accompanying map as Fountain stream. This comes to the surface
through a rudely circular opening, nearly 100 feet in diameter, surrounded
in part by ice. Owing to the pressure to which the waters are subjected,
they boil up violently, and are thrown into the air to the height of 12 or
15 feet, and send jets of spray several feet higher. The waters are brown
with sediment, and rush seaward with great rapidity, forming a roaring
stream fully 200 feet broad, which soon divides into many branches, and
is spreading a sheet of gravel and sand right and left into the adjacent
forest. Where Fountain stream rises the face of the glacier is steep and
covered with huge boulders, many of which are too large for the waters
to move. The finer material has been washed away, however, and a
slight recession in the face of the ice bluff has resulted. The largest
stream draining the glacier is the Yahtse. This river, as already stated,
rises in two principal branches at the base of the Chaix hills, and flowing
through a tunnel some six or eight miles long, emerges at the border of
the glacier as a swift brown flood fully 100 feet across and 15 or 20 feet
deep. The stream, after its subglacial course, spreads out into many
branches, and is building up an alluvial fan which has invaded and buried
several hundred acres of forest.
In traversing the coast from the Yahtse to Yakutat bay, we crossed a
large number of streams which drain the ice fields of the north, some of
which were large enough to be classed as rivers. When the streams, on
flowing away from the glacier, are large, they divide into many branches,
as do the Yahtse and Fountain, and enter the sea by several mouths.
When the streams are small, however, they usually unite to form large
rivers before entering the ocean. The Yahtse and Fountain, as we have
seen, are examples of the first, while Manby and Yahna streams are
examples of the second class. Manby stream rises in hundreds of small
springs along the margin of the glacier, which flow across a desolate,
GLACIERS OF ALASKA. 123
torrent-swept area, and unite just before reaching the ocean into one
broad, swift flood of muddy water, much too deep for one to wade.
On the border of the glacier facing Yakutat bay, however, the drain-
age is different. The flow of the ice is there eastward, although the
margin is probably stagnant, and instead of forming a bold, continuous
escarpment, ends irregularly and with a low frontal slope. The principal
streams on the eastern margin in 1891, were the Osar, Kame, and Kwik.
Each of these issues from a tunnel and flows for some distance between
walls of ice. Of the three streams mentioned, the most interesting is the
Kame, which issues as a swift brown flood partially choked with broken
ice from the mouth of a tunnel, and flows for half a mile in an open cut
between precipitous walls of dirty ice 80 to 100 feet high. This is the
longest open drainage channel that I have yet seen in the ice. It is about
50 feet broad where the stream rushes from the glacier, but soon widens
to several times this breadth. Its bottom is covered with rounded gravel
and sand, and along its sides are sand-flats and terraces of gravel resting
upon ice. The swift, muddy current was dotted with small bergs stranded
here and there in the center of the stream, showing that the water was
shallow. Evidently the stream has a long subglacial course, and carries
with it large quantities of stones, which are rounded as in ordinary rivers.
Gravel and sand are being rapidly deposited in the ice channel through
which it flows after emerging from its tunnel. Broad sand-flats are being
spread out in the lakes and swamps two or three miles to the east. The
stream is some four or five miles in length, and near Yakutat bay
meanders over a barren area perhaps a mile broad. I have called it Kame
stream because of a ridge of gravel running parallel with it, which was
deposited during a former stage, when the waters flowed about 100 feet
higher than now and built up a long ridge of gravel on the ice which
has all the characteristics of the kames in New England. In the more
definite classification of glacial sediments now adopted, this would more
properly be called an osar.
Near the shore of Yakutat bay, the streams from the glacier spread
out in lagoons and sand-flats, where much of the finer portion of the
material they carry is deposited. Sometimes this debris is spread out
above the ice, and forms level terraces of fine sand and mud which become
prominent as the glacier wastes away.
Osars. The drainage of the glacier has not been investigated as
fully as its importance demands, but the observations already made seem
124 GLACIERS OF NORTH AMERICA.
to warrant certain conclusions in reference to deposits made within the
glacier by subglacial or englacial streams.
When the streams from the north reach the glacier they invariably
flow into tunnels and disappear from view. The entrances to the tunnels
are frequently high arches, and the streams flowing into them carry along
great quantities of gravel and sand. About the southern and eastern
borders of the glacier, where the streams emerge, the arches of the tunnels
are low, owing principally to the accumulation of debris which obstructs
their discharge. In some instances, as at the head of Fountain stream,
the accumulation of debris is so. great that the water rises through a
vertical shaft in order to reach the surface, and rushes upward under
great pressure. The streams flowing from the glacier bring out large
quantities of well-rounded sand and gravel, much of which is immediately
deposited in alluvial cones. This much of the work of subglacial streams
is open to view, and enables one to infer what takes place within the
tunnels, and to analyze, to some extent, the processes of stream depo-
sition beneath glacial ice.
The streams issuing from the ice are overloaded, and besides, on
emerging, frequently receive large quantities of coarse debris from the
adjacent moraine-covered ice cliffs. The streams at once deposit the
coarser portion of their loads, thus building up their channels and obstruct-
ing the outlets of the tunnels. The blocking of the tunnels must cause
the subglacial streams to lose force and deposit sand and gravel on the
bottom of their channels ; this causes the water to flow at higher levels,
and, coming in contact with the roofs of the tunnels, enlarges them
upwards ; this in turn gives room for additional deposits within the ice
as the alluvial cones at the extremities of the tunnels grow in height.
In this way narrow ridges of gravel and sand, having, perhaps, some
stratification due to periodic variations in the volume of the streams, may
be formed within the ice. When the glacier melts, the gravel ridges con-
tained within it will be exposed at the surface, and as the supporting walls
melt away, the gravel at the top of the ridge will tend to slide down so
as to give the deposit a pseudo-anticlinal structure. Ridges of gravel
deposited in tunnels beneath the moraine-covered portion of the Malaspina
glacier would have boulders dropped upon them as the ice melts, but
where the glacier is free from surface debris there would be no angular
material left upon the ridges when the ice finally disappeared. Such a
system of deposition as is sketched above would result in the formation
of narrow, winding ridges of cross-bedded sand and gravel, corresponding,
GLACIERS OF NORTH AMERICA.
FIG. A. MORAINE-COVERED BORDER OF MALASPINA GLACIER, FROM
(Drawn from a Photograph.)
FIG. B. ENTRANCE TO TUNNEL IN MALASPINA GLACIER.
(Drawn from a Photograph.)
GLACIERS OF ALASKA. 125
seemingly, in every way to the osars of many glaciated regions. The
process of subglacial deposition pertains especially to stagnant ice sheets
of the Malaspina type, which are wasting away. In an advancing glacier
it is evident that the conditions would be different, and subglacial erosion
might take place instead of subglacial deposition.
Alluvial Cones. Below the outlets of the tunnels through which
Malaspina glacier is drained, there are immense deposits of boulders, gravel,
sand, and mud which have the form of segments of low cones. These
deposits are of the nature of the " alluvial cones " or " alluvial fans " so
common at the bases of mountains in arid regions, and are also related to
the " cones of dejection," deposited by torrents, and to the subaerial
portion of the deltas of swift streams. As deposits of this nature have
not been satisfactorily classified, I shall, for the present, call them " alluvial
As stated in speaking of osars, the streams issuing from tunnels in
Malaspina glacier at once begin to deposit. The larger boulders and
stones are first dropped, while gravel, sand, and silt are carried farther
and deposited in the order of their coarseness. The deposits originating
in this way have a conical form, the apex of each cone being at the mouth
of a tunnel. As the apexes of the cones are raised by the deposition of
coarse material, their peripheries expand in all directions, and, as the
region is densely forest-covered, great quantities of trees become buried
beneath them. As the ice at the head of an alluvial cone recedes, the
alluvial deposit follows it by deposition on the up-stream side. The growth
of the alluvial cones will continue so long as the glacier continues to
retreat, or until the streams which flow over them have their subglacial
courses changed. The material of the alluvial cones is as heterogeneous
as the material forming the moraines on the border of the glacier about
which they form, but the greater, and practically the entire, accumulation
is more or less rounded and waterworn. Cross stratification characterizes
the deposits throughout, and on the surface of many of the cones and
probably in their interior, also, there are large quantities of broken tree
trunks and branches. The coarse deposits first laid down on a growing
alluvial cone are buried beneath later deposits of finer material in such a
way that a somewhat regular stratification may result. A deep section
of one of these deposits should show a gradual change from fine material
at the top to coarse stones and subangular boulders at the bottom. Their
outer borders are of fine sand and mud, and when the distance of the
126 GLACIERS OF NORTH AMERICA.
ocean is sufficient, the streams flowing from them deposit large quantities
of silt on their flood plains. The very finest of the glacial mud is
delivered to the ocean and discolors its water for many miles from land.
The formation of alluvial cones about the border of a stagnant ice
sheet and the deposition of ridges of gravel within it, have an intimate
connection, and are, in fact, but phases of a single process. The growth
of an alluvial cone tends to obstruct the mouth of the tunnel through
which its feeding stream discharges ; this causes the stream to deposit
within the tunnel ; this again raises the stream and allows it to build its
alluvial cone still higher. In the case of Malaspina glacier, where this
process has been observed, the ice sheet is stagnant, at least on its border,
and is retreating. The ground on which it rests is low, but is thought to
be slightly higher on the southern margin of the glacier than under its
central portion. The best development of alluvial cones and osars would
be expected in a stagnant ice sheet resting on a gently inclined surface,
with high lands on the upper border from which abundant debris could
be derived. These ideal conditions are nearly reached in the example
Glacial and Ocean Records. Much has been written concerning
the character of the deposits made by glaciers when they meet the ocean,
but so far as can be judged from the conditions observed about the borders
of Malaspina ice sheet, the sea is much more powerful than the ice.
Where the two unite their action, the sea leaves the more conspicuous
records. The waters are active and aggressive, while the glacier is
passive. Where the glacier enters the ocean its records are at once
modified and to a great extent obliterated. The presence of large boulders
in marine sediments or in gravels and sands along the coast is about all
the evidence of glacial action that can be expected under the conditions
referred to. Where the swift streams from, the Malaspina glacier enter
the ocean, the supremacy of the waves, tides, and currents is even more
marked. The streams are immediately turned aside by the accumulation
of sandbars across their mouths, and nothing of the nature of stream-worn
channels beneath the level of the ocean can exist. All of the deposits
along the immediate shore between the Yahtse and Yakutat bay have the
characteristic topographic features resulting from the action of waves and
currents, and do not even suggest the proximity of a great glacier.
Recent Advance. On the eastern margin of Malaspina glacier,
about four miles north of Point Manby, there is a locality where the ice
GLACIERS OF ALASKA. 127
has recently advanced into the dense forest and cut scores of great spruce
trees short off and piled them in confused heaps. After this advance the
ice retreated, leaving the surface strewn with irregular heaps of boulders
and stones, and enclosing many basins, which, at the time of our visit, were
full of water to the brim. The glacier, during its advance, ploughed up
a ridge of blue clay in front of it, thus revealing, in a very satisfactory
manner, the character of the strata on which it rests. The clay is thickly
charged with sea-shells of living species, proving that the glacier, during
its former great advance, probably extended to the ocean, and that a rise
of the land has subsequently occurred. This is in harmony with many
other observations which show that the coast adjacent to Malaspina
glacier is now rising. The blue color of the subglacial strata is in marked
contrast with the browns and yellows of the moraines left on its surface
by the retreating ice, which, in common with the fringing moraines still
resting on the glacier, show considerable weathering. Among the shells
collected in the subglacial clay, Dr. W. H. Dall has identified the
Cardium gronlandicum, Gronl.
Cardium islandicum, L.
Kennerlia grandis, Dall.
Leda fossa, Baird.
Macoma sabulosa, Spengler.
Similar shells, all of living species, were previously found at an
elevation of 5000 feet on the crest of a fault scarp at Pinnacle pass,
showing that recent elevations of land, much greater than the one
recorded in the marine clay just noticed, have taken place. In fact there
are several indications that the coast in the vicinity has been rising, and
that the same process is still continuing.
SUBSOIL IcE. 1
On several occasions while traveling in central and northern Alaska,
I found, by removing a few inches of the moss which generally covers
the ground, that the subsoil was solidly frozen. This occurrence was
especially striking on summer days when the temperature of the air in
the shade was frequently between 90 and 100 of the Fahrenheit scale.
1 Observations on the subsoil ice of Alaska by the present writer, together with references
to a number of previously published papers on the same subject, may be found in the Bulletin
of the Geological Society of America, vol. 1 , pp. 125-133.
128 GLACIERS OF NORTH AMERICA.
Along the Yukon, from its mouth to near its source, one may fre-
quently see strata of clear ice, or more frequently of black, dirt-stained
ice and frozen gravel several feet thick, in the freshly cut banks of the
stream. In general, throughout the low-lying portions of central Alaska,
subsoil ice exists at a depth of but a few inches beneath the forest-covered
surface. The maximum thickness of this permanently frozen layer is not
known, but in a few instances of which I have authentic information, it
has been penetrated to a depth of 25 feet without reaching the bottom.
Explorations conducted by Lieut. J. C. Cantwell, 1 of the U. S.
Revenue Marine service, along the Kowak river, which flows into
Kotzebue sound, about 260 miles north of the mouth of the Yukon, and
just within the Arctic circle, have shown that a layer of subsoil ice from
100 to 200 feet thick has there been cut into by the streams so as to
leave steep bluffs of solid ice along their borders. The ice covers the
land like a stratum of rock, and has been dissected by stream erosion in
much the same manner that river channels are corroded in other regions.
Above the ice there is a thin covering of rich black soil supporting a
growth of mosses, grasses, and trees. Instructive illustrations of the Kowak
river flowing between precipitous, canon-like walls of ice are presented in
the report just referred to.
One of the most striking exposures of subsoil ice in Alaska, and one that
has been described by many travelers, exists on the shore of Eschscholtz
bay at the head of Kotzebue sound. The ice there forms a bold bluff,
and has been estimated to be from 150 to 300 feet thick. It is covered
with rich humus, on which grasses grow luxuriantly. In this instance, as
in several other similar examples that are known, the ice contains the
bones of the mammoth and other large animals that are now extinct.
The soil which accumulates as the ice melts, owing to the concentration
at the surface of the impurities it contains, has a strong odor of decaying
animal matter. We are thus assured that the subsoil ice, in certain
instances, and probably over ' extensive areas, was formed at a time so
remote that the animals then inhabiting the country in great numbers
have since become extinct.
The ice in the banks of the lower Yukon, and in the vicinity of
Kotzebue sound, is a part of a vast sheet of frozen subsoil that underlies
large portions of the low, marshy region fringing the shores of Bering
1 "A Narrative Account of the Exploration of the Kowak River, Alaska," in Report of
the Cruise of the Revenue Marine Steamer Corwin, in the Arctic Ocean, in the year 1885,
by Capt. M. A. Healy, Treasury Department, Washington, D.C., 1887.
GLACIERS OF ALASKA. 129
sea and the Arctic ocean. This tundra, as it is termed, covers many
thousands of square miles. It is bright with mosses and a profusion of
low, flowering plants during the short Arctic summer, but beneath its
luxuriant carpet of verdure the ground is always frozen.
Still more extensive tundras cover the low lands forming the Arctic
shores of Asia, and have there been penetrated to a depth of nearly 400
feet without reaching the bottom of the subsoil ice. It is in this deposit
that the complete carcasses of the mammoth and of the woolly rhinoceros
are found from time to time. The fossil ivory gathered along the banks
of the rivers in this Arctic region is said to be even more important as an
article of commerce than the elephant tusks obtained in the jungles of
It is not probable that all of the subsoil ice of northern regions has
been formed in one way. Along the flood plains and on the deltas of
rivers where layers of clear ice are interbedded with sheets of frozen
gravel and vegetable matter, as is frequently the case, it seems evident
that the growth of the deposit is due, in some instances, to the flooding
of previously frozen layers, and the freezing and subsequent burial of the
sediment thus added to their surfaces. When spring freshets spread out
sheets of debris over the flood plain of a river, as frequently happens when
streams in high latitudes flow northward, the previously frozen soil and
the ice of ponds and swamps may be buried and indefinitely preserved.
During the succeeding winter the surface layer thus added would itself
become frozen, and perhaps in its turn become buried beneath later
deposits of the same character at intervals of one or more years.
On the tundras the luxuriant growth of vegetation that starts into life
as soon as the winter's snow has disappeared, and grows rapidly during
the long, hot summer days, dies below and partially decays, but becomes
frozen and has its complete destruction arrested, while the dense mat of
roots and stems above continues to thrive. In this way an accumulation of
partially decayed vegetable matter is formed, which increases in thickness
from year to year by additions to its surface. The process is similar to
that by which peat bogs are formed in temperate latitudes, except that
the partially decomposed vegetation becomes solidly frozen. It is in
reality an example of cold storage on a grand scale. This slow accumu-
lation in northern regions of vegetable matter, together with the bones
and even complete carcasses of animals, is truly a wonderful process.
Under existing climatic conditions there does not seem to be any limit
to the depth such deposits may attain. The amount of carbonaceous
130 GLACIERS OF NORTH AMERICA.
material already accumulated in the tundras of America and Asia must
equal that of the most expensive coal field known. In view of these
facts it does not seem an unreasonable suggestion that some coal seams
might have originated from the vegetable matter accumulated in ancient
There is still another process by which frozen subsoil may be formed
in high latitudes : this is, the effects of the cold during the long winters
are not counteracted by the heat during the short summers. Under the
conditions now prevailing in northern Alaska, where the mean annual tem-
perature is below 32 Fahrenheit, the frozen layer tends to increase in thick-
ness from year to year just as the depth of frozen soil in more temperate
latitudes may increase from month to month during the winter season.
During the short northern summers, especially where the ground is moss-
covered, melting only extends a few inches below the surface.
Computations made by Prof. R. S. Woodward 1 have shown that the
freezing of even the deepest ice stratum yet discovered in Arctic regions
might have resulted in the course of a few thousand years from a mean
annual temperature no lower than that prevailing in northern Alaska at
the present time.
The subsoil ice described above lacks most of the characteristics of
glaciers and should not be included among them. It is so closely analo-
gous, however, to the condition reached by continental and piedmont
glaciers when they become stagnant and are wasting away, that the mode
of its formation needs to be understood in order that the two may not be
confounded. The climatic conditions admitting of the accumulation of
subsoil ice are similar, and probably identical, to those which initiate
glacial periods. In times immediately preceding the formation of conti-
nental glaciers, it is possible that subsoil ice like that of the tundras may
be formed over extensive regions before they become covered by a flowing
ice sheet. In such an instance the frozen subsoil might become a part of a
continental glacier when covered by the advancing ice. During the
amelioration of climate following an ice invasion, the tundra phase might
again return, so that a glacial period would be both preceded and followed
by a time when the mean annual temperature would favor the existence
of deeply frozen subsoil, and, at the same time, admit of the growth of
1 Geological Society of America, Bulletin, vol. 1, pp. 131, 132.
GLACIERS IN THE GREENLAND REGION.
Grinnell Land. Explorations in the extreme northeastern part of
North America have been carried on principally along the navigable water
ways. It is only recently that a knowledge has been gained of the vast
snow fields in which the glaciers descending to the sea have their origin.
The most instructive journeys that have been made on the islands to the
west of Baffin's bay, Davis strait, etc., were by members of the Lady
Franklin Bay expedition in 1882. Gen. A. W. Greely, in his admirable
account of " Three Years of Arctic Service," describes the United States
mountains in the northern part of Grinnell land, as being buried beneath
neve snow, and apparently presenting much the same appearance as the
desolate region to the north of Mount St. Elias, described in the preceding
The largest ice stream draining the snow fields of Grinnell land yet
discovered, known as the Henrietta Nesmith glacier, flows south and
terminates near Lake Hazen. The most striking feature of this great
glacier, and one that seems to be characteristic of many of the ice streams
of the far north, is the extremely precipitous slope in which it terminates.
As shown in an illustration published by General Greely, its extremity is