ASTRONOMICAL CONSIDERATIONS
The Great Ice Age—Extension of glaciation—Croll’s
ingenious hypothesis—Inadequacy of the explanation—The
ice caps of Mars—De Geer’s discovery of the annual layers
of glacial clay in Sweden—Application to the measurement
of time—Sederholm’s observations in Finland—Cyclic
sedimentation in the Cretaceous strata of Colorado—Its
correlation by Gilbert with an astronomical time unit.
In the last chapter mention was made of Kelvin’s work on tidal
friction and of Sir G. Darwin’s speculations as to the early history
of the earth-moon system. Their conclusions depended very largely on
a cosmogony which has failed to meet with general acceptance. A very
different appeal to astronomical causes was made by Croll in his famous
attempt to account for the anomalous conditions of the glacial period.
Here we may also consider the correlation by Baron De Geer and Mr. G.
K. Gilbert of certain unusual phases of sedimentation with the seasonal
and climatic changes brought about respectively by the earth’s motion
and its periodic fluctuations. All of these investigations find a
common basis in their direct reference to astronomical considerations.
In the course of its history the earth has undoubtedly passed through
several periods of extreme cold, periods when Arctic conditions swept
down from their polar strongholds and invaded the temperate zones and
even the tropics. Ice and snow have left their traces in many a grooved
and striated surface, and in gently rounded outlines the landscape
often betrays their former presence. Erratics and perched blocks,
terminal and lateral moraines, lakes and alluvial terraces, U-shaped
and hanging valleys all bear witness to the wide extension of the
vanished glaciers of the past. Relics of glaciation have been preserved
in the Cambrian or pre-Cambrian rocks of Norway, China, Australia
and South Africa. Still more remarkable are the records of a Permian
ice age found in the rocks of South America, South Africa, India and
Australia. No unequivocal evidence of glaciation in later periods is
forthcoming until the advent of the Pleistocene. The Great Ice Age
through which the earth has so recently passed was not, however, a
single epoch of glacial extension. Four episodes in the climatic cycle
can be recognised—a genial episode, a period of falling temperature
and glacial advance, a glacial episode, and finally, a period of rising
temperature and glacial retreat. That this cycle has been four times
repeated is the testimony of the Eastern Alps as interpreted by Prof.
Penck. In the British area the ebb and flow of temperature has not left
so clear a record, and a more continuous and persistent glaciation
appears to be indicated.
At the time of the maximum advance of the ice, all northern Europe lay
buried beneath an immense ice sheet, which was fed by enormous glaciers
slowly creeping down from the uplands. Curiously enough, Siberia, which
in parts experiences the coldest winter in the world, is not now and
was not then covered by the ice-field. Farther south the ice collected
in the Pyrenees, Alps, Caucasus and in the great Asiatic ranges
and descended in all directions far below the level of the present
snow-line. In Canada and the United States the ice mantle proceeded
from three great centres—from Labrador along the Atlantic coast,
from the Keewatin district by Hudson’s Bay and from the Cordilleras
along the Pacific coast. In the tropics and in the southern hemisphere
the story is the same. Down the slopes of Kenia, Kilima N’jaro and
Ruwenzori, the towering peaks of Central East Africa, the glaciers
descended 5000 feet below their present termination. Kosciusko in
New South Wales bears conspicuous traces of a former ice-cap which
completely shrouded all but the highest peaks of the plateau. From the
highlands of Tasmania and New Zealand, and from the Cordilleras and
Andes of Patagonia and Chili the evidences of a prolonged glaciation
are equally clear.
No satisfactory theory of climate has yet been propounded which affords
an adequate explanation of this universal lowering of the snow-line.
Whether or not the temperature fluctuations were synchronous all
over the earth is difficult to decide, but that a colder climate
characterised the southern no less than the northern hemisphere cannot
be doubted. The causes to which appeal has been made in accounting for
climatic changes are of three classes,—astronomical, geographical and
atmospheric. In general the various tendencies will be opposed to one
another and by their interference prevent the attainment of universal
extremities of climate. Occasionally, however, circumstances may
arise when their joint action will lead in a single direction. A wide
extension of tropical or polar conditions would then be expected to
follow.
Of the three contributory causes, only one, the astronomical, bears
any determinable relation with time. The theory that secular variation
of terrestrial climates results from the changing eccentricity of the
earth’s orbit, was first proposed by Adhémar. In 1868 James Croll
greatly elaborated and extended this theory, presenting it with such
a formidable array of quantitative data and yet in such an attractive
form that it exerted a considerable influence on the geological thought
of the day. Not only did it seem to offer an acceptable means of escape
from the maze of difficulties by which the problem was shrouded, but in
addition it promised a faithful chronology fixing the date and duration
of the Ice Age with almost ideal precision.
Laplace established the fact that the ellipticity of the earth’s
orbit is subject to periodic oscillations between certain limits.
Sometimes the path is nearly circular but at other times it becomes a
more flattened ellipse. If the earth were the only planet, its orbit
would suffer no change; that it does is due to the attractions of the
sister planets. Nevertheless, the orbit is essentially stable, and the
yearly journey always occupies the same time. Formulæ were devised by
Leverrier from which it was possible to calculate with some accuracy,
the actual value of the eccentricity at any given period in past or
future within a few million years of the present. Croll utilised these
formulæ to compute the dates of maximum and minimum eccentricity for
the past three million years. He found three important periods when
that factor was specially high, betraying a type of orbit more than
usually flattened. These were:
(_a_) from 2,500,000 to 2,600,000 years ago.
(_b_) ” 720,000 to 980,000 ” ”
(_c_) ” 80,000 to 240,000 ” ”
It was to the last of these that the Glacial Period was assigned.
When the earth is at perihelion, i.e. in that part of its orbit which
lies nearest to the sun, it enjoys a more generous radiation than
falls upon it at aphelion, when it is farthest from the sun. But this
alone does not determine the time of summer and winter. At present
the northern winter and southern summer occur when the earth is in
perihelion. As is well known, this is owing to the inclination of
the earth’s axis and the blanketing effect of the atmosphere. In the
northern hemisphere, the greater thickness of air presented to the
sun’s rays during the winter keeps out more heat than is gained by
the relative closeness to the sun. However, this condition is not
permanent. As discovered by Hipparchus in the year 134 B.C.,
the positions of summer and winter and of the equinoxes on the ecliptic
are subject to a slow forward movement. In 26,000 years they make the
complete circuit, and so in the course of time the relation of the
seasons to perihelion is slowly altered.
The _precession of the equinoxes_ was shown by Newton to be a
dynamical consequence of the spinning of the earth about a tilted axis.
While this change is progressing, the position of perihelion is also
shifting, and the resultant period is thereby reduced on the average
to 21,000 years. Thus, in about 10,000 years from now the northern
hemisphere will enjoy summer at perihelion. At the same time, the
northern winter will occur at aphelion. The cold season will then be
longer and more severe than now, and the annual accumulation of snow
correspondingly increased. If, in addition, we suppose the orbital
eccentricity to approach its maximum value, the northern hemisphere
would then be in the grip of winter for nearly four months of the year,
and the cold would become still more bitter than before. As a slight
compensation for the rigours of the winter, the summer, though short,
would be very much hotter. It was extremes such as these, accompanied
perhaps by favourable geographical conditions, that Croll postulated
for his glacial period. He believed that the snow and ice which would
collect during a long frosty winter would successfully resist the
evaporative powers of the summer, and that permanent snow-fields would
therefore arise. Once started, the snowy mantle would tend to continue.
A great deal of heat could be absorbed without raising the temperature
above freezing-point, and the result of evaporation would be the
creation of a thick blanket of fog, an effective agent in guarding the
ice against the ravages of the summer radiance.
A peculiarity of Croll’s hypothesis is that glacial epochs could not
exist in both hemispheres at the same time. South of the Equator the
summer would be longer than it is at present, and the winter would not
only be short, but also comparatively mild. The climate would be that
of a genial interglacial period. The alternation of glacial epochs
between north and south is a necessary consequence of Croll’s view, but
when the facts are examined they are found to be unfavourable to this
assumption. In Sweden the last period of extension of the ice appears
to have been synchronous with that in New South Wales, the close of
each being dated, if our present time-scale is reliable, at 15,000
to 20,000 years ago. Both date and coincidence are decisive against
the theory. Although Croll’s hypothesis achieved a great popularity,
geologists were not wanting who considered the suggested causes to
be utterly inadequate to produce so radical a change of climate. It
happened that during the penultimate period of extreme eccentricity,
which began nearly a million years ago, the astronomical conditions
were more favourable to glaciation than they have been since. The
question therefore arose why no traces had been preserved in the
deposits of that time. Indeed, since the Cambrian, many hundreds of
glacial periods should have come and gone. Had this been the case, and
corroborative evidence sufficiently convincing, the earth’s chronology
would have been written boldly in its rocks. But Croll’s brave attempt
to number the ages was unsuccessful, and his attractive theory no
longer holds the field.
[Illustration: FIG. 7.
At maximum: full extent of white.
At minimum: inner circle.
NORTH POLAR CAP.]
[Illustration: FIG. 8.
At maximum: full extent of white.
At minimum: disappears entirely.
SOUTH POLAR CAP.
MARS.]
A gigantic experiment illustrating the very conditions which Croll
postulated has been recently shown to us by Prof. Lowell. For his
working model we must look to the sky and carefully watch the changing
seasons of Mars. The eccentricity of the orbit of Mars is much higher
than that of the earth could ever have been, and moreover, the southern
winter falls near aphelion. The conditions for a permanent ice-cap
over the south polar regions are therefore ideal. The actual facts are
surprising, and on the accompanying plate Prof. Lowell has depicted
them admirably. During the winter a large snow-cap collects around the
south pole, but in summer it is entirely dispersed. Around the north
pole the winter snow-cap is less extensive, but all through the summer
it never quite disappears. While the long southern winter undoubtedly
makes possible a greater accumulation of snow, the hot short summer
more than compensates by its superior powers of evaporation. A
permanent and widely extended ice-field evidently could not originate.
This demonstration of Croll’s hypothesis in actual practice does not
present an altogether complete analogy to terrestrial conditions.
The surface temperature of Mars and the thermal properties of its
atmosphere may be widely different from ours, and the excessive rate
at which the polar caps diminish in the spring clearly indicates the
comparative thinness of the deposit. But in spite of these differences,
the remarkable conclusion stands unassailed—that the evaporative power
of the short hot summer of the south exceeds that of the long but
cooler summer of the north.
[Illustration: FIG. 9.
Directions of Glacial Movement in Southern Scandinavia.]
It was stated above that since the culmination of the last glacial
epoch more than 15,000 years have elapsed. Of the several methods which
have been employed to determine this period only one can be dealt with
here. The present genial climate has not greatly varied during the last
7000 years. As we trace back the record of temperature still farther a
gradual fall can be discerned, accompanied by a wide extension of the
ice. Fig. 9 depicts the directions of movement of the last continental
glacier of Scandinavia and its southern boundary across Jutland and
the Baltic provinces. As it retreated, it left the terminal moraine
known as the Baltic ridge, and its southern limit was gradually pushed
back till it extended to what is now the coast of Scania. From this
point its recession has been followed in great detail by De Geer, who
has made a careful study of the deposits which, extending from the
Baltic up to the Scandinavian ice-shed, mark the progress of its annual
retreat. Upon his observations he has founded a system of geological
chronology which is of the greatest importance in that it marks the
first effort towards absolute accuracy.
Each spring and summer, as the glacier thawed, a great deal of sand
and clay was set free and carried away in suspension by the numerous
streams which flowed from under the melting ice. The coarser material,
on reaching the sea, settled down almost at once, but the finest
particles of clay, able to remain in suspension much longer, were not
completely deposited. Then came the autumn and winter, and the freezing
of the streams. The sea received no further supply of sediment, and the
load of fine mud slowly settled on the sea-bottom to form a thin layer
of pure clay, sharply differentiated from the coarser bed below. The
following year the glacier retreated a few hundred feet to the north,
and the material then liberated was sorted out as before and again
deposited in two well-marked seasonal layers. As this process continued
year after year the area of deposit moved northwards with the ice, and
the annual layers of sediment thus became superimposed one upon the
other like wedge-shaped tiles on a roof. The width of each bed is
generally less than thirty miles, for even the finest mud cannot be
traced beyond that distance. This being the case, no vertical section
contains all the layers, and it is rarely that more than a hundred can
be counted in one place. The total thickness of the recessional deposit
seldom exceeds thirty feet.
De Geer successfully attempted the difficult task of counting the
annual bands of glacial clay deposited throughout the period of retreat
from the Scanian coast. Fortunately, any given set of layers can be
traced from one locality to another, and as each ribbon of sediment
dies out the higher beds are followed up in the same way until the
whole series has been examined from bottom to top. The late-glacial
beds number about 5000, and the time which has elapsed since the
ice border reached the eastern coast of Scania is therefore 12,000
years.[2] The time of recession from the Baltic ridge to the Scanian
coast remains to be estimated. Much of the record is hidden beneath the
waters of the Baltic, and in North Germany De Geer’s method has not yet
been applied. It appears, however, that the withdrawal of the ice was
not uniform. It began slowly and reluctantly, but towards the north
became more rapid. In the region of Stockholm the retreat was five
times as fast as in Scania. We may therefore assume that in the still
earlier stages the time taken was considerably more than that required
for a retreat over an equal distance in Scania. The latter would
have occupied rather more than 2500 years, and on this basis Sollas
provisionally accepts 5000 years as the period during which the ice
front was driven back to the south coast of Sweden. In this way the
whole interval which has passed since the culmination of the last
glacial episode is determined to be greater than 15,000 years, with
17,000 years as a probable value.
[Footnote 2: The post-glacial layers number about 7000.]
Amongst the Archean rocks of N.W. Finland, Prof. Sederholm has found
in the Bottnian schists and phyllites primary characters which are
strikingly similar to those of the banded glacial clays. Under the
microscope their textures are distinctly clastic, and each composite
stratum is sharply divided into two thin bands, the coarser one
originally of sand and marking the beginning of a new year of
deposition, the finer one originally of clay. Sederholm interprets
the phenomenon on a uniformitarian basis as indicating that, even in
those remote times, there was a marked difference in the seasons. By
measuring the thickness of many thousands of annual layers, he finds
their mean thickness to be about five inches. The total thickness of
the banded phyllites amounts to 10,000 feet and the time they appear to
represent is therefore only 24,000 years. This conclusion is probably
far from the truth, for, as Sederholm particularly emphasises, it is
difficult to know what is meant by the _thickness_ of a deposit,
so many arbitrary and misleading elements enter into its determination.
From De Geer’s work, it might have been said at first that a thickness
of thirty feet was deposited in a hundred years, but as the deposit
was traced over the country, it became equally evident, that with no
apparent thickening of the formation, deposition had gone on for 5000
years, and considering the growth northwards from the Baltic ridge,
perhaps for 10,000 years. The whole difficulty lies in determining
which parts of a formation are strictly contemporaneous. A measurement
of thickness is significant only in relation to the immediate area of
deposit and to the rate at which it moves landwards or seawards. The
_thickness_ of a formation and the _maximum thickness_ of the
layers formed in successive years may differ enormously, as De Geer’s
researches have made so evident. For the same reason the term _rate
of deposit_ is loose and misleading unless it is clear to what it
refers.
Another astronomical method of estimating time, though embracing much
longer periods, has been applied by Gilbert to certain formations
in Colorado. The basin of the Arkansas River is largely occupied by
Cretaceous sediments, a succession of which is given in the adjoining
diagram. At the four stages marked A, B, C and D the argillaceous
shales give place to a calcareous type, and in these there is a regular
alternation of thin layers of limestone and of calcareous shale. At A
the average thickness of a pair of beds is 1·5 feet, and the number
of repetitions is 15. At B the limestones are more massive, but the
parting shales are very thin. Here again the average thickness of two
adjacent beds is 1·5 feet. The limestones at C are less pure, and each,
with its associated layer of shale, amounts on the average to about 2·7
feet. At D the succession is similar.
[Illustration: FIG. 10.
Vertical Section of the Upper Cretaceous strata of Colorado. Below are
the Comanchian or Lower Cretaceous beds; above are the Transition beds
to the Eocene.]
To explain the remarkably regular alternation of conditions which
determined this uniform rhythm of sedimentation, purely terrestrial
causes appear to be insufficient. Upheaval and subsidence of the
earth’s crust, and the changing distribution of land and sea, are
characterised rather by their irregularity than by any rhythmic
sequence. While there is undoubtedly a rough periodicity in earth
movements, yet it is discernible only on the broadest scale and is out
of all proportion to the requirements of this case. Gilbert therefore
suggests an astronomical cause. Of the several cyclic changes to which
it is reasonable to appeal, the annual revolution of the earth, and
the variation of the eccentricity of its orbit, demand periods which
are in the first case too short (one year), and in the second too long
(91,000 years). As we have already seen, the relation of the seasons
to the position of perihelion repeats itself about every 21,000 years,
and this astronomical cycle seems better adapted to meet the case.
The climatic changes which accompany the precession of the equinoxes
might influence the character of sedimentation in many ways. With a
changing circulation of winds and currents, argillaceous material might
be transported and deposited at one time and calcareous at another. On
land, vegetation might predominate during part of the cycle, and the
surface waters would then dissolve more calcium carbonate than during
a period when vegetation became sparse. At the same time mechanical
erosion would be impeded in the first
instance, but would be more active in the second.
It is evident that the shale was deposited more rapidly than the
limestone, for when the principal deposit was calcareous the thickness
is 1·5 feet, whereas, in the case of the less calcareous beds, it
rises to 2·7 feet. In the normal shales a conservative estimate of the
equivalent thickness would be 4 feet. Adopting the astronomical time
unit of 21,000 years, the rate of deposit would then be of the order:
Limestone, 1 foot in 14,000 years.
Shale, 1 foot in 5,000 years.
On this basis the 3900 feet of shale in the Benton, Niobrara and Pierre
formations represent about 20 million years, for in this example
the term _thickness_ seems to have a definite meaning. If the
assumptions are correct, the duration of the whole Cretaceous period
must therefore be considerably greater than this.