antarctic
regions are in reality dunes of granular sand like snow, and
it seems likely that the margins of the inland-ice are broadly
moulded by this process (see Fig. 137).

" 2000 ~NIncIe - _^-- - ca ' 1900m _^~ - STntfT 2000 "

-1500-f-^ - ^^^^" - ^-^^^ - - -- j-1500 -

-1000 lls/^ - ^^-^ -- j-1000-

.500 -

FIG. 137. Section across the Vatnajokull of Iceland (after Spethmann and

Thoroddsen).

The relative parts played by wind transportation upon
the surface and by ice regelation and flow beneath it are yet
to be determined. The sharp contact of neves now with
blue glacier ice at the head of the Ferrar outlet appears to
bear upon this point.

High Level Cirrus Clouds the Source of Snow in the Interior
of Antarctica. It has heretofore been thought open to
question whether within the interior of Antarctica any snow
is precipitated in the ordinary sense. The fact that the winds
capable of transporting the snow all move outward upon the
plateau and that to the farthest point reached by Shackle-
ton what appeared to be new-fallen snow was encountered,
almost makes it necessary to assume that even there snow
reaches the surface of the plateau from the surrounding air
and is not all of it merely lifted and again deposited by the
wind. It is, however, clear that no moisture can there be
derived from surface currents, since all move outward from
the region. The only possible source of new snow is, there-

NOURISHMENT OF ANTARCTIC ICE MASSES 275

fore, the high level currents, which from cloud studies, as well
as from the observations of Mt. Erebus, are clearly shown to
supply the air of the Antarctic anticyclones.

As already stated, the dominant surface winds upon the
continental margins come from the south and southeast, with
the easterly component larger at the lower latitudes due to
deflection by earth rotation. The upper currents come, in
general, from the northwest quadrant and are more curved tow-
ard the south as they pass southward over the Ross Barrier.

Over the ice plateau the characteristic clouds which were
observed are the high broken cirrus and cirro-stratus. 35 At
times the peculiar " polar bands "or " Noah's Ark " clouds
were seen stretching across the sky and converging at oppo-
site points of the horizon, the direction of their movement
being here always southerly. 36 On the west of Ross Sea the
direction of these polar bands was from the north-northeast or
northeast curving round from the north. This is not in accord-
ance with the theory of the polar anticyclone, but conforms to that
of a continental (glacial) anticyclone, since the surface currents
on the plateau in this vicinity come from the westerly quarter.

In these high levels, clouds floating at an altitude certainly
in excess of 14,000, and probably 25,000 feet, the moisture
must be frozen, since the temperature of air ascending
through 6000 feet only is adiabatically lowered by about
35 F. There is, however, the probability that in general this
snow or ice is adiabatically melted and vaporized during its
descent to the plateau, and subsequently congealed as it
mixes with the cold air above the plateau surface. This
would explain the clear skies which are so general over both
Greenland and Antarctica during snows in the higher levels.
It is of course true that the latent heat of fusion and
vaporization of ice, abstracted as it is from the air during
its descent within the eye of the anticyclone, will counter-
act to some extent the warming adiabatic effect; and it is

276 CHARACTERISTICS OF EXISTING GLACIERS

not improbable that the long duration of Antarctic blizzards
and their somewhat sudden terminations accompanied by
snowfall are explained in part by the transformations of
latent and sensible heat.

Additional evidence for the continental and glacial rather
than the polar nature of the Antarctic anticyclone is derived
from the strong blizzards observed at the British winter
quarters on McMurdo Sound. Whereas the lighter gales came
from the southeast and indicate a control by local conditions, 37 a
blizzard of the first magnitude was not thus influenced, and al-
ways swept down from the southwest that is, from the high
plateau, and not from the pole, since otherwise the earth y s rota-
tion would have given it an easterly direction. When its
powers begin to wane, it is once more controlled by local
conditions and the wind again comes from the southeasterly
quarter. 38

An apparent confirmation of this theory of the alimenta-
tion of inland-ice masses is to be found in Nordenskjold's
narrative of the sledge journey across Northeast Land
(Spitzbergen) in 1873. While Nordenskjold did not at that
early day and on the basis of the single journey discover
the important law of atmospheric circulation above an in-
land-ice mass, which the subsequent explorations, particu-
larly of Peary, Scott, and Shackleton, have revealed, yet the
presence there of essentially the same conditions is probable
from his narrative. 39 The fine, hard snow was found to be
almost constantly in motion along the surface, which was
glazed and polished by its action. Under ordinary circum-
stances, this stream of rounded snow grains rose a few feet
only into the air, but even then it was so troublesome as to be
likened to the desert sand in the Sahara.

After the first day upon the inland-ice, during which the
weather was clear, either snow-storms or dense snow mists
were the rule, and several times a quite remarkable phenom-

NOURISHMENT OF ANTARCTIC ICE MASSES 277

enon was observed which we may best describe in Norden-
skjold's own words as rendered by Leslie : 40

During our journey over the inland ice, we several times had
a highly peculiar fall of -

1. Small, round snowflakes, sometimes resembling stars, of a
woolly appearance.

2. Grains falling simultaneously, of about the same size as
the snowflakes, but formed of a translucent, irregular ice kernel,
surrounded by a layer of water, which, however, froze in a few
moments after the fall of ice, and in a short time covered our sledge-
sail, etc., with a thin and smooth crust, or fastened itself to our hair
and clothes as small translucent ice-drops. During one such fall
on the 5th June there were seen simultaneously a faint halo and a
common rainbow, the temperature being 4 to 5 C. under the
freezing point. [See Fig. 138.1

We have thus here to do with irregular ice grains envel-
oped in water, falling in sunlight near the glacier surface in
a temperature 4 to 5 centigrade degrees below the freezing
point, and in association with freshly
precipitated felty snow-flakes. Since all
the material of the glacier surface is
snow, the source of the ice kernels must
be within the upper atmospheric regions.
We know of no source there save only
the ice grains composing the cirrus FlG . iss. Section of
clouds. The water envelope about the one of the irregular ice

A . grains enveloped in

ice kernels would be explained by the wa t er which was pre-
adiabatic rise in temperature during the cipitated together with

snow-flakes upon the

descent of these grains to the plateau iniand-ice of Northeast

within the eye of the anticvclone, and LarK | <*?*?; A - E -

Nordenskiold).

the sudden subsequent freezing would
be explained by the arrest of the downward motion and
the mixing with cold layers of air lying in contact with the
glacier. The associated snow-flakes would be derived by

278 CHARACTERISTICS OF EXISTING GLACIERS

similar changes of temperature from those smaller ice
grains which during their descent to the plateau had been
entirely melted and vaporized, as well as from the vapor
of the water envelopes about the ice kernels. It will be
interesting to learn when the central areas of the Greenland
and Antarctic glaciers have been similarly penetrated,
whether a like phenomenon is characteristic of them. 41 The
much higher altitudes and the lower temperatures make it,
however, rather unlikely. While it thus appears to be true
that the inland-ice of Northeast Land is able to induce a
local glacial anticyclone within the atmospheric envelope,
the somewhat smaller ice mass of the Vatnajokull in Iceland
produces apparently no such disturbance. 42 This is proba-
bly not alone to be explained by the somewhat smaller
dimensions of the Icelandic ice mass, but quite as much by
the fact that it is near the centre of a fixed low pressure area
of the atmosphere with disturbances of large extent and of
exceptional violence. The problem of the size of ice cara-
pace which under normal conditions is just able to induce a
local anticyclone within the atmosphere is one of great
interest, for with the initiation of this gas engine we reach
an important turning point in the processes of glacier ali-
mentation and depletion.

At the beginning of this volume it was stated that air tem-
perature has come to be recognized as the chief factor in the
initiation of glaciation. While this is true, the factor of
temperature loses its dominance and becomes secondary in
importance to wind currents so soon as the local anticyclone
has been strongly developed. As we have seen, this anti-
cyclone not only puts a stop to the nourishment of the glacier
by moisture-laden air currents, but, in addition, it constantly
transports the snow of the central portion to the margins.
This process tends to reduce the altitude of the central por-
tion of the mass as it extends the margins. With a continu-

NOURISHMENT OF ANTARCTIC ICE MASSES 279

ance of the process, the vigorous outward-blowing currents
of the anticyclone extend the margins toward the sea, where-
upon large amounts of snow are there deposited, dissipated,
and consequently lost to the inland-ice. Were it not for the
fact that the same engine pulls down the ice grains in the
cirrus cloud masses so as to in part replace its earlier nourish-
ment by the snow of low-level currents, the local anticyclone
must quickly induce a receding hemicycle of glaciation, re-
sulting eventually in its own extinction. It is, in fact, quite
possible that a limit is set upon glacier size by the refrigerated
air engine thus
brought into exist-
ence, and that the
great Antarctic
glacier, now in a
waning stage, is
an illustration of
this fact.

The warming of
the air observed
toward the close
of an Antarctic
blizzard, and the
appearance at the

rr fiTYiA nf tViP FlG ' 139 -~ Sketch map showing the glaciated and the

higher non-glaciated surfaces of the rock masses which

SOft and newly protrude through the ice in the vicinity of McMurdo

fallen snow in Sound (after Ferrar) '

place of the lifted and driven snow of the earlier stages,
both testify to the existence of the system of currents
above described. Thus an adequate explanation is found
for the disappearance of moisture congealed in the ice
grains of cirrus clouds.

Former Extent of Antarctic Glaciation. Studies of Green-
landic and Antarctic glaciation alike show that we live in a

SKETCH MAP or Ice DISTRIBUTION

O S
E

280 CHARACTERISTICS OF EXISTING GLACIERS

receding hemicycle of glaciation. According to Scott the
surface of the great inland-ice mass of Victoria Land was
once from 400 to 500 feet above its present level. 43 The
Ross Barrier has been at least 800 feet higher than now, since
Dr. Wilson discovered moraines on the slopes of Mt. Terror
at that altitude (see Fig. 139). The barrier must, therefore,
have been aground, and in the view of Scott, it once filled all
the Ross Sea as far out as Cape Adare. The Bellany Islands,
much farther out and near the Antarctic Circle, are more
heavily glaciated than is Victoria Land. Since 1840, when
Ross sailed along its edge, the Ross Barrier has receded in
places a distance of from twenty to thirty miles (see Fig. 113,
p. 217). The Nordenskjold shelf-ice tongue and the shelf-
ice of Lady Newnes Bay are remnants of the older shelf-ice
of Ross Sea and are now no longer adequately nourished.
The mountain glaciers which now lie on the east slope of the
mountain rampart of Victoria Land indicate clearly that they
were once much more important than now. Most interest-
ing of all are the ice slabs dead piedmont aprons of which
the feeders have disappeared. 44

To-day it is highly probable that far more snow is blown
from the borders of the continent out upon the sea-ice, and
hence eventually melted in the water, than falls upon the con-
tinent and its shelf-ice margin. The recession is to-day in
progress.

On 1 the summit of the Gaussberg in Kaiser Wilhelm Land,
and at a height of 350 metres (about 1140 feet) above the
surface of the surrounding inland-ice, erratic blocks of gneiss
were found, from which we conclude that this mountain was
once entirely submerged beneath the inland-ice. 45

In Belgica Strait (Gerlache Channel) of West Antarctica in
the low latitude of 64 S. are evidences that this great trench,
fully ten miles in width, was once completely filled by a great
glacier tongue which pushed westward into the Pacific. The

NOURISHMENT OF ANTARCTIC ICE MASSES 281

lateral moraines of this ice mass, fifteen to twenty feet in
height, are found to-day between sixty-five and eighty feet
above the sea level, and the depth of the channel is in the
neighborhood of 2000 feet. 46 Roches montonnees and erratic
boulders found upon the islands of the Palmer archipelago to
the west of Belgica Strait afford further confirmation of the
once much greater extension of the ice. Other data from
the same region have been furnished by J. G. Andersson, 47
Otto Nordenskjold 48 and Gourdon. 49 The last-mentioned
observer is, however, of the opinion that the present rate of
recession as estimated from the retreat of the Ross Barrier
since 1840 has been given too much weight. That the pres-
ent is, however, a hemicycle of recession, all observers are
agreed.

There is an interesting theoretical problem connected
with a possible future extinction of the Greenland and
Antarctic continental glaciers. For the latter, at least, the
dimensions of the superimposed fixed anticyclone are such
that the highest surface of the snow-ice dome may be re-
garded as in some sense an eccentric earth pole in the
wind system comparable to one of the eccentric magnetic
poles. The high-level air currents traveling as anti-trades
are at this point drawn down to the surface from all sides
and here begin their return journey equatorward as sur-
face currents. Were the glacier removed entirely, certain
changes in the wind system of this part of the globe would
certainly be brought about. The glacial studies herein set
.forth show quite conclusively that it is the dome-like sur-
face of the snow-ice mass with its universal outward grades
always increasing in value toward the periphery, quite as
much as its refrigerating property, that is responsible for
the vigor of the glacial anticyclone.

We are still without knowledge concerning the elevations
or the configuration of the underlying Antarctic basement,

282 CHARACTERISTICS OF EXISTING GLACIERS

though well convinced that it must constitute an upland
of some sort. The studies of v. Ficker 50 in the Eastern Alps
and of Bigelow and others in the American Cordillera,
indicate that on bare mountain slopes there is at times a
sliding off of refrigerated surface air under essentially
anticyclonic conditions. It seems likely, therefore, that a
continental anticyclone would persist over an outward slop-
ing Antarctic continent after the total extinction of the ice
mass, even though greatly reduced in vigor.

REFERENCES

1 Proc. Am. Phil Soc., vol. 49, pp. 96-110.

2 Racovitza, I.e., p. 416.

3 Arctowski, in "Through the First Antarctic Night," p. 431.

4 Borchgrevink, I.e., p. 305.

5 E. Gourdon, in J. Charcot, Expedition Antarctique Frangaise (1903-
1905), Geographic physique, glaciologie, petrographie, Paris, 1908, pp.
71, 74.

6 Mawson, I.e., pp. 335-336.

7 David, Narrative in Shackleton's "Heart of the Antarctic," vol. 2,
pp. 178-179.

8 Scott, I.e., vol. 2.

9 E. von Drygalski, " Zum Kontinent, etc.," I.e., p. 394.

10 David and Adams, Meteorology in Shackleton's "Heart of the Ant-
arctic," vol. 2, p. 377.

, u Arctowski, in Cook's "Through the First Antarctic Night," pp. 429-
431.

12 Bernacchi, in Borchgrevink' s "First on the Antarctic Continent,''
p. 306.

13 C. W. Royds, " On the Meteorology of the part of the Antarctic
regions where the ' Discovery ' wintered," Geogr. Jour., vol. 25, 1905, pp.
387-392.

14 David and Adams, I.e., pp. 378-379.

15 E. von Drygalski, " Zum Kontinent," I.e., p. 268.

16 Reprinted in Smithsonian Report for 1893, 1894, pp. 353-373.

17 Wm. Ferrel, "A popular treatise on the winds," New York, 1889.

18 Wm. M. Davis, "Elementary Meteorology," Boston, 1894, pp. 101,
103-104, 110-111.

19 A. Buchan, Smithsonian Report for 1897, 1898, pp. 429-432.

20 L. Bernacchi, "To the South Polar Regions," London, 1901, pp. 294-
295.

21 Shackleton, vol. 2, p. 18.

22 Proc. Am. Phil. Soc., vol. 49, pp. 99-104.

NOURISHMENT OF ANTARCTIC ICE MASSES 283

23 David, Narrative, pp. 176-184.

24 Shackleton, I.e., vol. 1, pp. 341-348.

25 David, I.e., pp. 381-383.

26 E. von Drygalski, " Zum Kontinent," pp. 418-419.

27 David, I.e., p. 164.

28 O. Nordenskiold, " Ueber die Natur des West-Antarktisclien Eisre-
gionen," Zeit. d. Gesell. f. Erdkunde z. Berlin, 1908, p. 616.

29 N. v. Ficker, " Innsbrucker Fohnstudien, IV; Weitere Beitrage zur
Dynamik der Fohns," Holder, Vienna, 1910, pp. 37-38. (Reviewed in
Nature of September 22, 1910, pp. 368-369.)

30 Science, N.S., vol. 32, Oct. 7, 1910, p. 460.

31 See among others : Royds, Geogr. Jour., vol. 25, 1905, p. 387 ; O.
Nordenskiold, Zeit. f. Gletscherk., vol. 3, 1909, p. 325.

32 "The air is entirely filled with drifting snow, which strikes you like
a sand-blast. You can not face it but have to stumble on to wherever
you may be going with your head down and arms protecting your face,
and even could you face it, you are not able to see a yard all around you.'*
(Lieut. Royds, in Geogr. Jour., vol. 25, 1905, p. 389.)

"Nothing more appalling than these frightful winds, accompanied by
tons of drift snow from the mountains above, can be imagined." (Ber-
nacchi in Borchgrevink, I.e., p. 306.)

"During snow storms it was characteristic that the snow did not drive
high. The masts of the 'Gauss' were frequently free, while the snow
below was so thick that nothing could be seen." (E. von Drygalski,
"Zum Kontinent, etc.," p. 372.)

"Nothing is more trying in the torment than this powder of murderous
crystals which whip the face and eyes and prevent one from keeping his
direction. Walking is, therefore, at times impossible and the traveller
must bury himself in a hole in the snow until the blizzard is over." (E.
Gourdon, in Charcot, Expedition Antarctique Franc,aise, 1903-1905, p. 74.)

33 Shackleton, vol. 2, p. 19.

34 Gourdon, I.e., pp. 74-75.

35 Royal Soc., National Antarctic Expedition, 1901-1904. Album of
photographs and sketches. Description of Plate 155. See also Racovitza,
I.e., p. 416; David and Adams, I.e., p. 379; David, Narrative, I.e., p. 91.

36 David, Narrative, I.e., pp. 91, 168, 171, 175.

37 The winter quarters were located in a gully or "gap" running down
from the barrier surface toward the sound in a direction from southeast
to northwest. (Royds, I.e., p. 387.) (See fig. 139.)

38 David and Adams, I.e., pp. 379-383.

39 A. E. Nordenskjold, " Die Schlittenfahrt der schwedischen Expedition
im nordostlichen Theile von Spitzbergen, 24 April-15 Juni 1873," Pet.
Mitt., vol. 19, 1873, pp. 451-452. Also A. Leslie, " The Arctic Voyages
of Adolf Erik Nordenskjold 1858-1879," with illustrations and maps,
London, 1879, p. 257. Also A. E. Nordenskjold, " Redogorelse for den
svenska polarexpeditionen ar 1872-1873," Bihang till K. Svenska Vet.
'Akad. Handlingar, vol. 2, no. 18, 1873, pp. 1-132, map and plate.

40 A. Leslie, I.e.

284 CHARACTERISTICS OF EXISTING GLACIERS

41 Shackleton's journey has made clear that the boss of the ice plateau
is far to the southwest of his route.

42 Personal communication from Dr. Th. Thoroddsen.

43 Scott, vol. 2, p. 423.

44 Scott, I.e., vol. 2, pp. 422-425.

45 E. von Drygalski, Zeit. f. Gletscherk., I.e., p. 311. Also Philippi, I.e.,
p. 7.

46 H. Arctowski, " The Antarctic voyage of the 'Belgica ' during the years
1897, 1898, 1899," Geogr. Jour., vol. 18, 1901, pp. 372-373.

47 J. G. Andersson, Bull. Geol Inst. Upsala, vol. 7, 1906, pp. 53-57.

48 O. Nordenskiold, Zeit. f. Gletscherk., vol. 3, 1909, pp. 329-331.

49 Gourdon, I.e. (1908), pp. 116-121.
^VonFicker, I.e.

AFTERWORD

The Two Contrasted Glacier Types. We have seen that
existing glaciers illustrate two widely different types in-
land-ice and mountain glaciers with the small ice-caps in
an intermediate and transitional position. As regards the
inland-ice, the Arctic and Antarctic continental glaciers dif-
fer mainly in degree ; the smaller Arctic form being entirely
restricted to the land area, whereas the Antarctic ice-mass
being more amply nourished is locally extended by a mar-
ginal terrace floating upon the sea. Similarly mountain
glaciers on the basis of their alimentation fall naturally into
a series of sub-types ranging on the one hand from those
which spread out beyond the margin of the upland the
piedmont glacier to those puny forms which in the last
stage of a progressive extinction are crowded hard against
the mountain summits.

Physiographic Form. As regards their physiographic
form the inland-ice masses adhere to a definite model --a
flat dome which in the case of the Antarctic example is ex-
tended by a lower marginal terrace of shelf ice. The moun-
tain glaciers, on the other hand, conform to no regular model,
but have a relief directly determined by the forms of the sup-
porting upland. In respect to form, the ice-cap is allied to
the inland-ice, having, in common with it, small irregularities
in the surface of its supporting base if these be but compared
to its general dimensions.

285

286 CHARACTERISTICS OF EXISTING GLACIERS

Denuding Processes. As concerns denuding processes
mountain glaciers are in common with inland-ice character-
ized by the capacity to lower the level of their beds through
the operation of the processes of abrasion and plucking ; but
they have in addition the power to denude rapidly by cirque
recession extraordinarily rapid sapping through daily
summer frost work at the base of the bergschrund. Whereas
inland-ice reduces the irregularities and softens the outlines
of its rock bed, mountain glaciers by contrast increase the
accent of the relief, and, in fact, develop a more sharply
rugged topography than does any other known geological
process. In respect to denudation, ice-caps are intermediate
between the two main glacier types. While their main deg-
radational process is apparently abrasion (plate 34 A) they
may, when aided by uplift of the land under specially
favorable conditions, develop the sharply peaked mountains
known as tinds. Unlike the peaks carved by mountain
glaciers, these sharp peaks are developed not above the
higher but near the lower ice levels (plate 34 B).

Alimentation. In respect to nourishing processes the two
main glacier types are no less sharply differentiated. The
ice-caps, which in their physiographic form are allied to the
inland-ice, are here no less clearly to be classed with moun-
tain glaciers.

Mountain glaciers are nourished by moisture-laden surface
currents of air, which encountering an upland area are forced
to rise, and are thereby cooled adiabatically and by contact
with the upland, so that their burden is deposited in the form
of snow. Inland-ice, on the contrary, if we neglect for the
moment the marginal terrace sometimes present, appears to
be regularly fed from high-level currents through the opera-
tion of a refrigerating air engine of which the ice mass and its
atmospheric cover are the essential parts. :

Through the rhythmic action of this engine the congealed

PLATE 34.

A. View of the high surfaces of the Jotunheim from the Galdho, showing effect
of abrasion beneath ice-cap glaciers (after Fritz Machacek).

B. The Maelkevoldsbrae of the Jostefjelcl, showing the development of tinds about
the borders of a Norwegian ice-cap through the erosional work of outlet glaciers
(after Fritz Machacek).

AFTERWORD 287

moisture derived from the ocean surface within moderate or
low latitudes and carried to the polar region in the high level
cirrus clouds, is pulled down to the surface of the glacier in the
eye of a great glacial anticyclone which is centred above it.
During their descent from high levels the ice grains of the
clouds are melted and vaporized by adiabatic warming, and
on reaching the cold surface layer of air next the ice, are
quickly congealed to form flakes of fresh snow. The progres-
sive warming of the air adiabatically both during its descent
to the central area of the ice mass and on the further slide out-
ward to the peripheral portions, gradually damps and eventu-
ally stops the sliding centrifugal motion of the surface air-
layer. Thus the engine comes to rest or, as we may say,
has reached the end of its stroke. The great calm which en-
sues allows heat to be again slowly abstracted from the sur-
face layer of air, thereby lowering its temperature and raising
its density until gravity again starts the engine, which now
acquires the steadily accelerating velocity characteristic of
bodies sliding on inclined planes. The tempest which is
eventually engendered is succeeded by a rapid rise of air
temperature, a fall of fresh snow, and another stopping of the
engine.

The fierce violence of the surface air currents when at their
maximum, and the fall of the snow for the most part as the
engine is slowing down, together make of this glacial anti-
cyclone a gigantic snow broom. The snow deposited as it
were between strokes of the engine is by the next sweep of
the broom brushed largely clear from all central portions of
the glacier, and the sweepings are deposited near and about
the margins of the mass (see Figure 140). Since the con-
tinental glacier must grow during the advancing hemicycle
in a vertical as well as in a horizontal direction, there must
be accretions of snow upon the central areas, which layers
adhere to the surface so as not to be removed by the

288 CHARACTERISTICS OF EXISTING GLACIERS

rhythmic engine strokes. Thus are produced the alternat-
ing layers of incoherent and marble-like snow which the
crude sections of the surface material have revealed. We

A N T I c C L O.N

FIG. 140. Diagram to illustrate the growth of an inland-ice mass through the
rhythmic action of the anticyclonic air engine.

are still without sufficient knowledge of the conditions
which give rise to the coherent layers. In this manner, then,
the great glacier is enlarged and shaped through periodic
deposit of snow in successive layers upon its surface but
especially by more frequent deposit of sweepings about its
margin. This process plays no part in the shaping of
mountain glaciers.

The marginal shelf ice appears to be in some cases at least
partially nourished by the overflow of glacier ice from the
neighboring continental area, but more largely it is fed
through the continued thickening of field ice the frozen
sea surface by deposits of precipitated snow upon its sur-
face. To this precipitated snow there is added a portion of
the sweepings carried outward from the surface of the inland-
ice. That portion of the nourishment which is derived from
glacier overflow at its inner margin, becomes covered by the
surface deposits, and so is segregated toward its lower and
inner margins.

The formation of shelf ice is greatly favored in more or less
sheltered bights from which the sea ice is less easily dislodged
and hence does not take part in the annual drift of the pack
toward lower latitudes. Further it is locally favored by the
stranding or freezing into the sea ice of a fleet of icebergs

AFTERWORD 289

from the areas of inland-ice; since these furnish lanes be-
tween them within which snow sweepings are more easily
caught and retained.

Marginal Contours. As concerns the moulding of surface
contours at the glacier margin, the determining factors in the
case of mountain glaciers seem to be an upturning of the ice
layers in this region and surface ablation or melting. In the
case of inland-ice the important factor is the deposition upon
the surface of snow borne by the wind from the interior of the
mass.

INDEX

Aarschlucht near Meiringen, 63.

Ablation, on Greenland glacier, 162.

Accordance, of side valleys, cause of, 67 ;
of summit levels, 31.

Adams, Lieutenant, cited, 282.

Adelie Land, 193.

Adiabatic refrigeration of air, 36, 150.

Adiabatic warming of air, during Ant-
arctic blizzard, 269.

Admiralty Range, 193, 253.

Advancing hemicycle of glaciation, 6.

Agassiz, Louis, cited, 3, 10, 167.

Aiguille type of mountain ridge, 32.

Air circulation over Greenland glacier,
146.

Air, humidity of, in Antarctica, 262 ; in
Greenland, 145-146.

Air temperatures, relation to glaciation,
4, 278 ; over inland-ice of Greenland,
145.

Alaska, icebergs of, 178.

Albs, 53, 64.

Alexander I Land, 211.

Alimentation, of glaciers, 286 ; of Green-
land glacier, 131, 143; of mountain
glaciers in polar regions, 260.

Alpine glaciers, 52 ; of Antarctica, 259.

Amundsen, Roald, cited, 206, 211, 213.

Ancestry of glacial theories, 1.

Andersson, J. Gunnar, cited, 96, 141,
210, 211, 243, 252, 281, 284.

"Antarctica," crushing of, in pack ice,
205 ; sinking in pack ice, view of, 205.

Antarctica, exploration of, 190 ; maps
of, 194, 195 ; rock basement of, 281.

Antarctic glacier, larger than land base,
187 ; where unconfined, 245 ; where
unconfined, view from sea, 246.

Antarctic region, characterized, 98 ;
climatic symmetry of, 99.

Anticyclone, glacial, essentials of, 148.

Apron, outwash, 87.

Aprons, ice, below outlets, 257 ; of shelf-
ice tongues, 234.

Arctic glacier type, 97.

Arctic region, characterized, 98 ; climatic
asymmetry of, 99 ; contrasted with
Antarctic, 186.

Arctowski, Henryk, cited, 189, 197, 198,

202, 210, 211, 212, 213, 238, 243, 264,

282, 284.
Arete, 32.
Asakak glacier outlet (Greenland), 125,

134.

Asulkan glacier, 54, 55.
Asymmetry of Greenland glacier, 131,

146.
Atmospheric depression, fixed areas of,

99.

Atwood, W. W., cited, 26, 39, 60.
Austmann valley, Greenland, moraines

of, 138.

Backstairs passage, Victoria Land, 254-
272.

Baffin Land, map of, 117.

Baffin's Bay, currents in, 162.

Bagnoires, 166.

Baird glacier, Alaska, 46.

Balch, E. S., cited, 210.

Baltoro glacier, 47, 50.

Barchans, of snow, 156.

Barrier-ice. See Shelf-ice.

Basin, tongue-like, before mountain
front, 83.

Basins of exudation above outlets, Green-
land glacier, 132.

Beardmore outlet, 223, 233, 253, 254,
256, 257, 272 ; map of, 258.

"Belgica" expedition, 189, 191, 196, 197,
200, 203, 206, 262, 265.

Belgica Strait, former glaciation of,
280.

Belleny Islands, 192, 209, 280, pi. 27.

Benard, Charles, cited, 118.

Benedict glacier, 169, pi. 25.

Bergen railway, Norway, melting of
snow, on, 166, 176.

Bergschrund, 15, 61 ; explored by John-
son, 16 ; in relation to cirque, 14 ;
time of its appearance, 22.

Bering glacier, Alaska, 43.

Bernacchi, L., cited, 189, 210, 212, 242,
264, 265, 282, 283.

Biafo glacier, 47.

Bigelow, F. H., cited, 282.

291

292

INDEX

Bighorn Mountains, cirque cutting in, 19.

Birth of tabular bergs, 235.

Biscoe, John, cited, 192, 196, 215.

Bishops glacier, Frontispiece.

Blackwelder, E., cited, 56.

Blizzards, Antarctic, 264, 268, 272;

duration of, in relation to latent heat

transformation, 276 ; sequence of

events during, 269 ; termination of,

276.

Blue glacier, 259.

Bonney, T. G., cited, 12, 15, 22, 23.
Borchgrevink, C. E., cited, 191, 204, 211,

216, 230, 231, 242, 282, 283.
Border lakes, 83.
Bouvet Island, 209, pi. 28.
Braided streams, flowing from glacier

front, 85-86.

"Break up" of sea-ice, 206.
Brooks, Alfred H., cited, 56, 57.
Brown, et al. (of "Scotia" expedition),

cited, 211, 242, 252.

Bruce, W. S., cited, 189, 193, 212, 216.
Bruckner, E., cited, 10, 11, 40, 56, 60,

62, 69, 84, 85, 96, 212.
Bryant glacier, pi. 22.
Buchan, A., cited, 100, 265, 282.
Butler, B. F., cited, 58, 185.

Calhoun, F. H. H., cited, 56, 88, 96.
"Canals," on inland-ice of North East

Land, 115; view of, 114; hypotheti-
cal section of, 115.
Cape Adair, 189 f 191, 199, 223, 240, 249,

262, 264.

Cape Armitage, 189, 190.
Cape Carr, 192.
Cape Royds, 220, 223, 264.
Capps, Stephen R., Jr., cited, 96.
Carbon dioxide, content of, in air over

Greenland glacier, 145.
Cascade stairway, 59.
Case, E. C., cited, 137.
Cauldron glaciers, 51.
"Challenger" expedition, 190, 197, 198,

211, 212, 237, 241.
Chamberlin, T. C., cited, 12, 22, 56, 137,

138, 139, 140, 141, 142, 153, 160, 170,

176, 209.
Characteristic profiles, from high latitude

glaciation, 74.
Characteristic section, from successive

landslides in canyon, 92.
Charcot, J., 191,211, 212, 252, 282.
Charpentier, Jean de, cited, 3.
Chasse neige, 273.
Childs glacier, Alaska, 45.
"Chimneys," 30.
"Chinese Wall," on Grinnell Land, 116,

128; view of, 116.

Chistochina glacier, 48.

Chun, cited, 213.

Cirque, definition, 10 ; figure after
Richter, 14; form of, in different
stages, 31 ; its initiation, 18 ; its re-
cession, 12 ; relation to Bergschrund,
14.

Cirques, in Victoria Land, 259 ; loca-
tion of, in early stages, 32 ; maturity
of, 29 ; multiple, 30 ; nourishment of,
by snow, 31 ; on lee side only of moun-
tain range in Colorado, 28 ; rock flows
from abandoned, 94.

Cirrus clouds, nature of moisture in,
275 ; snow of, 158 ; source of Green-
landic snowfall, 158 ; source of snow
in interior of Antarctica, 274.

Cliff glaciers, 54.

Climatic conditions, affecting nourish-
ment of Antarctic ice masses, 216.

Clouds, over Greenland ice, type of, 159 ;
"Polar bands" of Antarctica, 275.

Coats Land, 193, 195, 197, 216, 245, 270 ;
view of, 216.

Col, formed by intersection of cirques,
34 ; typical example from Selkirks,
pi. 9.

Cols, characteristic high levels of, 35.

Comb-ridges, 32.

Conditions which bring on glaciation, 5.

Continental (glacial) anticyclone of An-
tarctica, 265, 266.

Continental glacier, of Greenland, cross
section of , 122; physiography of , 119;
of Victoria Land, 253.

Continental platform, Antarctica, 196.

Contrast of northern and southern polar
areas, 98.

Convict Lake, view of, 82.

Conway, Sir Martin. See W. M. Con-
way.

Conway, W. M., cited, 57, 98, 111, 112,
117, 118.

Cook, Captain James, cited, 190, 192,
196, 215.

Cornell glacier, 172, pi. 24, pi. 26.

Cornish, Vaughan, cited, 155, 157, 161.

Corries, of Scottish highlands, 34.

Cracks, tide, 208.

Crevasses, in Greenland glacier, 129 ; in
inland-ice of North East Land, 115;
rectangular, in Antarctic glacier, 247.

"Cryaconite" wells, on Antarctic glacier,
248.

Cryohydrates, from sea-ice formation,
199.

Cutting effect, of drift snow, 154.

Cycle of glaciation, 6.

Cyclonic air circulation, over south
margin of Greenland glacier, 149.

INDEX

293

Dalagers nunataks, Greenland, scape
colks at, 135 ; map of, 136.

Daly, R. A., cited, 31, 40.

Danco Land, 211.

Daubree, A., cited, 201, 213, 236.

David outlet, 233, 254.

David, T. W. E., cited, 213, 242, 243,
255, 260, 266, 267, 269, 270, 272, 282,
283.

Davidson glacier, Alaska, 45.

Davis, W. M., cited, 6, 9, 11, 25, 39, 66,
67, 265, 282.

Dawson glacier, pi. 6.

Delta, in ice-dammed lake, pi. 26.

Dendritic glaciers, 47.

Denuding processes, of glaciers, 286.

Depletion, of glaciers, special causes of,
36 ; of Greenland glacier, from surface
melting, 162.

Depot "A," Victoria Land, 222, 223.

Deserts and inland-ice compared, 150,
272, 273-276.

Devil's Thumb, Greenland, glacier mar-
gin at, 126.

Diagram, of shore line of marginal lakes,
172 ; to illustrate air circulation over
Greenland glacier, 146 ; to illustrate
birth of icebergs (Reid), 181 ; to illus-
trate birth of icebergs (Russell), 180;
to illustrate differential melting about
rock fragments, 166, 167 ; to illustrate
formation of col, 34 ; to illustrate
formation of horns, 33 ; to illustrate
formation of lakes in drift ice, 202 ;
to illustrate formation of zigzag leads,
202 ; to illustrate growth of inland-ice
mass, 288 ; to illustrate regular cracks
in drift ice, 201 ; showing longitudinal
section of glaciated valley, 60 ; show-
ing manner of formation of "West
ice," 230 ; showing serial subsurface
temperatures in Greenland glacier,
164 ; to^how "biscuit cutting" effects,
26.

Differential surface melting of ice, 165.

Di Filippi, Filippo, cited, 40.

Dimples, above outlets of Greenland
glacier, 132 ; on Antarctic glacier, 257 ;
on Greenland ice near Disco Bay, 134.

Direction of nearest land determined by
winds over ice, 147.

Dirk Gerritz Archipelago, 211.

Dissection of upland, by mountain
glaciers, stages of, pi. 4.

"Docks," in North East Land, 116.

Drainage, on Greenland glacier, 170.

Drift ice, pressures in, 200.

Drift site, in Lapland, figure after pho-
tograph by Von Zahn, 21.

Drift sites, in Bighorn Mountains, 19.

Drift snow, over Greenland glacier, 151.

Drumlins, position of, in site of ice apron,
83.

Drygalski, E. v., cited, 11, 125, 131, 134,
138, 141, 142, 146, 153, 160, 161, 162,
163, 169, 174, 176, 177, 179, 183, 184,
189, 191, 192, 193, 203, 210, 211, 227,
228, 230, 239, 241, 242, 243, 244, 246,
249, 250, 252, 270, 282, 283, 284.

Drygalski shelf-ice tongue, 223, 231, 233,
254, 270.

Dugdale glacier, 230, 231.

Duke of the Abruzzi, cited, 106, 108,
118.

Dumond d'Urville, J. S. C., cited, 190,
193, 210.

Dunes, snow, 274.

Dusen, P., cited, 124.

Dust wells, 166, 167.

Ebeling, Max, cited, 118.

Effects of wind drift on snow density,
157.

Ellesmere Land, inland-ice of, 116; map
of, 115.

Enderby Land, 192, 195, 196, 215.

Engell, M. C., cited, 179, 184, 185.

Englacial drainage, on Greenland glacier,
170.

Englacial streams, Antarctica, 234.

Environment, importance of in evolu-
tion of science, 2.

Erebus, Mount, behavior during bliz-
zard, 269, 275.

Ericksen, Mylius, expedition of, in
Greenland, 121.

Ericksen 's route across inland-ice of
northeast Greenland (map), 127.

Erosion by drift snow, 154, 155.

Eskers, 87 ; manner of formation of, 88.

Evaporation, over inland-ice of Green-
land, 145.

Exfoliation, its part in formation of
tinds, 79.

Expanded-foot glaciers, 45, pi. 10.

Experiments in glacier motion, 137.

Explorations, Antarctic, 3.

Fairbanks, H. W., cited, 82, 273.

Features within marginal zone of Green-
land glacier, 129.

Feeder basins (dimples) on Greenland
ice, 134.

Feilden, H. W., cited, 24, 71, 80.

Ferrar glacier, 253.

Ferrar, H. T., cited, 212, 213, 242, 259,
260, 279.

Ferrel, William, cited, 265, 282.

Ficker, Heinrich v., cited, 271, 282, 283,
284.

294

INDEX

Field-ice, defined, 198 ; manner of forma-
tion of, 199.

Filchner, Wilhelm, cited, 212.

Fixed areas of atmospheric depression,
99.

Fjords of western Norway, 73.

Flatly grooved glacier valleys, 72.

Flimser Bergstiirz, 93.

Fluvio-glacial deposits, 88 ; of Green-
land, 142.

Foehn blizzards, of Antarctica, 268.

Foehn winds, 234, 268, 270; drying
effect of, 270 ; local intensification of,
in Antarctica, 270 ; of Antarctica,
271 ; of Bavarian Highlands, 271 ; on
borders of Greenland, 149.

Foetal Glacier outlets, view of, 226.

Former extent of Antarctic glacier,
279.

Form of tongue-like basin, diagram,
88.

Foster glacier, Alaska, 45, pi. 10.

Franz Josef Land, 106 ; map of, 107.

Fretted upland, 29, pis. 4, 6, 7 ; com-
pared with etched faces on crystals,
40 ; in part submerged, pi. 17 ; East
Greenland, pi. 22.

Fricker, Karl, cited, 210, 212.

Friederickshaab glacier, Greenland, 44 ;
map of, 171.

Friedrichsen, Max, cited, 57.

Fringing glaciers, 274, pi. 27 ; Green-
land, 151.

Fringing ice-foot, 209.

Front of Greenland glacier, 127-128.

"Frost snow," Greenland, 144, 145, 153;
Antarctica, 263.

Future condition of Antarctica, possible
effect upon wind system, 281.

Gains and losses of glaciers, how con-
trolled, 36.

Gannett, Henry, cited, 15, 23, 242.

Garde, J. V., cited, 120, 121, 129, 130,
140, 141, 255.

Garde's route, map of, in southern Green-
land, 121.

Garwood, E. J., cited, 23, 48, 57, 96.

Gastaldi, B., cited, 13, 22.

"Gauss" expedition, 189, 198, 203, 228.

Gaussberg, 247, 248, 270, 280, pis. 32,
33.

Gehangegletscher, 209.

Geikie, A., cited, 13, 15, 23.

"Gendarmes, 1 ' 30.

Gerlache, Adrian de, cited, 191, 211,
213.

Gilbert, G. K., cited, 11, 18, 23, 26, 28,
39, 52, 58.

Glacial abrasion, 9.

Glacial amphitheatre. See Cirque.

Glacial anticyclone, essentials of, 148 ;
Antarctic, 265, 266.

Glacial cycle, of Davis, 6.

Glacial features due to deposition, 81.

Glacial sculpture, pi. 15 ; by Norwegian
glaciers, pi. 34; high latitude, 70;
high level, 25 ; in moderate latitudes,
low level, 59.

Glacial theories, ancestry of, 1.

Glacial trough, overdeepened by over-
flow glacier, view, 77.

Glaciated surface, pi. 16 ; furrowed by
shallow channels, map showing, 73.

Glaciated valleys, too large for present
streams, 67.

Glaciation, cycle of, 6.

Glacier channels, directed by rock
structures, 75.

"Glacier docks," 116.

Glacierets, hanging, 50, 53.

"Glacier run," 105.

Glaciers, Arctic type, 97 ; classification
of, 7, 41, 97, 285; dependent upon
alimentation, 41 ; first appear on lee
side of ranges due to wind distribu-
tion, 28; ice-foot, 209; in Caucasus
Mountains, 38 ; initial forms of, 37 ;
life history of, 36 ; mountain, nourish-
ment of in polar regions, 260 ; moun-
tain, types of, pi. 11; "new-born,"
37 ; nivation, 37 ; nourishing processes
8 ; outlet, 104 ; rock, of Alaska, 96 ;
size of, in relation to land masses, 101 ;
slope, 209 ; Spitzbergen type, 210.

Glacier stars, 166, 168.

Glacier Tongue, 231.

Gletscher sterne, 166, 168.

Glint lakes, 136.

Gorge of the Albula River, view of,
90.

Gorges in glaciated valleys, how formed,
66.

Corner glacier, 52.

Gourdon, E., cited, 5, 158, 161, 200, 206,
209, 212, 213, 237, 243, 244, 251, 252,
272, 281, 282, 283, 284.

Gradation, from nivation to glaciation,
20.

Graham Land, 211.

Grat, 32.

Great Aletsch glacier, 50, 52, pis. 5, 12 ;
size compared to Beardmore outlet,
258.

Great Karajak glacier, 178.

Great Ross Barrier. See Ross Barrier.

Greely, A. W., cited, 116, 118, 210.

Greenhalgh Mountain, rock streams on,
pi. 19.

Greenland, map of, 120.

INDEX

295

Greenland glacier, air circulation over,
146 ; asymmetry of, 162 ; east and
west shores compared, 162 ; nourish-
ment of, 143; outlines of, 119.

Gregory, John W., cited, 111.

Grinnell Land, map of, 115.

Grooved upland, 29, pis. 4, 6.

Grossmann, Karl, cited, 7, 11.

Hanging glacier, defined, 57.

Hanging glacierets, 50, 54, pis. 12, 14.

Hanging valley, 48, 61, 66, 67, pi. 13.

Hardangerjokull, 102, pi. 17.

Harker, Alfred, cited, 40.

Hauthal, R., cited, 177.

Hayes, C. W., cited, 57.

Hayes, I. I., cited, 121.

Head-wall erosion, 10.

Heat transfer, between poles and equa-
tor, 99.

Holland, Amund, cited, 13, 14, 22. 141.

Hemicycle, advancing, 35 ; receding, 89.

Hemicycles of glaciation, 6.

Hess, H., cited, 7, 40, 56, 118, 122, 137,
141, 238, 240, 244.

High latitude glacial sculpture, 70.

High level clouds, bring snow to interior
of Antarctica, 274.

High level glacial sculpture, 8.

Hispar glacier, 47, 50.

Hobbs, W. H., cited, 24, 96, 117, 210,
260.

Hofer, Hanns, cited, 118.

Hofs Jokull, 102, 103.

Hollander, L. M., cited, 78.

Holmes, W. H., cited, 81.

Horn, defined, 33.

Horns, in relation to neves, 33.

Horseshoe Glacier, the, in the Canadian
Rockies, 54.

Horseshoe glaciers, 53 ; concavity of
frontal margin, 54 ; of Antarctica,
259.

Howe, Ernest, cited, 95, 96.

Hugi, F. G., cited, 3.

Humidity of air, absolute, in Greenland,
145-146; in Antarctica, 262; rela-
tive, in Greenland, 145, 146.

Hummocks, in sea-ice, 204.

Huntington, Ellsworth, cited, 10.

Hydraulic ram effect, at termination of
Antarctic blizzard, 270.

Hyperbolic form, of col, 34.

Ice aprons, below outlets, 257.

Ice barrier, surface of, pi. 30.

Iceberg, in Melville Bay, view of, 182.

Icebergs, Antarctic, beauty of, 240, 241 ;
Antarctic, debris in, 241 ; Antarctic,
drift of, 240; Antarctic, melting of,

241 ; Arctic, manner of birth of, 178 ;
blue, of Antarctica, 239, 248; of
Antarctica, in parallel ranges, 203 ;
of Greenland, 182, 183 ; of ice-dammed
lakes, 179; the anchorage of "West
ice," 250; rock debris in, 249; von
Drygalski's classification of, 249 ;
tabular, deformation of, 239 ; tabular,
embryonic forms of, 251 ; tabular,
Antarctica, 234 ; tabular, forming
from Ross Barrier, view of, 235 ;
tabular, rectangular plan of, 237 ;
tabular, stratification of, 239 ; tabu-
lar, views of, 236, 237.

Ice blink, defined, 141.

Ice-cap, of Eyriksjokull, 7 ; suddenly
melted by lava, 105 ; transitional
position of, 7.

Ice-cap glaciers, 42 ; of East Greenland,
124.

Ice-caps, of Iceland, 8, 102 ; of Norway,
8, 104 ; on volcanic peaks, 8.

Ice-caves, 208.

Ice-cliff, at fjord heads, Greenland, 178.

Ice-cones, debris covered, 168.

Ice crystals, in glacial anticyclone, 270.

Ice-dammed lakes, pis. 25, 26 ; in Green
Mountains, 172.

Ice dams, in extraglacial drainage, 174.

Ice face, of Greenland glacier, 127, pi. 23.

Ice flowers (rosette-like ice crystals), 199.

Ice-foot, 208.

Ice-foot glaciers, 208, 209, pi. 27.

Ice front, of Greenland glacier, pi. 24.

Ice grains in water, precipitated in
sunlight, 277.

Ice grottoes, about nunataks, 169.

Ice island, 208, pi. 28 ; views of, 209, pi.
28.

Ice plateau, of Antarctica, monotony of,
256.

Ice slabs, 259, 280.

Ice-tongues (outlet glaciers), 125-126.

"Icy barrier," how used by Wilkes, 215.

Ideal cross-section, of U-valley and Albs,
64. ,

Illecillewaet glacier, 50.

Inherited basin glacier, 50.

Initiation of glaciation, 5.

Inland Forts, Victoria Land, 259.

Inland-ice, contrasted with mountain
glaciers, 285 ; ideal section across, 7 ;
in relation to basement, 7 ; physio-
graphic form of, 7 ; of Kaiser Wilhelm
Land, pi. 32 ; of Spitz bergen, views of,
111-112.

Insolation, over Antarctic glacier, 262,
263.

International cooperative expeditions to
Greenland, desirable, 143.

296

INDEX

Intersecting crevasses in Antarctic

glaciers, map of, 247.
Isblink, defined, 141.
Ivory Gate, the, on Spitz bergen, 111.

Jackson, F. G., cited, 118.

Jamieson, T. F., cited, 172, 177.

Jensen, J. A. D., cited, 120, 140, 170,
171, 176.

Johnson, Willard D., cited, 15, 17, 18, 23,
25, 39, 68; his exploration of a
Bergschrund, 16.

Joint planes, in connection with land-
slides, 92.

Jokulhlaup, 105.

Jostedalsbraen, pi. 18 ; map of, pi. 20.

Jotenheim, pi. 34.

Kaiser Wilhelm Land, 193, 195, 203, 216,
239, 244, 246, 250, 253, 263, 265, 272,
280, pi. 32.

Kames, in Greenland, 140.

Karajak glaciers, sections of, 179.

Karling, defined, 32; in North Wales,
pi. 8.

Karso trough valley, in Northern Lap-
land, view of, 71.

Kemp, cited, 192, 197.

Kemp Land, 192, 195, 215.

Kennicott glacier, 48.

Kilimandjaro, ice-cap of, 43.

King Edward Land, 193, 195, 209, 216.

King Oscar Land, 211, 216; ice terraces
of, 224 ; map of ice terraces, 225.

Klutlan glacier, Alaska, 46.

Knobs rising from dome as result of
ice-cap glaciation, view of, 76.

Kornerup, cited, 171.

Krech, cited, 209.

Lake Argentine, 174, pi. 25.

Lake Constance, origin of, 83, 84.

Lake Garda, formation of, 84.

Lake ice, manner of formation of, 226.

Lake Mono glaciers, former, pi. 15.

Lake Tyndall, 174.

Lakes, border, 83 ; in drift ice, lozenge-
shaped, 203; marginal, to Greenland
glacier, 171-173 ; morainal, 82.

Landslide, of Elm in Canton Glarus, 93 ;
of Frank, Alberta, 92.

Landslides, in Colorado, 92 ; in valley
vacated by glaciers, 91.

Lang Jokull, 102, 103.

Lapland, former glaciers of, 39.

Lapp's Gate, Lapland, 73.

Larsen Bay, sea-ice of, 225.

Latent heat transformations, during
Antarctic blizzard, 269.

Lateral moraines, Victoria Land, 259.

Lateral streams, of outlet glaciers, 169.
Laurentian district of North America,

temperature necessary for glaciation,

Law of adjusted cross-sections, 60, 137.
Lawson, A. C., cited. 25, 26, 30, 39, 57.
Leads, 200, 206.
Lefroy glacier, Selkirks, 50.
Lendenfeld, R. v., cited, 57.
Leslie, A., cited (translator), 277, 283.
Leverett, Frank, cited, 10.
Life history of a glacier (Russell), 6.
Little Cottonwood canyon, U-valley,

pi. 16.

Lockwood, Lieutenant, cited, 128.
Lofoten Islands. Norway, 31, pi. 7.
Low level glacial sculpture, 8, 59.
Lozenge-shaped lakes, in drift ice, 202.

Machacek, Fritz, cited, 80, pi. 34.

McMurdo Sound, 218, 231, 253, 259, 264,
267, 276, 279.

Maine, gulf of, probable former shelf-
ice in, 214.

Malaspina glacier, 43 ; evolution of, 37.

Map, Antarctica, 194, 195 ; Asulkan
glacier, 54 ; Baird glacier, 46 ; Beard-
more outlet, 258 ; David's route to
south magnetic pole, 267 ; Greenland,
showing inland-ice, 120; Hispar gla-
cier, 47 ; Hofsjokull and Langjokull,
102; Illecillewaet glacier, 49; King
Oscars and Kaiser Franz Josef fjords,
124 ; Lake Garda, 84 ; North Green-
land, 133 ; of area near Tornetrask,
Swedish Lapland, 72 ; areas of heavy
glaciation, northern hemisphere, 100 ;
braided stream, from Iceland, 86 ;
fixed "lows" in northern hemisphere,
100 ; glaciated and unglaciated rock
near McMurdo Sound, 279 ; Great
Ross Barrier, 217 ; morainic ridges in
front of Wasatch Range, 82 ; Ross
Barrier (outline), 220; margin of
Ross Barrier, 2 17-; Northeast Fore-
land, Greenland, 127 ; shelf-ice tongues
of Ross Sea, 232; soundings, "Bel-
gica," 196 ; radiating glacier of the
Nicolai valley, 53 ; shelf-ice tongues,
Robertson Bay, 230; Sheridan gla-
cier, 45 ; Storfjord with joint directed
valleys, 75 ; terraces of King Oscar
Land, 225 ; showing dimple above
Ferrar glacier, Antarctica, 256 ; show-
ing superglacial streams on Greenland
glacier, 165 ; Tasman glacier, 48 ;
Victoria and Lefroy glaciers, 50;
Wenkchemna glacier, 55.

Marginal contours, of contrasted glacier
types, 289.

INDEX

297

Marginal cross-sections, Antarctic gla-
cier, 253.

Marginal lakes, Greenland glacier, 171-
173.

Marginal moraines, Greenland, 138, 139.

Marginal physiography, of Greenland
glacier, 123.

Marguerite Bay, field ice of, 251.

Miirjelensee, 175.

Markham, Sir Clements, cited, 210.

Martin, G. C., cited, 45, 46, 56, 57.

Martin, Lawrence, cited, 46, 57.

Martonne, E. de, cited, 13, 23, 61, 62, 63,
64, 65, 66, 67, 68, 69.

Matterhorn, 33, pi. 9.

Matthes, Francois E., cited, 15, 19, 20,
22, 23, 26, 39, 40, 57.

Mawson, Douglas, cited, 207, 212, 213,
282.

Mawson outlet, 234.

Medial moraines, Greenland, 138 ; Vic-
toria Land, 259.

Melting of Antarctic ice, 234.

Melting of Greenland glacier toward
interior, 146.

Melville Bay, 178 ; ice margin at, 132.

Mendenhall glacier, Alaska, 45.

Mendenhall, W. C., cited, 57.

Mer de glace, 52, 53.

Merwin, H. E., cited, 177.

Meyer, Hans, cited, 56, 57.

Miles glacier, Alaska, 45.

Mill, H. R., cited, 210, 212.

Mills (Moulins) on Greenland glacier,
166.

Minna Bluff, 222, pi. 30.

Moat, Antarctic, view of, pi. 33.

Moats, Antarctic, 257 ; Greenlandic,
169.

Mohri, H., cited, 140, 142, 160, 176.

Mont Blanc, significance of its dome
form, 32.

Moraines, ground, 81 ; lateral, Belgica
Strait, 281 ; lateral, Victoria Land,
259; marginal, 138, 139; medial, 81,
138; medial, Victoria Land, 259;
of mountain glaciers, 81 ; on flanks
of Sawatch Range, view of, 81 ; re-
cessional, 82, 87; terminal, about
apron sites, 83, 84 ; terminal, of
Greenland glacier, pi. 24.

Moreno, Francisco P., cited, 177.

Mossman, R. C., cited, 210.

Mt. Assiniboine, 33.

Mt. Lyell glacier, 55.

Mt. Ranier, pi. 13.

Mt. Sir Donald, pi. 9.

Mountain foreland, apron sites on, 83,
84 ; stream action of, 85.

Mountain glacier, ideal section across, 7.

Mountain glaciers, alimentation of, 36;
cauldron type, 51 ; contrasted with
inland ice, 285; "dead," 51 ; defined,
6 ; dendritic type, 47 ; evolution of r
37 ; form sensitive to temperature
changes, 41 ; horseshoe type, 53 ;
inherited-basin type, 50 ; in relation,
to basement, 7; "living," 51; mar-
ginal to Greenland, view of, 122 ;
nivation type, 42 ; nourishment of, in-
polar regions, 260 ; of Antarctica, 259 ;
on volcanic cone, pi. 13 ; on volcanie
peaks in low latitudes, 51 ; piedmont
type, 43 ; radiating (Alpine) type, 52 ;
relation to bed, 41 ; tide-water sub-
type, 51 ; transection type, 44 ; types
of, 42, pi. 11.

Mountain rampart, of Victoria Land,
253 ; of Antarctica, mountain glaciers
on, 259.

Movement, of Antarctic glacier, 248 ;
of Antarctic glacier, in Kaiser Wil-
helm Land, 222 ; of Pleistocene gla-
cier in Scandinavia, 136 ; rate of, in
glacier outlets, 134.

Multiple cirques, pi. 5.

Murray, George, cited, 210.

Murray, Sir John, cited, 212, 241, 242,
244, 265.

Nansen, Fritjof, cited, 4, 120, 122, 123,

124, 129, 131, 132, 138, 140, 141, 142,

144, 145, 148, 150, 160, 161, 162, 176,

207, 255, 263.

Nathorst, A. G., cited, 141.
Neu-Haufen dyke, Danube, map of, 136.
Neumayr, Georg v., cited, 210.
Neve snow of Greenland, 153.
Nivation, 18 ; in Yellowstone National

Park, view of, pi. 2 ; on Quadrant

Mountain, Yellowstone National Park,

20 ; in Swedish Lapland, 20.
Nivation glaciers, 37, 42.
"Noah's Ark" clouds, 275.
Nordenskiold, A. E., cited, 4, 113-118,

122, 123, 144, 150, 160, 161, 166, 169,

170, 176, 255, 276, 277, 283.
Nordenskiold, Gustav, cited, 112.
Nordenskiold, Otto, cited, 24, 96, 189,

191, 209, 210, 211, 213, 224, 225, 243,

252, 270, 281, 283, 284.
Nordenskiold shelf-ice tongue, 231, 234,

251, 280.

Northeast Foreland, Greenland, 178.
North East Land, inland-ice of, 112-113 ;

map of, 110; peculiar precipitations

over, 276.
Northern Lapland, surface features of,

71.
North Wales, 32.

298

INDEX

Norway, cirques of, 13.
Norwegian ice-caps, 101.
Norwegian tind, formation of, 78.
Nourishment of Greenland glacier, 131,

143.

Nova Zembla, 110 ; map of, 109.
Nunataks, 73 ; in surface of Folgefond,

view of, 76 ; of Greenland, 125.
Nussbaum, cited, 68, 69.

Ocean currents, in relation to glaciers, 99.

Olriks Bay, Greenland, kames of, 140.

Osar, formation of, 89.

Outlet glacier, denned, 221.

Outlet glaciers, 104; dead, 253; of

Greenland, 126 ; Greenland, map of,

125.

Outlets, from Antarctic glaciers, 221.
Outwash apron, 87.

Overdeepening, in glaciated valleys, 66.
Overthrusting, on margins of Greenland

ice, 140, pi. 23.

Pack-ice, Antarctica, 198, 200, 203.

Palander, cited, 4.

Palisade ridge (comb-ridge), 32.

Palmer Land, 211.

Pancake ice, 208.

Parallel crevasses, in Greenland glacier,
view of, 129.

"Parallel roads," of Scottish Glens, 172.

Passarge, S., cited, 10.

"Paternoster" Lakes, 60.

Peary, Robert E., cited, 4, 120, 123, 126,
129, 130, 131, 132, 133, 134, 141, 145,
146, 150, 153, 154, 155, 160, 164, 166,
169, 170, 176, 207, 209, 252, 255, 266,
276.

Penck, Albrecht, cited, 10, 11, 13, 18, 23,
40, 56, 60, 61, 69, 83, 84, 88, 96, 137,
212.

Philippi, Emil, cited, 198, 212, 227, 241,
243, 252.

Physiographic form of glaciers, 285.

Physiography of Greenland continental
glacier, 119.

"Pie crust" snow, surface, 263.

Piedmont aprons, dead, 280.

Piedmont glaciers, 43, pi. 10.

"Piedmont" (ice-foot) glaciers, 209.

"Piedmonts afloat," 231.

Pillsbury, Admiral John E., cited, 212.

"Planks," in sea-ice, 204.

Playfair, Sir John, cited, 67.

Pleistocene glaciation, characteristic ero-
sion from, 72.

Plucking, 9.

Polar regions, contrasted, 186.

Posadowsky Bay, 203, 241.

Poudrin, 273.

Precipitous rock face, characteristic of

sculpture by mountain glaciers, 91.
Pressure, in pack-ice, 200.
Pressure ridges, in sea-ice, 204 ; views

of, 204-205.

Priestley, R. E., 242, 243.
Prince Rudolph Island, ice-cap, 106, 107 ;

view of, 108.
Profiles, of sub-aerial and glaciated

valleys, 68.

"Protection" by glaciers, 28, 66.
Purity Range, of Selkirks, Frontispiece.

Quadrant Mountain, Yellowstone Na-
tional Park, views of, pis. 2, 3 ; map
of, 27 ; nivation upon, 20.

Quensel, P. D., cited, 177.

Rabot, cited, 56.

Racovitza, E., cited, 199, 211, 212, 262,

282, 283.

Radiating glaciers, 52.
Rainbow with halo, 277.
Ramsey, cited, 40.
Randsee, 174.
Rate of movement, Greenland glacier

outlets, 134.
Receding hemicycle of glaciation, 6, 89,

280; Greenland, 143-144.
Receding of Ross Barrier, 227.
Recessional moraines, 87.
Recession, of cirque, 12.
Reconstructed glaciers, 50.
Rectangular crevasses in ice of Kaiser

Wilhelm Land, 130.
Refrigerating air engine, over continental

glacier, 269, 286.
Reid, H. F., cited, 180, 181, 185, 238,

243.

Reid's theory of iceberg formation, 181.
Relation of cirque to Bergschrund, 14.
Remnantal tableland, figure after At-

wood, 27.

Retirement of glacier, up valley, 89.
Reusch, H., cited, 15, 23.
Rhone glacier, 91.
Ribbon falls, 48.

Richter, E., cited, 13, 14, 15, 17, 23.
Richtofeneis, of Kerguelen Island, 43.
Riegel, 63, 90.
Rimaye. See Bergschrund.
Rink, Henry, cited, 149, 160, 175, 177.
Robertson Bay, 230, 231.
Roches moutonnees, 39, 64 ; of Antarc-
tica, 281.
Rock bars, 63, 90.
Rock basement, beneath Greenland

glacier, 125.
Rock basin lakes, 60.
Rock flows, from abandoned cirques, 94.

INDEX

299

Rock glaciers, 94 ; of Alaska, 96.

Rock pedestals (enclosed by fjords), 75.

Rock slides, near Flims, view of, 93.

Rock streams, in vacated valley, 91, pi.
19 ; in San Juan Mountains, map of,
95.

Rocky Mountains, foehn winds of, 271.

Ross Barrier, 193, 196, 216; evidence
for floating of, 221 ; face of, pis. 29,
31 ; inner margins of, 221 ; map of
margin of, 217 ; margins, views of,
219; material of, 218; movement of,
222, 224; nourishment of, 221, 272;
old and new faces on, 235 ; outline
map of, 220 ; recession of, 227 ; surface
of, 220.

Ross, Sir James, cited, 190, 193, 198,
215, 216, 218, 238, 262, 264, 265.

Royds, C. W., cited, 220, 282, 283.

Bundling, 39.

Russell's theory of iceberg formation,
180.

Russell, I. C., cited, 6, 11, 13, 23, 43, 56,
57, 58, 59, 68, 88, 96, 180, 184, 185.

Russian Lapland, glaciation of, 72.

Ryder, C. H., 141, 161, 170, 176.

Sago snow, 262.

"Sahara of snow," view of, 151.

Salisbury, R. D., cited, 12, 22, 60, 118,
142, 153, 160, 170, 185.

Sand dune, marginal view of, 273.

San Juan Mountains, rock flows in, 94.

San Rafael glacier, Chili, 44.

Sapper, Carl, cited, 118.

Sapping process, in cirque recession, 91.

Sastrugi, 154, 158 ; Antarctica, 203, 267,
268, 273; on schollen ice, view of,
204 ; on shelf -ice, pi. 30.

Saussure, H. B. de, cited, 2.

Scape colks, 135, 140.

Scattered knobs, a result of high latitude
glaciation, 72.

Scheuchzer, cited, 3.

Schollen ice, 200, 203.

Schrader, F. C., cited, 57.

Schrund-line, 18 ; continued down valley,
64 ; view of, after Gilbert, 18.

Scoresby, cited, 213.

Scott, R. F., cited, 189, 190, 191, 193,
200, 209, 210, 211, 212, 213, 216, 218,
220, 242, 243, 244, 254, 255, 257, 260,
266, 267, 276, 280, 282, 284.

Scottish highlands, temperature neces-
sary for glaciation, 5.

Sea-ice, Antarctica, 186, 198; forma-
tion of, 199, 206, 207; manner of
growth of, 208; thickening of, 226,
250 ; thickness of, 199, 200.

Seal Islands, 224.

Section, across Vatna Jokull, 274; of
ice grains in water, precipitated in
North East Land, 277 ; of Great Ross
Barrier, 217; marginal portion of ice-
cap, 101.

Sections, across Antarctic glacier margin,
253, 254 ; across margins of Green-
land glacier, 123 ; comparative, across
Greenlandic and Antarctic continental
glaciers, 255.

Selkirks, 30.

Seter, 172.

Seven Sisters, view of, 77.

Shackleton, Sir Ernest, cited, 147, 189,
191, 193, 200, 210, 211, 220, 232, 242,
243, 254, 256, 258, 266, 272, 274, 276,
282, 283, 284.

Shaping of Antarctic glacier margins, by
wind, 273.

Shelf-ice, 214 ; alimentation of, 221, 226,
227; density of, 218; how formed,
288 ; nature and distribution of, 214.

Shelf-ice tongue, supposed section of,
233.

Shelf-ice tongues, 230, 231.

Sheridan glacier, in Alaska, 45.

Sherzer, W. H., cited, 55, 57, 58.

Sierra Nevadas, California, glacial sculp-
ture in, pi. 15.

Sinking of "Antarctica," 205.

Sir John Murray glacier, 230, 231, 233.

Sjogren, O., cited, 71, 73.

Sketch map of north border of Alpine
Highland, 85.

Skottsberg, C. J., cited, 213.

Sky, in interior of Greenland, 144 ;
nature of, during snowfall, 262.

Slabs, ice, 259.

Sledge journeys, of Peary in north Green-
land, map, 133.

Slope glaciers, 209.

Snow, Antarctic, in summer season, 5 ;
blown off Antarctica into sea, 280 ;
compressed, in Ross Barrier, 218 ; den-
sity of, 157 ; drifting, over Greenland
glacier, 151; "pie crust," 263; pre-
cipitated through mixing of surface
air with descending currents, 277 ;
smooth-sledging type, 263 ; structure
of, on surface of Greenland glacier,
153 ; transported by wind, 271.

Snow barchans, 156.

Snowdrift forms, 154.

Snowdrift site, figured after Matthes, 19.

Snow dunes, on margin of Greenland
glacier, 132.

Snowfall, character of, what dependent
upon, 155 ; Greenlandic, source of in
cirrus clouds, 158 ; in Antarctica, 223 ;
in Greenland, 144 ; in interior of Ant-

300

INDEX

arctica, 264 ; of Antarctica, in sum-
mer months, 226 ; nature of, in Ant-
arctica, 262.
Snowflakes, nature of in relation to

temperature of precipitation, 262.
Snow Hill Island, 189, 226, 270.
Snow-line, defined, 5.
Snow precipitation, at end of blizzard,

279.
Snow sweepings, from Antarctic glacier,

251.

Sobral, J. M., cited, 224.
Solifluction, process of, 21, 94.
Soundings, about Antarctica, 196, 197,

198, 218, 245; Antarctica, map of,

196.

Spethmann, Hans, cited, 104, 118, 274.
Spitzbergen, expedition to in 1858, 4 ;

inland-ice of, 111; map of, 110.
Spitzbergen type of glacier, 210.
Staircase, due to successive landslides,

92.

Steffen, Hans, cited, 177.
Stein, Robert, cited, 160, 161, 170, 176.
Steps, in glaciated valley, 61.
Stille, H., cited, 243.
Stone rivers, 94.
Stratification, in ice island, pi. 28 ; of

continental glacier, 248, pi. 22.
Stream action, in valley while glacier

retires, 89 ; on mountain foreland,

85.
Streams, braided, 86 ; lateral, of outlet

glaciers, 169.
Sturge Island, 209.
Subglacial drainage, on Greenland, glacier,

170.

Subglacial streams, Antarctica, 234.
Submarine wells, in fjord heads, 175.
Submerged continental platform, about

Antarctica, 196.

Suess, E., cited, 135, 136, 142, 172, 176.
Supposed south polar anticyclone, 265.
Superglacial debris, on Antarctic glaciers,

259.
Superglacial streams, on Greenland

glacier, map of, 165.
Surface moraines, Greenland, 135 ; cross

section of, 139 ; view of, 139.
Sverdrup expedition, 117.
Sverdrup, Otto, cited, 118.
Swedish polar expedition of 1872-1873,

4.
Swirl colks (ice eddies), 137.

Tabular icebergs, views of, 236, 237.

" Tapioca " snow, 262.

Tarr, R. S., cited, 46, 48, 56, 57, 58, 88,

105, 128, 138, 141, 142, 162, 173, 176,

177, 185.

Temperature, its relation to glaciation,

36.
Temperatures, air, Antarctic, 188-189,

262; in relation to glaciation, 278;

over inland-ice of Greenland, 145 ;

serial subsurface, in Greenland glacier,

163 ; serial subsurface, in Ross Barrier,

221.
Terraced margin, of Greenland glacier,

130.
"Terraces," of King Oscar Land, 224;

West Antarctica, origin of, 250.
Thomson, Wyville, cited, 236, 238, 240,

243, 264.
Thoroddsen, Th., cited, 56, 102, 103,

104, 118, 274, 284.
Tide-cracks, 208.
Tide-water glaciers, 51.
Tinds, development of, 78, 79, 286 ; pis.

18, 34; remarkable circular one from

Lofoten Islands, view of, 78.
Tongue-like basin, before mountain front,

83, 88.

Torell, Otto, cited, 4.
Transection glacier, former over Grimsel

pass, 45.

Transection glaciers, 44.
Transportation of snow, by wind, 271.
Tresca, cited, 201.
Triest glacier, pi. 12.
Trogthal, 62.
Trolle, Lieutenant A., cited, 127, 141, 160,

163, 176.

Tschirwinsky, P. N., cited, 161.
Tuktoo glacier, pi. 23.
Turner glacier, Alaska, 52.
Turtle Mountain, landslide from, 92.
Tyndall, John, cited, 12, 22, 91, 96.
Tyrrell, J.'B., cited 236, 243.

Uinta Mountains, cirque cutting in, 26.

Umanak Fjord glacier outlet, Greenland,
125.

Upland, dissected by glaciers, 25.

Uplifts in connection with glacial sculp-
ture, 74.

Upper air currents, function in nourish-
ing Antarctic glaciers, 269.

Urville, J. S. C. Dumond de, cited, 190,
193, 210.

U-valleys, 63; initiation of, 20; over-
emphasis upon, 8 ; Wasatch Range,
pi. 16.

"Valley" glaciers, 47.

Vatna Jokull, 43, 103, 274 ; air circula-
tion over, 278 ; cross section of, 104 ;
map of, 103 ; map of margin of, pi.
21.

Venetz, cited, 3.

INDEX

301

Vernagt glacier, 91.
Victoria glacier, 50.

Wallace, A. R., cited, 13, 23.
Wandel Island, 199, 200, 207.
Warm season, effect of, on Greenland

glacier, 163.
Wasatch Range, pi. 16.
Water basins, on Greenland glacier, 166.
Water fountain, on Greenland glacier,

170.

Water sky, in ribbons, cause of, 202.
Weddell, James, cited, 192.
Weddell Sea, 189, 193, 223, 244.
Wenkchemna glacier, pi. 14.
Werth, Emil, cited, 56.
West Antarctica, 192, 199, 209, 238, 251.
"West-ice," junction with sea-ice, view

of, 228; of Kaiser Wilhelm Land,

227; origin of, 250; stranded, 229;

surface of, 229 ; view from sea, 228.
Wheeler, A. O., cited, 49, 50.
White Island, 189.
Whymper, Edward, cited, 150.
Widening of glacier valleys at mouths,

66.

Wilkes, Captain Charles, cited, 190, 192,
193, 211, 212, 215, 238, 242, 243, 244,
264.

Wilkes Land, 192, 193, 195, 240.

Wilson, E. A., cited, 280.

Wind directions, over Antarctic glacier,
266, 267.

Wind poles, 281.

Wind transportation of snow, over
Greenland ice, 150.

Winds, Antarctic, sweep inland-ice clear
of snow, 247 ; importance of in dis-
tribution of snow, 28, 151, 271 ; in
relation to alimentation of Antarctic
glacier, 226 ; on border of Greenland,
149 ; prevailing, on margins of Ant-
arctica, 263, 264.

Workman, Fanny Bullock, cited, 57.

Workman, Wm. Hunter, cited, 57.

Yoho glacier, pi. 3.

Zigzag leads, in drift ice, 202.
Zungenbecken, 83.
Zusammengesetzte Gletscher, 52.

RETURN EARTH SCIENCES LIBRARY
TO* 642-2997

LOAN PERIOD 1
1 MONTH

ALL BOOKS MAY BE RECALLED AFTER 7 DAYS

Books needed for class reserve are subject to immediate recall

DUE AS STAMPED BELOW

FORM NO. DD8

UNIVERSITY OF CALIFORNIA, BERKELEY
BERKELEY, CA 94720

U.C. BERKELEY LIBRARIES