Produced by Hugh Rance





THE BIRTH-TIME OF THE WORLD AND OTHER SCIENTIFIC ESSAYS

by

J. JOLY, M.A., Sc.D., F.R.S.,
PROFESSOR OF GEOLOGY AND MINERALOGY IN THE UNIVERSITY OF DUBLIN

E. P. DUTTON AND COMPANY
681 FIFTH AVENUE NEW YORK


Cover

Title page

CONTENTS PAGE

I. THE BIRTH-TIME OF THE WORLD - - - - - - - - - - - 1

II. DENUDATION  - - - - - - - - - - - - - - - - - - 30

III. THE ABUNDANCE OF LIFE  - - - - - - - - - - - - 60

IV. THE BRIGHT COLOURS OF ALPINE FLOWERS - - - - - 102

V. MOUNTAIN GENESIS  - - - - - - - - - - - - - - - 116

VI. ALPINE STRUCTURE - - - - - - - - - - - - - - - 146

VII. OTHER MINDS THAN OURS - - - - - - - - - - - - 162

VIII. THE LATENT IMAGE - - - - - - - - - - - - - - 202

IX. PLEOCHROIC HALOES  - - - - - - - - - - - - - - 214

X. THE USE OF RADIUM IN MEDICINE - - - - - - - - - 244

XI. SKATING  - - - - - - - - - - - - - - - - - - - 260

XII. A SPECULATION AS TO A PRE-MATERIAL UNIVERSE - 288

LIST OF ILLUSTRATIONS

PLATE I. LAKE OF LUCERNE, LOOKING WEST FROM BRUNNEN -
Frontispiece

PLATE II. "UPLIFTED FROM THE SEAS." CLIFFS OF THE TITLIS,
SWITZERLAND - to face p. 4

PLATE III. AN ALPINE TORRENT AT WORK--VAL D'HERENS, SWITZERLAND -
to face p. 31

PLATE IV. EARTH PILLARS--VAL D'HERENS - to face p. 34

PLATE V. "SCENES OF DESOLATION." THE WEISSHORN SEEN FROM BELLA
TOLA, SWITZERLAND - to face p. 40

PLATE VI. ALLUVIAL CONE--NICOLAI THAL, SWITZERLAND. MORAINE ON
ALETSCH GLACIER SWITZERLAND - to face p. 50

PLATE VII. IN THE REGION OF THE CROCI; DOLOMITES. THE ROTHWAND
SEEN FROM MONTE PIANO - to face p. 60

PLATE VIII. FIRS ASSAILING THE HEIGHTS OF THE MADERANER THAL,
SWITZERLAND - to face p. 73

PLATE IX. LIFE NEAR THE SNOW LINE; THE BOG-COTTON IN POSSESSION.
NEAR THE TSCHINGEL PASS, SWITZERLAND - to face p. 80

PLATE X. THE JOY OF LIFE. THE AMPEZZO THAL; DOLOMITES - to face
p. 93

PLATE XI. "PINES SOLEMNLY QUIET." DÜSSISTOCK; MADERANER THAL - to
face p. 100

PLATE XII. ALPINE FLOWERS IN THE VALLEYS - to face p. 105

PLATE XIII. ALPINE FLOWERS ON THE HEIGHTS - to face p. 106

PLATE XIV. MOUNTAIN SOLITUDES; VAL DE ZINAL. FROM LEFT TO RIGHT
ROTHHORN; BESSO; OBERGABELHORN; MATTERHORN; PIC DE ZINAL (THROUGH
CLOUD); DENT BLANCHE - to face p. 116

ix

PLATE XV. SECTOR OF THE EARTH RISE OF ISOGEOTHERMS INTO A DEPOSIT
EVOLVING RADIOACTIVE HEAT - to face p. 118

PLATE XVI. "THE MOUNTAINS COME AND GO." THE DENT BLANCHE SEEN
FROM THE SASSENEIRE - to face p. 133

PLATE XVII. DIAGRAMMATIC SECTIONS OF THE HIMALAYA - to face p.
140

PLATE XVIII. RESIDUES OF DENUDATION. THE MATTERHORN SEEN FROM THE
SUMMIT OF THE ZINAL ROTHHORN - to face p. 148

PLATE XIX. THE FOLDED ROCKS OF THE MATTERHORN, SEEN FROM NEAR
HÖHBALM. SKETCH MADE IN 1906 - to face p. 156

PLATE XX. SCHIAPARELLI'S MAP OF MARS OF 1882, AND ADDITIONS (IN
RED) OF 1892 - to face p. 166

PLATE XXI. GLOBE OF MARS SHOWING PATH OF IN-FALLING SATELLITE -
to face p. 188

PLATE XXII. CANALS MAPPED BY LOWELL COMPARED WITH CANALS FORMED
BY IN-FALLING SATELLITES - to face p. 192

PLATE XXIII. HALOES IN MICA; CO. CARLOW. HALO IN BIOTITE
CONTAINED IN GRANITE - to face p. 224

PLATE XXIV. RADIUM HALO, MUCH ENLARGED. THORIUM HALO AND RADIUM
HALO IN MICA - to face p. 228

PLATE XXV. HALO ROUND CAPILLARY IN GLASS TUBE. HALOES ROUND
TUBULAR PASSAGES IN MICA - to face p. 230

PLATE XXVI. ALETSCH GLACIER, SWITZERLAND - to face p. 282

PLATE XXVII. THE MIDDLE ALETSCH GLACIER JOINING THE GREAT ALETSCH
GLACIER. GLACIERS OF THE LAUTERBRUNNEN THAL - to face p. 285

PLATE XXVIII. PERCHED BLOCK ON THE ALETSCH GLACIER. GRANITE
ERRATIC NEAR ROUNDWOOD, CO. WICKLOW; NOW BROKEN UP AND REMOVED -
to face p. 286

And Fifteen Illustrations in the Text.

x

PREFACE

Tins volume contains twelve essays written at various times
during recent years. Many of them are studies contributed to
Scientific Reviews or delivered as popular lectures. Some are
expositions of views the scientific basis of which may be
regarded as established. Others--the greater number--may be
described as attempting the solution of problems which cannot be
approached by direct observation.

The essay on The Birth-time of the World is based on a lecture
delivered before the Royal Dublin Society. The subject has
attracted much attention within recent years. The age of the
Earth is, indeed, of primary importance in our conception of the
longevity of planetary systems. The essay deals with the
evidence, derived from the investigation of purely terrestrial
phenomena, as to the period which has elapsed since the ocean
condensed upon the Earth's surface. Dr. Decker's recent addition
to the subject appeared too late for inclusion in it. He finds
that the movements (termed isostatic) which geologists recognise
as taking place deep in the Earth's crust, indicate an age of the
same order of magnitude

xi

as that which is inferred from the statistics of denudative
history.[1]

The subject of _Denudation_ naturally arises from the first essay.
In thinking over the method of finding the age of the ocean by
the accumulation of sodium therein, I perceived so long ago as
1899, when my first paper was published, that this method
afforded a means of ascertaining the grand total of denudative
work effected on the Earth's surface since the beginning of
geological time; the resulting knowledge in no way involving any
assumption as to the duration of the period comprising the
denudative actions. This idea has been elaborated in various
publications since then, both by myself and by others.
"Denudation," while including a survey of the subject generally,
is mainly a popular account of this method and its results. It
closes with a reference to the fascinating problems presented by
the inner nature of sedimentation: a branch of science to which I
endeavoured to contribute some years ago.

_Mountain Genesis_ first brings in the subject of the geological
intervention of radioactivity. There can, I believe, be no doubt
as to the influence of transforming elements upon the
developments of the surface features of the Earth; and, if I am
right, this source of thermal energy is mainly responsible for
that local accumulation of wrinkling which we term mountain
chains. The

[1] Bull. Geol. Soc. America, vol. xxvi, March 1915.

xii

paper on _Alpine Structure_ is a reprint from "Radioactivity and
Geology," which for the sake of completeness is here included. It
is directed to the elucidation of a detail of mountain genesis: a
detail which enters into recent theories of Alpine development.
The weakness of the theory of the "horst" is manifest, however,
in many of its other applications; if not, indeed, in all.

The foregoing essays on the physical influences affecting the
surface features of the Earth are accompanied by one entitled _The
Abundance of Life._ This originated amidst the overwhelming
presentation of life which confronts us in the Swiss Alps. The
subject is sufficiently inspiring. Can no fundamental reason be
given for the urgency and aggressiveness of life? Vitality is an
ever-extending phenomenon. It is plain that the great principles
which have been enunciated in explanation of the origin of
species do not really touch the problem. In the essay--which is an
early one (1890)--the explanation of the whole great matter is
sought--and as I believe found--in the attitude of the organism
towards energy external to it; an attitude which results in its
evasion of the retardative and dissipatory effects which prevail
in lifeless dynamic systems of all kinds.

_Other Minds than Ours_? attempts a solution of the vexed question
of the origin of the Martian "canals." The essay is an abridgment
of two popular lectures on the subject. I had previously written
an account of my views which carried the enquiry as far as it was
in

xiii

my power to go. This paper appeared in the "Transactions of the
Royal Dublin Society, 1897." The theory put forward is a purely
physical one, and, if justified, the view that intelligent beings
exist in Mars derives no support from his visible surface
features; but is, in fact, confronted with fresh difficulties.

_Pleochroic Haloes_ is a popular exposition of an inconspicuous but
very beautiful phenomenon of the rocks. Minute darkened spheres--a
microscopic detail--appear everywhere in certain of the rock
minerals. What are they? The discoveries of recent radioactive
research--chiefly due to Rutherford--give the answer. The
measurements applied to the little objects render the explanation
beyond question. They turn out to be a quite extraordinary record
of radioactive energy; a record accumulated since remote
geological times, and assuring us, indirectly, of the stability
of the chemical elements in general since the beginning of the
world. This assurance is, without proof, often assumed in our
views on the geological history of the Globe.

Skating is a discourse, with a recent addition supporting the
original thesis. It is an illustration of a common experience--the
explanation of an unimportant action involving principles the
most influential considered as a part of Nature's resources.

The address on _The Latent Image_ deals with a subject which had
been approached by various writers before the time of my essay;
but, so far as I know, an explanation

xiv

based on the facts of photo-electricity had not been attempted.
Students of this subject will notice that the views expressed are
similar to those subsequently put forward by Lenard and Saeland
in explanation of phosphorescence. The whole matter is of more
practical importance than appears at first sight, for the
photoelectric nature of the effects involved in the radiative
treatment of many cruel diseases seems to be beyond doubt.

It was in connection with photo-electric science that I was led
to take an interest in the application of radioactivity in
medicine. The lecture on _The Use of Radium in Medicine_ deals with
this subject. Towards the conclusion of this essay reference will
be found to a practical outcome of such studies which, by
improving on the methods, and facilitating the application, of
radioactive treatment, has, in the hands of skilled medical men,
already resulted in the alleviation of suffering.

Leaving out much which might well appear in a prefatory notice, a
word should yet be added respecting the illustrations of scenery.
They are a small selection from a considerable number of
photographs taken during my summer wanderings in the Alps in
company with Henry H. Dixon. An exception is Plate X, which is by
the late Dr. Edward Stapleton. From what has been said above, it
will be gathered that these illustrations are fitly included
among pages which owe so much to Alpine inspiration. They
illustrate the

xv

subjects dealt with, and, it is to be hoped, they will in some
cases recall to the reader scenes which have in past times
influenced his thoughts in the same manner; scenes which in their
endless perspective seem to reduce to their proper insignificance
the lesser things of life.

My thanks are due to Mr. John Murray for kindly consenting to the
reissue of the essay on _The Birth-time of the World_ from the
pages of _Science Progress_; to Messrs. Constable & Co. for leave
to reprint _Pleochroic Haloes_ from _Bedrock_, and also to make some
extracts from _Radioactivity and Geology_; and to the Council of
the Royal Dublin Society for permission to republish certain
papers from the Proceedings of the Society.

_Iveagh Geological Laboratory, Trinity College, Dublin._

July, 1915.

xvi

THE BIRTH-TIME OF THE WORLD [1]

LONG ago Lucretius wrote: "For lack of power to solve the
question troubles the mind with doubts, whether there was ever a
birth-time of the world and whether likewise there is to be any
end." "And if" (he says in answer) "there was no birth-time of
earth and heaven and they have been from everlasting, why before
the Theban war and the destruction of Troy have not other poets
as well sung other themes? Whither have so many deeds of men so
often passed away, why live they nowhere embodied in lasting
records of fame? The truth methinks is that the sum has but a
recent date, and the nature of the world is new and has but
lately had its commencement."[2]

Thus spake Lucretius nearly 2,000 years ago. Since then we have
attained another standpoint and found very different limitations.
To Lucretius the world commenced with man, and the answer he
would give to his questions was in accord with his philosophy: he
would date the birth-time of the world from the time when

[1] A lecture delivered before the Royal Dublin Society, February
6th, 1914. _Science Progress_, vol. ix., p. 37

[2] _De Rerum Natura_, translated by H. A. J. Munro (Cambridge,
1886).

1

poets first sang upon the earth. Modern Science has along with
the theory that the Earth dated its beginning with the advent of
man, swept utterly away this beautiful imagining. We can, indeed,
find no beginning of the world. We trace back events and come to
barriers which close our vista--barriers which, for all we know,
may for ever close it. They stand like the gates of ivory and of
horn; portals from which only dreams proceed; and Science cannot
as yet say of this or that dream if it proceeds from the gate of
horn or from that of ivory.

In short, of the Earth's origin we have no certain knowledge; nor
can we assign any date to it. Possibly its formation was an event
so gradual that the beginning was spread over immense periods. We
can only trace the history back to certain events which may with
considerable certainty be regarded as ushering in our geological
era.

Notwithstanding our limitations, the date of the birth-time of
our geological era is the most important date in Science. For in
taking into our minds the spacious history of the universe, the
world's age must play the part of time-unit upon which all our
conceptions depend. If we date the geological history of the
Earth by thousands of years, as did our forerunners, we must
shape our ideas of planetary time accordingly; and the duration
of our solar system, and of the heavens, becomes comparable with
that of the dynasties of ancient nations. If by millions of
years, the sun and stars are proportionately venerable. If by
hundreds or thousands of millions of

2

years the human mind must consent to correspondingly vast epochs
for the duration of material changes. The geological age plays
the same part in our views of the duration of the universe as the
Earth's orbital radius does in our views of the immensity of
space. Lucretius knew nothing of our time-unit: his unit was the
life of a man. So also he knew nothing of our space-unit, and he
marvels that so small a body as the sun can shed so much, heat
and light upon the Earth.

A study of the rocks shows us that the world was not always what
it now is and long has been. We live in an epoch of denudation.
The rains and frosts disintegrate the hills; and the rivers roll
to the sea the finely divided particles into which they have been
resolved; as well as the salts which have been leached from them.
The sediments collect near the coasts of the continents; the
dissolved matter mingles with the general ocean. The geologist
has measured and mapped these deposits and traced them back into
the past, layer by layer. He finds them ever the same;
sandstones, slates, limestones, etc. But one thing is not the
same. _Life_ grows ever less diversified in character as the
sediments are traced downwards. Mammals and birds, reptiles,
amphibians, fishes, die out successively in the past; and barren
sediments ultimately succeed, leaving the first beginnings of
life undecipherable. Beneath these barren sediments lie rocks
collectively differing in character from those above: mainly
volcanic or poured out from fissures in

3

the early crust of the Earth. Sediments are scarce among these
materials.[1]

There can be little doubt that in this underlying floor of
igneous and metamorphic rocks we have reached those surface
materials of the earth which existed before the long epoch of
sedimentation began, and before the seas came into being. They
formed the floor of a vaporised ocean upon which the waters
condensed here and there from the hot and heavy atmosphere. Such
were the probable conditions which preceded the birth-time of the
ocean and of our era of life and its evolution.

It is from this epoch we date our geological age. Our next
purpose is to consider how long ago, measured in years, that
birth-time was.

That the geological age of the Earth is very great appears from
what we have already reviewed. The sediments of the past are many
miles in collective thickness: yet the feeble silt of the rivers
built them all from base to summit. They have been uplifted from
the seas and piled into mountains by movements so slow that
during all the time man has been upon the Earth but little change
would have been visible. The mountains have again been worn down
into the ocean by denudation and again younger mountains built
out of their redeposited materials. The contemplation of such
vast events

[1] For a description of these early rocks, see especially the
monograph of Van Hise and Leith on the pre-Cambrian Geology of
North America (Bulletin 360, U.S. Geol. Survey).

4

prepares our minds to accept many scores of millions of years or
hundreds of millions of years, if such be yielded by our
calculations.

THE AGE AS INFERRED FROM THE THICKNESS OF THE SEDIMENTS

The earliest recognised method of arriving at an estimate of the
Earth's geological age is based upon the measurement of the
collective sediments of geological periods. The method has
undergone much revision from time to time. Let us briefly review
it on the latest data.

The method consists in measuring the depths of all the successive
sedimentary deposits where these are best developed. We go all
over the explored world, recognising the successive deposits by
their fossils and by their stratigraphical relations, measuring
their thickness and selecting as part of the data required those
beds which we believe to most completely represent each
formation. The total of these measurements would tell us the age
of the Earth if their tale was indeed complete, and if we knew
the average rate at which they have been deposited. We soon,
however, find difficulties in arriving at the quantities we
require. Thus it is not easy to measure the real thickness of a
deposit. It may be folded back upon itself, and so we may measure
it twice over. We may exaggerate its thickness by measuring it
not quite straight across the bedding or by unwittingly including
volcanic materials. On the other hand, there

5

may be deposits which are inaccessible to us; or, again, an
entire absence of deposits; either because not laid down in the
areas we examine, or, if laid down, again washed into the sea.
These sources of error in part neutralise one another. Some make
our resulting age too long, others make it out too short. But we
do not know if a balance of error does not still remain. Here,
however, is a table of deposits which summarises a great deal of
our knowledge of the thickness of the stratigraphical
accumulations. It is due to Sollas.[1]

Feet.

Recent and Pleistocene  - -    4,000
Pliocene                 - -   13,000
Miocene                  - -   14,000
Oligocene                - -     2,000
Eocene                   - -  20,000
                              63,000

Upper Cretaceous        - -   24,000
Lower Cretaceous        - -   20,000
Jurassic                 - -    8,000
Trias                    - -   7,000
                              69,000

Permian                  - -    2,000
Carboniferous           - -    29,000
Devonian                 - -   22,000
                              63,000

Silurian                 - -   15,000
Ordovician              - -   17,000
Cambrian                 - -   6,000
                              58,000

Algonkian--Keeweenawan   - -   50,000
Algonkian--Animikian     - -   14,000
Algonkian--Huronian      - -   18,000
                              82,000

Archæan                   - -   ?

Total                     - - 335,000 feet.

[1] Address to the Geol. Soc. of London, 1509.

6

In the next place we require to know the average rate at which
these rocks were laid down. This is really the weakest link in
the chain. The most diverse results have been arrived at, which
space does not permit us to consider. The value required is most
difficult to determine, for it is different for the different
classes of material, and varies from river to river according to
the conditions of discharge to the sea. We may probably take it
as between two and six inches in a century.

Now the total depth of the sediments as we see is about 335,000
feet (or 64 miles), and if we take the rate of collecting as
three inches in a hundred years we get the time for all to
collect as 134 millions of years. If the rate be four inches, the
time is soo millions of years, which is the figure Geikie
favoured, although his result was based on somewhat different
data. Sollas most recently finds 80 millions of years.[1]

THE AGE AS INFERRED FROM THE MASS OF THE SEDIMENTS

In the above method we obtain our result by the measurement of
the linear dimensions of the sediments. These measurements, as we
have seen, are difficult to arrive at. We may, however, proceed
by measurements of the mass of the sediments, and then the method
becomes more definite. The new method is pursued as follows:

[1] Geikie, _Text Book of Geology_ (Macmillan, 1903), vol. i., p.
73, _et seq._ Sollas, _loc. cit._ Joly, _Radioactivity and Geology_
(Constable, 1909), and Phil. Mag., Sept. 1911.

7

The total mass of the sediments formed since denudation began may
be ascertained with comparative accuracy by a study of the
chemical composition of the waters of the ocean. The salts in the
ocean are undoubtedly derived from the rocks; increasing age by
age as the latter are degraded from their original character
under the action of the weather, etc., and converted to the
sedimentary form. By comparing the average chemical composition
of these two classes of material--the primary or igneous rocks and
the sedimentary--it is easy to arrive at a knowledge of how much
of this or that constituent was given to the ocean by each ton of
primary rock which was denuded to the sedimentary form. This,
however, will not assist us to our object unless the ocean has
retained the salts shed into it. It has not generally done so. In
the case of every substance but one the ocean continually gives
up again more or less of the salts supplied to it by the rivers.
The one exception is the element sodium. The great solubility of
its salts has protected it from abstraction, and it has gone on
collecting during geological time, practically in its entirety.
This gives us the clue to the denudative history of the
Earth.[1]

The process is now simple. We estimate by chemical examination of
igneous and sedimentary rocks the amount of sodium which has been
supplied to the ocean per ton of sediment produced by denudation.
We also calculate

[1] _Trans. R.D.S._, May, 1899.

8

the amount of sodium contained in the ocean. We divide the one
into the other (stated, of course, in the same units of mass),
and the quotient gives us the number of tons of sediment. The
most recent estimate of the sediments made in this manner affords
56 x 1016 tonnes.[1]

Now we are assured that all this sediment was transported by the
rivers to the sea during geological time. Thus it follows that,
if we can estimate the average annual rate of the river supply of
sediments to the ocean over the past, we can calculate the
required age. The land surface is at present largely covered with
the sedimentary rocks themselves. Sediment derived from these
rocks must be regarded as, for the most part, purely cyclical;
that is, circulating from the sea to the land and back again. It
does not go to increase the great body of detrital deposits. We
cannot, therefore, take the present river supply of sediment as
representing that obtaining over the long past. If the land was
all covered still with primary rocks we might do so. It has been
estimated that about 25 per cent. of the existing continental
area is covered with archæan and igneous rocks, the remainder
being sediments.[2] On this estimate we may find valuable

[1] Clarke, _A Preliminary Study of Chemical Denudation_
(Washington, 1910). My own estimate in 1899 (_loc. cit._) made as a
test of yet another method of finding the age, showed that the
sediments may be taken as sufficient to form a layer 1.1 mile
deep if spread uniformly over the continents; and would amount to
64 x 1018 tons.

[2] Van Tillo, _Comptes Rendues_ (Paris), vol. cxiv., 1892.

9

major and minor limits to the geological age. If we take 25 per
cent. only of the present river supply of sediment, we evidently
fix a major limit to the age, for it is certain that over the
past there must have been on the average a faster supply. If we
take the entire river supply, on similar reasoning we have what
is undoubtedly a minor limit to the age.

The river supply of detrital sediment has not been very
extensively investigated, although the quantities involved may be
found with comparative ease and accuracy. The following table
embodies the results obtained for some of the leading rivers.[1]

               Mean annual   Total annual   Ratio of
               discharge in  sediment in    sediment
               cubic feet    thousands      to water
               per second.   of tons.       by weight.

Potomac -        20,160          5,557        1 : 3.575
Mississippi -   610,000       406,250        1 : 1,500
Rio Grande -      1,700          3,830        1 : 291
Uruguay -       150,000         14,782       1 : 10,000
Rhone -          65,850         36,000       1 : 1,775
Po -             62,200         67,000       1 : 900
Danube -        315,200        108,000       1 : 2,880
Nile -          113,000         54,000       1 : 2,050
Irrawaddy -     475,000       291,430        1 : 1,610

Mean -          201,468        109,650       1 : 2,731

We see that the ratio of the weight of water to the

[1] Russell, _River Development_ (John Murray, 1888).

10

weight of transported sediment in six out of the nine rivers does
not vary widely. The mean is 2,730 to 1. But this is not the
required average. The water-discharge of each river has to be
taken into account. If we ascribe to the ratio given for each
river the weight proper to the amount of water it discharges, the
proportion of weight of water to weight of sediment, for the
whole quantity of water involved, comes out as 2,520 to 1.

Now if this proportion holds for all the rivers of the
world--which collectively discharge about 27 x 1012 tonnes of
water per annum--the river-born detritus is 1.07 x 1010 tonnes. To
this an addition of 11 per cent. has to be made for silt pushed
along the river-bed.[1] On these figures the minor limit to the
age comes out as 47 millions of years, and the major limit as 188
millions. We are here going on rather deficient estimates, the
rivers involved representing only some 6 per cent. of the total
river supply of water to the ocean. But the result is probably
not very far out.

We may arrive at a probable age lying between the major and minor
limits. If, first, we take the arithmetic mean of these limits,
we get 117 millions of years. Now this is almost certainly
excessive, for we here assume that the rate of covering of the
primary rocks by sediments was uniform. It would not be so,
however, for the rate of supply of original sediment must have
been continually diminishing

[1] According to observations made on the Mississippi (Russell,
_loc. cit._).

11

during geological time, and hence we may assume that the rate of
advance of the sediments on the primary rocks has also been
diminishing. Now we may probably take, as a fair assumption, that
the sediment-covered area was at any instant increasing at a rate
proportionate to the rate of supply of sediment; that is, to the
area of primary rocks then exposed. On this assumption the age is
found to be 87 millions of years.

THE AGE BY THE SODIUM OF THE OCEAN

I have next to lay before you a quite different method. I have
already touched upon the chemistry of the ocean, and on the
remarkable fact that the sodium contained in it has been
preserved, practically, in its entirety from the beginning of
geological time.

That the sea is one of the most beautiful and magnificent sights
in Nature, all admit. But, I think, to those who know its story
its beauty and magnificence are ten-fold increased. Its saltness
it due to no magic mill. It is the dissolved rocks of the Earth
which give it at once its brine, its strength, and its buoyancy.
The rivers which we say flow with "fresh" water to the sea
nevertheless contain those traces of salt which, collected over
the long ages, occasion the saltness of the ocean. Each gallon of
river water contributes to the final result; and this has been
going on since the beginning of our era. The mighty total of the
rivers is 6,500 cubic miles of water in the year!

12

There is little doubt that the primeval ocean was in the
condition of a fresh-water lake. It can be shown that a primitive
and more rapid solution of the original crust of the Earth by the
slowly cooling ocean would have given rise to relatively small
salinity. The fact is, the quantity of salts in the ocean is
enormous. We are only now concerned with the sodium; but if we
could extract all the rock-salt (the chloride of sodium) from the
ocean we should have enough to cover the entire dry land of the
Earth to a depth of 400 feet. It is this gigantic quantity which
is going to enter into our estimate of the Earth's age. The
calculated mass of sodium contained in this rock-salt is 14,130
million million tonnes.

If now we can determine the rate at which the rivers supply
sodium to the ocean, we can determine the age.[1] As the result
of many thousands of river analyses, the total amount of sodium
annually discharged to the ocean

[1] _Trans. R.D.S._, 1899. A paper by Edmund Halley, the
astronomer, in the _Philosophical Transactions of the Royal
Society_ for 1715, contains a suggestion for finding the age of
the world by the following procedure. He proposes to make
observations on the saltness of the seas and ocean at intervals
of one or more centuries, and from the increment of saltness
arrive at their age. The measurements, as a matter of fact, are
impracticable. The salinity would only gain (if all remained in
solution) one millionth part in Too years; and, of course, the
continuous rejection of salts by the ocean would invalidate the
method. The last objection also invalidates the calculation by T.
Mellard Reade (_Proc. Liverpool Geol. Soc._, 1876) of a minor limit
to the age by the calcium sulphate in the ocean. Both papers were
quite unknown to me when working out my method. Halley's paper
was, I think, only brought to light in 1908.

13

by all the rivers of the world is found to be probably not far
from 175 million tonnes.[1] Dividing this into the mass of
oceanic sodium we get the age as 80.7 millions of years. Certain
corrections have to be applied to this figure which result in
raising it to a little over 90 millions of years. Sollas, as the
result of a careful review of the data, gets the age as between
80 and 150 millions of years. My own result[2] was between 80 and
90 millions of years; but I subsequently found that upon certain
extreme assumptions a maximum age might be arrived at of 105
millions of years.[3] Clarke regards the 80.7 millions of years
as certainly a maximum in the light of certain calculations by
Becker.[4]

The order of magnitude of these results cannot be shaken unless
on the assumption that there is something entirely misleading in
the existing rate of solvent denudation. On the strength of the
results of another and

[1] F. W. Clarke, _A Preliminary Study of Chemical Denudation_
(Smithsonian Miscellaneous Collections, 1910).

[2] _Loc. cit._

[3] "The Circulation of Salt and Geological Time" (Geol. Mag.,
1901, p. 350).

[4] Becker (loc. cit.), assuming that the exposed igneous and
archæan rocks alone are responsible for the supply of sodium to
the ocean, arrives at 74 millions of years as the geological age.
This matter was discussed by me formerly (Trans. R.D.S., 1899,
pp. 54 _et seq._). The assumption made is, I believe, inadmissible.
It is not supported by river analyses, or by the chemical
character of residual soils from sedimentary rocks. There may be
some convergence in the rate of solvent denudation, but--as I
think on the evidence--in our time unimportant.

14

entirely different method of approaching the question of the
Earth's age (which shall be presently referred to), it has been
contended that it is too low. It is even asserted that it is from
nine to fourteen times too low. We have then to consider whether
such an enormous error can enter into the method. The
measurements involved cannot be seriously impugned. Corrections
for possible errors applied to the quantities entering into this
method have been considered by various writers. My own original
corrections have been generally confirmed. I think the only point
left open for discussion is the principle of uniformitarianism
involved in this method and in the methods previously discussed.

In order to appreciate the force of the evidence for uniformity
in the geological history of the Earth, it is, of course,
necessary to possess some acquaintance with geological science.
Some of the most eminent geologists, among whom Lyell and
Geikie[1] may be mentioned, have upheld the doctrine of
uniformity. It must here suffice to dwell upon a few points
having special reference to the matter under discussion.

The mere extent of the land surface does not, within limits,
affect the question of the rate of denudation. This arises from
the fact that the rain supply is quite insufficient to denude the
whole existing land surface. About 30 per cent. of it does not,
in fact, drain to the

[1] See especially Geikie's Address to Sect. C., Brit. Assoc.
Rep., 1399.

15

ocean. If the continents become invaded by a great transgression
of the ocean, this "rainless" area diminishes: and the denuded
area advances inwards without diminution. If the ocean recedes
from the present strand lines, the "rainless" area advances
outwards, but, the rain supply being sensibly constant, no change
in the river supply of salts is to be expected.

Age-long submergence of the entire land, or of any very large
proportion of what now exists, is negatived by the continuous
sequence of vast areas of sediment in every geologic age from the
earliest times. Now sediment-receiving areas always are but a
small fraction of those exposed areas whence the sediments are
supplied.[1] Hence in the continuous records of the sediments we
have assurance of the continuous exposure of the continents above
the ocean surface. The doctrine of the permanency of the
continents has in its main features been accepted by the most
eminent authorities. As to the actual amount of land which was
exposed during past times to denudative effects, no data exist to
show it was very different from what is now exposed. It has been
estimated that the average area of the North American continent
over geologic time was about eight-tenths of its existing
area.[2] Restorations of other continents, so far as they have
been attempted, would not

[1] On the strength of the Mississippi measurements about 1 to 18
(Magee, _Am. Jour. of Sc._, 1892, p. 188).

[2] Schuchert, _Bull. Geol. Soc. Am._, vol. xx., 1910.

16

suggest any more serious divergency one way or the other.

That climate in the oceans and upon the land was throughout much
as it is now, the continuous chain of teeming life and the
sensitive temperature limits of protoplasmic existence are
sufficient evidence.[1] The influence at once of climate and of
elevation of the land may be appraised at their true value by the
ascertained facts of solvent denudation, as the following table
shows.

                 Tonnes removed in      Mean elevation.
                 solution per square    Metres.
                 mile per annum.
North America -         79                  700
South America -         50                  650
Europe -                100                  300
Asia -                   84                  950
Africa -                 44                  650

In this table the estimated number of tonnes of matter in
solution, which for every square mile of area the rivers convey
to the ocean in one year, is given in the first column. These
results are compiled by Clarke from a very large number of
analyses of river waters. The second column of the table gives
the mean heights in metres above sea level of the several
continents, as cited by Arrhenius.[2]

Of all the denudation results given in the table, those relating
to North America and to Europe are far the

[1] See also Poulton, Address to Sect. D., Brit. Assoc. Rep.,
1896.

[2] _Lehybuch dev Kosmischen Physik_, vol. i., p. 347.

17

most reliable. Indeed these may be described as highly reliable,
being founded on some thousands of analyses, many of which have
been systematically pursued through every season of the year.
These show that Europe with a mean altitude of less than half
that of North America sheds to the ocean 25 per cent. more salts.
A result which is to be expected when the more important factors
of solvent denudation are given intelligent consideration and we
discriminate between conditions favouring solvent and detrital
denudation respectively: conditions in many cases
antagonistic.[1] Hence if it is true, as has been stated, that we
now live in a period of exceptionally high continental elevation,
we must infer that the average supply of salts to the ocean by
the rivers of the world is less than over the long past, and
that, therefore, our estimate of the age of the Earth as already
given is excessive.

There is, however, one condition which will operate to unduly
diminish our estimate of geologic time, and it is a condition
which may possibly obtain at the present time. If the land is, on
the whole, now sinking relatively to the ocean level, the
denudation area tends, as we have seen, to move inwards. It will
thus encroach upon regions which have not for long periods
drained to the ocean. On such areas there is an accumulation of
soluble salts which the deficient rivers have not been able to
carry to the ocean. Thus the salt content of certain of

[1] See the essay on Denudation.

18

the rivers draining to the ocean will be influenced not only by
present denudative effects, but also by the stored results of
past effects. Certain rivers appear to reveal this unduly
increased salt supply those which flow through comparatively arid
areas. However, the flowoff of such tributaries is relatively
small and the final effects on the great rivers apparently
unimportant--a result which might have been anticipated when the
extremely slow rate of the land movements is taken into account.

The difficulty of effecting any reconciliation of the methods
already described and that now to be given increases the interest
both of the former and the latter.

THE AGE BY RADIOACTIVE TRANSFORMATIONS

Rutherford suggested in 1905 that as helium was continually being
evolved at a uniform rate by radioactive substances (in the form
of the alpha rays) a determination of the age of minerals
containing the radioactive elements might be made by measurements
of the amount of the stored helium and of the radioactive
elements giving rise to it, The parent radioactive substances
are--according to present knowledge--uranium and thorium. An
estimate of the amounts of these elements present enables the
rate of production of the helium to be calculated. Rutherford
shortly afterwards found by this method an age of 240 millions of
years for a radioactive mineral of presumably remote age. Strutt,
who carried

19

his measurements to a wonderful degree of refinement, found the
following ages for mineral substances originating in different
geological ages:

Oligocene -               8.4 millions of years.
Eocene -                 31    millions of years.
Lower Carboniferous -  150   millions of years.
Archæan -               750    millions of years.

Periods of time much less than, and very inconsistent with, these
were also found. The lower results are, however, easily explained
if we assume that the helium--which is a gas under prevailing
conditions--escapes in many cases slowly from the mineral.

Another product of radioactive origin is lead. The suggestion
that this substance might be made available to determine the age
of the Earth also originated with Rutherford. We are at least
assured that this element cannot escape by gaseous diffusion from
the minerals. Boltwood's results on the amount of lead contained
in minerals of various ages, taken in conjunction with the amount
of uranium or parent substance present, afforded ages rising to
1,640 millions of years for archæan and 1,200 millions for
Algonkian time. Becker, applying the same method, obtained
results rising to quite incredible periods: from 1,671 to 11,470
millions of years. Becker maintained that original lead rendered
the determinations indefinite. The more recent results of Mr. A.
Holmes support the conclusion that "original" lead may be present
and may completely falsify results derived

20

from minerals of low radioactivity in which the derived lead
would be small in amount. By rejecting such results as appeared
to be of this character, he arrives at 370 millions of years as
the age of the Devonian.

I must now describe a very recent method of estimating the age of
the Earth. There are, in certain rock-forming minerals,
colour-changes set up by radioactive causes. The minute and
curious marks so produced are known as haloes; for they surround,
in ringlike forms, minute particles of included substances which
contain radioactive elements. It is now well known how these
haloes are formed. The particle in the centre of the halo
contains uranium or thorium, and, necessarily, along with the
parent substance, the various elements derived from it. In the
process of transformation giving rise to these several derived
substances, atoms of helium--the alpha rays--projected with great
velocity into the surrounding mineral, occasion the colour
changes referred to. These changes are limited to the distance to
which the alpha rays penetrate; hence the halo is a spherical
volume surrounding the central substance.[1]

The time required to form a halo could be found if on the one
hand we could ascertain the number of alpha rays ejected from the
nucleus of the halo in, say, one year, and, on the other, if we
determined by experiment just how many alpha rays were required
to produce the same

[1] _Phil. Mag._, March, 1907 and February, 1910; also _Bedrock_,
January, 1913. See _Pleochroic Haloes_ in this volume.

21

amount of colour alteration as we perceive to extend around the
nucleus.

The latter estimate is fairly easily and surely made. But to know
the number of rays leaving the central particle in unit time we
require to know the quantity of radioactive material in the
nucleus. This cannot be directly determined. We can only, from
known results obtained with larger specimens of just such a
mineral substance as composes the nucleus, guess at the amount of
uranium, or it may be thorium, which may be present.

This method has been applied to the uranium haloes of the mica of
County Carlow.[1] Results for the age of the halo of from 20 to
400 millions of years have been obtained. This mica was probably
formed in the granite of Leinster in late Silurian or in Devonian
times.

The higher results are probably the least in error, upon the data
involved; for the assumption made as to the amount of uranium in
the nuclei of the haloes was such as to render the higher results
the more reliable.

This method is, of course, a radioactive method, and similar to
the method by helium storage, save that it is free of the risk of
error by escape of the helium, the effects of which are, as it
were, registered at the moment of its production, so that its
subsequent escape is of no moment.

[1] Joly and Rutherford, _Phil. Mag._, April, 1913.

22

REVIEW OF THE RESULTS

We shall now briefly review the results on the geological age of
the Earth.

By methods based on the approximate uniformity of denudative
effects in the past, a period of the order of 100 millions of
years has been obtained as the duration of our geological age;
and consistently whether we accept for measurement the sediments
or the dissolved sodium. We can give reasons why these
measurements might afford too great an age, but we can find
absolutely no good reason why they should give one much too low.

By measuring radioactive products ages have been found which,
while they vary widely among themselves, yet claim to possess
accuracy in their superior limits, and exceed those derived from
denudation from nine to fourteen times.

In this difficulty let us consider the claims of the radioactive
method in any of its forms. In order to be trustworthy it must be
true; (1) that the rate of transformation now shown by the parent
substance has obtained throughout the entire past, and (2) that
there were no other radioactive substances, either now or
formerly existing, except uranium, which gave rise to lead. As
regards methods based on the production of helium, what we have
to say will largely apply to it also. If some unknown source of
these elements exists we, of course, on our assumption
overestimate the age.

23

As regards the first point: In ascribing a constant rate of
change to the parent substance--which Becker (loc. cit.) describes
as "a simple though tremendous extrapolation"--we reason upon
analogy with the constant rate of decay observed in the derived
radioactive bodies. If uranium and thorium are really primary
elements, however, the analogy relied on may be misleading; at
least, it is obviously incomplete. It is incomplete in a
particular which may be very important: the mode of origin of
these parent bodies--whatever it may have been--is different to
that of the secondary elements with which we compare them. A
convergence in their rate of transformation is not impossible, or
even improbable, so far as we known.

As regards the second point: It is assumed that uranium alone of
the elements in radioactive minerals is ultimately transformed to
lead by radioactive changes. We must consider this assumption.

Recent advances in the chemistry of the radioactive elements has
brought out evidence that all three lines of radioactive descent
known to us--_i.e._ those beginning with uranium, with thorium,
and with actinium--alike converge to lead.[1] There are
difficulties in the way of believing that all the lead-like atoms
so produced ("isotopes" of lead, as Soddy proposes to call them)
actually remain as stable lead in the minerals. For one

[1] See Soddy's _Chemistry of the Radioactive Elements_ (Longmans,
Green & Co.).

24

thing there is sometimes, along with very large amounts of
thorium, an almost entire absence of lead in thorianites and
thorites. And in some urano--thorites the lead may be noticed to
follow the uranium in approximate proportionality,
notwithstanding the presence of large amounts of thorium.[1] This
is in favour of the assumption that all the lead present is
derived from the uranium. The actinium is present in negligibly
small amounts.

On the other hand, there is evidence arising from the atomic
weight of lead which seems to involve some other parent than
uranium. Soddy, in the work referred to, points this out. The
atomic weight of radium is well known, and uranium in its descent
has to change to this element. The loss of mass between radium
and uranium-derived lead can be accurately estimated by the
number of alpha rays given off. From this we get the atomic
weight of uranium-derived lead as closely 206. Now the best
determinations of the atomic weight of normal lead assign to this
element an atomic weight of closely

[1] It seems very difficult at present to suggest an end product
for thorium, unless we assume that, by loss of electrons, thorium
E, or thorium-lead, reverts to a substance chemically identical
with thorium itself. Such a change--whether considered from the
point of view of the periodic law or of the radioactive theory
would involve many interesting consequences. It is, of course,
quite possible that the nature of the conditions attending the
deposition of the uranium ores, many of which are comparatively
recent, are responsible for the difficulties observed. The
thorium and uranium ores are, again, specially prone to
alteration.

25

207. By a somewhat similar calculation it is deduced that
thorium-derived lead would possess the atomic weight of 208. Thus
normal lead might be an admixture of uranium- and thorium-derived
lead. However, as we have seen, the view that thorium gives rise
to stable lead is beset with some difficulties.

If we are going upon reliable facts and figures, we must, then,
assume: (a) That some other element than uranium, and genetically
connected with it (probably as parent substance), gives rise, or
formerly gave rise, to lead of heavier atomic weight than normal
lead. It may be observed respecting this theory that there is
some support for the view that a parent substance both to uranium
and thorium has existed or possibly exists. The evidence is found
in the proportionality frequently observed between the amounts of
thorium and uranium in the primary rocks.[1] Or: (b) We may meet
the difficulties in a simpler way, which may be stated as
follows: If we assume that all stable lead is derived from
uranium, and at the same time recognise that lead is not
perfectly homogeneous in atomic weight, we must, of necessity,
ascribe to uranium a similar want of homogeneity; heavy atoms of
uranium giving rise to heavy

[1] Compare results for the thorium content of such rocks
(appearing in a paper by the author Cong. Int. _de Radiologie et
d'Electricité_, vol. i., 1910, p. 373), and those for the radium
content, as collected in _Phil. Mag._, October, 1912, p. 697.
Also A. L. Fletcher, _Phil. Mag._, July, 1910; January, 1911, and
June, 1911. J. H. J. Poole, _Phil. Mag._, April, 1915

26

atoms of lead and light atoms of uranium generating light atoms
of lead. This assumption seems to be involved in the figures
upon, which we are going. Still relying on these figures, we
find, however, that existing uranium cannot give rise to lead of
normal atomic weight. We can only conclude that the heavier atoms
of uranium have decayed more rapidly than the lighter ones. In
this connection it is of interest to note the complexity of
uranium as recently established by Geiger, although in this case
it is assumed that the shorter-lived isotope bears the relation
of offspring to the longer-lived and largely preponderating
constituent. However, there does not seem to be any direct proof
of this as yet.

From these considerations it would seem that unless the atomic
weight of lead in uraninites, etc., is 206, the former complexity
and more accelerated decay of uranium are indicated in the data
respecting the atomic weights of radium and lead[1]. As an
alternative view, we may assume, as in our first hypothesis, that
some elementally different but genetically connected substance,
decaying along branching lines of descent at a rate sufficient to
practically remove the whole of it during geological time,
formerly existed. Whichever hypothesis we adopt

[1] Later investigation has shown that the atomic weight of lead
in uranium-bearing ores is about 206.6 (see Richards and Lembert,
_Journ. of Am. Claem. Soc._, July, 1914). This result gives support
to the view expressed above.

27

we are confronted by probabilities which invalidate
time-measurements based on the lead and helium ratio in minerals.
We have, in short, grave reason to question the measure of
uniformitarianism postulated in finding the age by any of the
known radioactive methods.

That we have much to learn respecting our assumptions, whether we
pursue the geological or the radioactive methods of approaching
the age of our era, is, indeed, probable. Whatever the issue it
is certain that the reconciling facts will leave us with much
more light than we at present possess either as respects the
Earth's history or the history of the radioactive elements. With
this necessary admission we leave our study of the Birth-Time of
the World.

It has led us a long way from Lucretius. We do not ask if other
Iliads have perished; or if poets before Homer have vainly sung,
becoming a prey to all-consuming time. We move in a greater
history, the landmarks of which are not the birth and death of
kings and poets, but of species, genera, orders. And we set out
these organic events not according to the passing generations of
man, but over scores or hundreds of millions of years.

How much Lucretius has lost, and how much we have gained, is
bound up with the question of the intrinsic value of knowledge
and great ideas. Let us appraise knowledge as we would the
Homeric poems, as some-

28

thing which ennobles life and makes it happier. Well, then, we
are, as I think, in possession today of some of those lost Iliads
and Odysseys for which Lucretius looked in vain.[1]

[1] The duration in the past of Solar heat is necessarily bound
up with the geological age. There is no known means (outside
speculative science) of accounting for more than about 30 million
years of the existing solar temperature in the past. In this
direction the age seems certainly limited to 100 million years.
See a review of the question by Dr. Lindemann in Nature, April
5th, 1915.

29

DENUDATION

THE subject of denudation is at once one of the most interesting
and one of the most complicated with which the geologist has to
deal. While its great results are apparent even to the most
casual observer, the factors which have led to these results are
in many cases so indeterminate, and in some cases apparently so
variable in influence, that thoughtful writers have even claimed
precisely opposite effects as originating from, the same cause.
Indeed, it is almost impossible to deal with the subject without
entering upon controversial matters. In the following pages I
shall endeavour to keep to broad issues which are, at the present
day, either conceded by the greater number of authorities on the
subject, or are, from their strictly quantitative character, not
open to controversy.

It is evident, in the first place, that denudation--or the wearing
away of the land surfaces of the earth--is mainly a result of the
circulation of water from the ocean to the land, and back again
to the ocean. An action entirely conditioned by solar heat, and
without which it would completely cease and further change upon
the land come to an end.

To what actions, then, is so great a potency of the

30

circulating water to be traced? Broadly speaking, we may classify
them as mechanical and chemical. The first involves the
separation of rock masses into smaller fragments of all sizes,
down to the finest dust. The second involves the actual solution
in the water of the rock constituents, which may be regarded as
the final act of disintegration. The rivers bear the burden both
of the comminuted and the dissolved materials to the sea. The mud
and sand carried by their currents, or gradually pushed along
their beds, represent the former; the invisible dissolved matter,
only to be demonstrated to the eye by evaporation of the water or
by chemical precipitation, represents the latter.

The results of these actions, integrated over geological time,
are enormous. The entire bulk of the sedimentary rocks, such as
sandstones, slates, shales, conglomerates, limestones, etc., and
the salt content of the ocean, are due to the combined activity
of mechanical and solvent denudation. We shall, later on, make an
estimate of the magnitude of the quantities actually involved.

In the Swiss valleys we see torrents of muddy water hurrying
along, and if we follow them up, we trace them to glaciers high
among the mountains. From beneath the foot of the glacier, we
find, the torrent has birth. The first debris given to the river
is derived from the wearing of the rocky bed along which the
glacier moves. The river of ice bequeaths to the river of
water--of which it is the parent--the spoils which it has won from
the rocks

31

The work of mechanical disintegration is, however, not restricted
to the glacier's bed. It proceeds everywhere over the surface of
the rocks. It is aided by the most diverse actions. For instance,
the freezing and expansion of water in the chinks and cracks in
those alpine heights where between sunrise and sunset the heat of
summer reigns, and between sunset and sunrise the cold of winter.
Again, under these conditions the mere change of surface
temperature from night to day severely stresses the surface
layers of the rocks, and, on the same principles as we explain
the fracture of an unequally heated glass vessel, the rocks
cleave off in slabs which slip down the steeps of the mountain
and collect as screes in the valley. At lower levels the
expansive force of vegetable growth is not unimportant, as all
will admit who have seen the strong roots of the pines
penetrating the crannies of the rocks. Nor does the river which
flows in the bed of the valley act as a carrier only. Listening
carefully we may detect beneath the roar of the alpine torrent
the crunching and knocking of descending boulders. And in the
potholes scooped by its whirling waters we recognise the abrasive
action of the suspended sand upon the river bed.

A view from an Alpine summit reveals a scene of remarkable
desolation (Pl. V, p. 40). Screes lie piled against the steep
slopes. Cliffs stand shattered and ready to fall in ruins. And
here the forces at work readily reveal themselves. An occasional
wreath of white smoke among

32

the far-off peaks, followed by a rumbling reverberation, marks
the fall of an avalanche. Water everywhere trickles through the
shaly _débris_ scattered around. In the full sunshine the rocks are
almost too hot to bear touching. A few hours later the cold is
deadly, and all becomes a frozen silence. In such scenes of
desolation and destruction, detrital sediments are actively being
generated. As we descend into the valley we hear the deep voice
of the torrents which are continually hurrying the disintegrated
rocks to the ocean.

A remarkable demonstration of the activity of mechanical
denudation is shown by the phenomenon of "earth pillars." The
photograph (Pl. IV.) of the earth pillars of the Val d'Hérens
(Switzerland) shows the peculiar appearance these objects
present. They arise under conditions where large stones or
boulders are scattered in a deep deposit of clay, and where much
of the denudation is due to water scour. The large boulders not
only act as shelter against rain, but they bind and consolidate
by their mere weight the clay upon which they rest. Hence the
materials underlying the boulders become more resistant, and as
the surrounding clays are gradually washed away and carried to
the streams, these compacted parts persist, and, finally, stand
like walls or pillars above the general level. After a time the
great boulders fall off and the underlying clay becomes worn by
the rainwash to fantastic spikes and ridges. In the Val d'Hérens
the earth pillars are formed

33

of the deep moraine stuff which thickly overlies the slopes of
the valley. The wall of pillars runs across the axis of the
valley, down the slope of the hill, and crosses the road, so that
it has to be tunnelled to permit the passage of traffic. It is
not improbable that some additional influence--possibly the
presence of lime--has hardened the material forming the pillars,
and tended to their preservation.

Denudation has, however, other methods of work than purely
mechanical; methods more noiseless and gentle, but not less
effective, as the victories of peace ate no less than those of
war.

Over the immense tracts of the continents chemical work proceeds
relentlessly. The rock in general, more especially the primary
igneous rock, is not stable in presence of the atmosphere and of
water. Some of the minerals, such as certain silicates and
carbonates, dissolve relatively fast, others with extreme
slowness. In the process of solution chemical actions are
involved; oxidation in presence of the free oxygen of the
atmosphere; attack by the feeble acid arising from the solution
of carbon dioxide in water; or, again, by the activity of certain
acids--humous acids--which originate in the decomposition of
vegetable remains. These chemical agents may in some instances,
_e.g._ in the case of carbonates such as limestone or
dolomite--bring practically the whole rock into solution. In other
instances--_e.g._ granites, basalts, etc.--they may remove some of
the

34

constituent minerals completely or partially, such as felspar,
olivine, augite, and leave more resistant substances to be
ultimately washed down as fine sand or mud into the river.

It is often difficult or impossible to appraise the relative
efficiency of mechanical and chemical denudation in removing the
materials from a certain area. There can be, indeed, little doubt
that in mountainous regions the mechanical effects are largely
predominant. The silts of glacial rivers are little different
from freshly-powdered rock. The water which carries them but
little different from the pure rain or snow which falls from the
sky. There has not been time for the chemical or solvent actions
to take place. Now while gravitational forces favour sudden shock
and violent motions in the hills, the effect of these on solvent
and chemical denudation is but small. Nor is good drainage
favourable to chemical actions, for water is the primary factor
in every case. Water takes up and removes soluble combinations of
molecules, and penetrates beneath residual insoluble substances.
It carries the oxygen and acids downwards through the soils, and
finally conveys the results of its own work to the rivers and
streams. The lower mean temperature of the mountains as well as
the perfect drainage diminishes chemical activities.

Hence we conclude that the heights are not generally favourable
to the purely solvent and chemical actions. It is on the
lower-lying land that soils tend to accumulate,

35

and in these the chief solvent and the chief chemical denudation
of the Earth are effected.

The solvent and chemical effects which go on in the
finely-divided materials of the soils may be observed in the
laboratory. They proceed faster than would be anticipated. The
observation is made by passing a measured quantity of water
backwards and forwards for some months through a tube containing
a few grammes of powdered rock. Finally the water is analysed,
and in this manner the amount of dissolved matter it has taken up
is estimated. The rock powder is examined under the microscope in
order to determine the size of the grains, and so to calculate
the total surface exposed to the action of the water. We must be
careful in such experiments to permit free oxidation by the
atmosphere. Results obtained in this way of course take no
account of the chemical effects of organic acids such as exist in
the soils. The quantities obtained in the laboratory will,
therefore, be deficient as compared with the natural results.

In this manner it has been found that fresh basalt exposed to
continually moving water will lose about 0.20 gramme per square
metre of surface per year. The mineral orthoclase, which enters
largely into the constitution of many granites, was found to lose
under the same conditions 0.025 gramme. A glassy lava (obsidian)
rich in silica and in the chemical constituents of an average
granite, was more resistant still; losing but 0.013 gramme per
square metre per year. Hornblende, a mineral

36

abundant in many rocks, lost 0.075 gramme. The mean of the
results showed that 0.08 gramme was washed in a year from each
square metre. Such results give us some indication of the rate at
which the work of solution goes on in the finely divided
soils.[1]

It might be urged that, as the mechanical break up of rocks, and
the production in this way of large surfaces, must be at the
basis of solvent and chemical denudation, these latter activities
should be predominant in the mountains. The answer to this is
that the soils rarely owe their existence to mechanical actions.
The alluvium of the valleys constitutes only narrow margins to
the rivers; the finer _débris_ from the mountains is rapidly
brought into the ocean. The soils which cover the greater part of
continental areas have had a very different origin.

In any quarry where a section of the soil and of the underlying
rock is visible, we may study the mode of formation of soils. Our
observations are, we will suppose, pursued in a granite quarry.
We first note that the material of the soil nearest the surface
is intermixed with the roots of grasses, trees, or shrubs.
Examining a handful of this soil, we see glistening flakes of
mica which plainly are derived from the original granite. Washing
off the finer particles, we find the largest remaining grains are
composed of the all but indestructible quartz.

[1] Proc. Roy. Irish Acad., VIII., Ser. A, p. 21.

37

This also is from the granite. Some few of the grains are of
chalky-looking felspar; again a granitic mineral. What is the
finer silt we have washed off? It, too, is composed of mineral
particles to a great extent; rock dust stained with iron oxide
and intermixed with organic remains, both animal and vegetable.
But if we make a chemical analysis of the finer silt we find that
the composition is by no means that of the granite beneath. The
chemist is able to say, from a study of his results, that there
has been, in the first place, a large loss of material attending
the conversion of the granite to the soil. He finds a
concentration of certain of the more resistant substances of the
granite arising from the loss of the less resistant. Thus the
percentage amount of alumina is increased. The percentage of iron
is also increased. But silica and most other substances show a
diminished percentage. Notably lime has nearly disappeared. Soda
is much reduced; so is magnesia. Potash is not so completely
abstracted. Finally, owing to hydration, there is much more
combined water in the soil than in the rock. This is a typical
result for rocks of this kind.

Deeper in the soil we often observe a change of texture. It has
become finer, and at the same time the clay is paler in colour.
This subsoil represents the finer particles carried by rain from
above. The change of colour is due to the state of the iron which
is less oxidised low down in the soil. Beneath the subsoil the
soil grows

38

again coarser. Finally, we recognise in it fragments of granite
which ever grow larger as we descend, till the soil has become
replaced by the loose and shattered rock. Beneath this the only
sign of weathering apparent in the rock is the rusty hue imparted
by the oxidised iron which the percolating rain has leached from
iron-bearing minerals.

The soil we have examined has plainly been derived in situ from
the underlying rock. It represents the more insoluble residue
after water and acids have done their work. Each year there must
be a very slow sinking of the surface, but the ablation is
infinitesimal.

The depth of such a soil may be considerable. The total surface
exposed by the countless grains of which it is composed is
enormous. In a cubic foot of average soil the surface area of the
grains may be 50,000 square feet or more. Hence a soil only two
feet deep may expose 100,000 square feet for each square foot of
surface area.

It is true that soils formed in this manner by atmospheric and
organic actions take a very long time to grow. It must be
remembered, however, that the process is throughout attended by
the removal in solution: of chemically altered materials.

Considerations such as the foregoing must convince us that while
the accumulation of the detrital sediments around the continents
is largely the result of activities progressing on the steeper
slopes of the land, that is,

39

among the mountainous regions, the feeding of the salts to the
ocean arises from the slower work of meteorological and organic
agencies attacking the molecular constitution of the rocks;
processes which best proceed where the drainage is sluggish and
the quiescent conditions permit of the development of abundant
organic growth and decay.

Statistics of the solvent denudation of the continents support
this view. Within recent years a very large amount of work has
been expended on the chemical investigation of river waters of
America and of Europe. F. W. Clarke has, at the expense of much
labour, collected and compared these results. They are expressed
as so many tonnes removed in solution per square mile per annum.
For North America the result shows 79 tonnes so removed; for
Europe 100 tonnes. Now there is a notable difference between the
mean elevations of these two continents. North America has a mean
elevation of 700 metres over sea level, whereas the mean
elevation of Europe is but 300 metres. We see in these figures
that the more mountainous land supplies less dissolved matter to
the ocean than the land of lower elevation, as our study has led
us to expect.

We have now considered the source of the detrital sediments, as
well as of the dissolved matter which has given to the ocean, in
the course of geological time, its present gigantic load of
salts. It is true there are further solvent and chemical effects
exerted by the sea water

40

upon the sediments discharged into it; but we are justified in
concluding that, relatively to the similar actions taking place
in the soils, the solvent and chemical work of the ocean is
small. The fact is, the deposited detrital sediments around the
continents occupy an area small when contrasted with the vast
stretches of the land. The area of deposition is much less than
that of denudation; probably hardly as much as one twentieth.
And, again, the conditions of aeration and circulation which
largely promote chemical and solvent denudation in the soils are
relatively limited and ineffective in the detrital oceanic
deposits.

The summation of the amounts of dissolved and detrital materials
which denudation has brought into the ocean during the long
denudative history of the Earth, as we might anticipate, reveals
quantities of almost unrealisable greatness. The facts are among
the most impressive which geological science has brought to
light. Elsewhere in this volume they have been mentioned when
discussing the age of the Earth. In the present connection,
however, they are deserving of separate consideration.

The basis of our reasoning is that the ocean owes its saltness
mainly if not entirely to the denudative activities we have been
considering. We must establish this.

We may, in the first place, say that any other view at once
raises the greatest difficulties. The chemical composition of the
detrital sediments which are spread over

41

the continents and which build up the mountains, differs on the
average very considerably from that of the igneous rocks. We know
the former have been derived from the latter, and we know that
the difference in the composition of the two classes of materials
is due to the removal in solution of certain of the constituents
of the igneous rocks. But the ocean alone can have received this
dissolved matter. We know of no other place in which to look for
it. It is true that some part of this dissolved matter has been
again rejected by the ocean; thus the formation of limestone is
largely due to the abstraction of lime from sea water by organic
and other agencies. This, however, in no way relieves us of the
necessity of tracing to the ocean the substances dissolved from
the igneous rocks. It follows that we have here a very causa for
the saltness of the ocean. The view that the ocean "was salt from
the first" is without one known fact to support it, and leaves us
with the burden of the entire dissolved salts of geological time
to dispose of--Where and how?

The argument we have outlined above becomes convincingly strong
when examined more closely. For this purpose we first compare the
average chemical composition of the sedimentary and the igneous
rocks. The following table gives the percentages of the chief
chemical constituents: [1]

[1] F. W. Clarke: _A Preliminary Study of Chemical Denudation_,
p. 13

42

                          Igneous.        Sedimentary.
Silica (SiO2) -             59.99           58.51
Alumina (Al2O3) -           15.04           13.07
Ferric oxide (F2O3) -       2.59            3.40
Ferrous oxide (FeO) -       3.34            2.00
Magnesia (MgO) -            3.89            2.52
Lime (CaO) -                 4.81            5.42
Soda (Na2O) -                3.41            1.12
Potash (K2O) -               2.95            2.80
Water (H2O) -                1.92            4.28
Carbon dioxide (CO2) -       --             4.93
Minor constituents -        2.06            1.95
                          100.00          100.00

In the derivation of the sediments from the igneous rocks there
is a loss by solution of about 33 per cent; _i.e._ 100 tons of
igneous rock yields rather less than 70 tons of sedimentary rock.
This involves a concentration in the sediments of the more
insoluble constituents. To this rule the lime-content appears to
be an exception. It is not so in reality. Its high value in the
sediments is due to its restoration from the ocean to the land.
The magnesia and potash are, also, largely restored from the
ocean; the former in dolomites and magnesian limestones; the
latter in glauconite sands. The iron of the sediments shows
increased oxidation. The most notable difference in the two
analyses appears, however, in the soda percentages. This falls
from 3.41 in the igneous rock to 1.12 in the average sediment.
Indeed, this

43

deficiency of soda in sedimentary rocks is so characteristic of
secondary rocks that it may with some safety be applied to
discriminate between the two classes of substances in cases where
petrological distinctions of other kinds break down.

To what is this so marked deficiency of soda to be ascribed? It
is a result of the extreme solubility of the salts of sodium in
water. This has not only rendered its deposition by evaporation a
relatively rare and unimportant incident of geological history,
but also has protected it from abstraction from the ocean by
organic agencies. The element sodium has, in fact, accumulated in
the ocean during the whole of geological time.

We can use the facts associated with the accumulation of sodium
salts in the ocean as a means of obtaining additional support to
the view, that the processes of solvent denudation are
responsible for the saltness of the ocean. The new evidence may
be stated as follows: Estimates of the amounts of sedimentary
rock on the continents have repeatedly been made. It is true that
these estimates are no more than approximations. But they
undoubtedly _are_ approximations, and as such may legitimately be
used in our argument; more especially as final agreement tends to
check and to support the several estimates which enter into
them.

The most recent and probable estimates of the sediments on the
land assign an average thickness of one mile of

44

secondary rocks over the land area of the world. To this some
increase must be made to allow for similar materials concealed in
the ocean, principally around the continental margins. If we add
10 per cent. and assign a specific gravity of 2.5 we get as the
mass of the sediments 64 x 1016 tonnes. But as this is about 67
per cent. of the parent igneous rock--_i.e._ the average igneous
rock from which the sediments are derived--we conclude that the
primary denuded rock amounted to a mass of about 95 x 1016
tonnes.

Now from the mean chemical composition of the secondary rocks we
calculate that the mass of sediments as above determined contains
0.72 x1016 tonnes of the sodium oxide, Na2O. If to this amount we
add the quantity of sodium oxide which must have been given to
the ocean in order to account for the sodium salts contained
therein, we arrive at a total quantity of oxide of sodium which
must be that possessed by the primary rock before denudation
began its work upon it. The mass of the ocean being well
ascertained, we easily calculate that the sodium in the ocean
converted to sodium oxide amounts to 2.1 x 1016 tonnes. Hence
between the estimated sediments and the waters of the ocean we
can account for 2.82 x 1016 tonnes of soda. When now we put this
quantity back into the estimated mass of primary rock we find
that it assigns to the primary rock a soda percentage of 3.0. On
the average analysis given above this should be 3.41 per cent.
The agreement,

45

all things considered, more especially the uncertainty in the
estimate of the sediments, is plainly in support of the view that
oceanic salts are derived from the rocks; if, indeed, it does not
render it a certainty.

A leading and fundamental inference in the denudative history of
the Earth thus finds support: indeed, we may say, verification.
In the light of this fact the whole work of denudation stands
revealed. That the ocean began its history as a vast fresh-water
envelope of the Globe is a view which accords with the evidence
for the primitive high temperature of the Earth. Geological
history opened with the condensation of an atmosphere of immense
extent, which, after long fluctuations between the states of
steam and water, finally settled upon the surface, almost free of
matter in solution: an ocean of distilled water. The epoch of
denudation then began. It will, probably, continue till the
waters, undergoing further loss of thermal energy, suffer yet
another change of state, when their circulation will cease and
their attack upon the rocks come to an end.

From what has been reviewed above it is evident that the sodium
in the ocean is an index of the total activity of denudation
integrated over geological time. From this the broad facts of the
results of denudation admit of determination with considerable
accuracy. We can estimate the amount of rock which has been
degraded by solvent and chemical actions, and the amount of
sediments which has been derived from it. We are,

46

thus, able to amend our estimate of the sediments which, as
determined by direct observation, served to support the basis of
our argument.

We now go straight to the ocean for the amount of sodium of
denudative origin. There may, indeed, have been some primitive
sodium dissolved by a more rapid denudation while the Earth's
surface was still falling in temperature. It can be shown,
however, that this amount was relatively small. Neglecting it we
may say with safety that the quantity of sodium carried into the
ocean by the rivers must be between 14,000 and 15,000 million
million tonnes: _i.e._ 14,500 x 1012 tonnes, say.

Keeping the figures to round numbers we find that this amount of
sodium involves the denudation of about 80 x 1016 tonnes of
average igneous rock to 53 x 1016 tonnes of average sediment.
From these vast quantities we know that the parent rock denuded
during geological time amounted to some 300 million cubic
kilometres or about seventy million cubic miles. The sediments
derived therefrom possessed a bulk of 220 million cubic
kilometres or fifty million cubic miles. The area of the land
surface of the Globe is 144 million square kilometres. The parent
rock would have covered this to a uniform depth of rather more
than two kilometres, and the derived sediment to more than 1.5
kilometres, or about one mile deep.

The slow accomplishment of results so vast conveys some idea of
the great duration of geological time.

47

The foregoing method of investigating the statistics of solvent
denudation is capable of affording information not only as to the
amount of sediments upon the land, but also as to the quantity
which is spread over the floor of the ocean.

We see this when we follow the fate of the 33 per cent. of
dissolved salts which has been leached from the parent igneous
rock, and the mass of which we calculate from the ascertained
mass of the latter, to be 27 x 1016 tonnes. This quantity was at
one time or another all in the ocean. But, as we saw above, a
certain part of it has been again abstracted from solution,
chiefly by organic agencies. Now the abstracted solids have not
been altogether retained beneath the ocean. Movements of the land
during geological time have resulted in some portion being
uplifted along with other sediments. These substances constitute,
mainly, the limestones.

We see, then, that the 27 x 1016 tonnes of substances leached
from the parent igneous rocks have had a threefold destination.
One part is still in solution; a second part has been
precipitated to the bottom of the ocean; a third part exists on
the land in the form of calcareous rocks.

Observation on the land sediments shows that the calcareous rocks
amount to about 5 per cent. of the whole. From this we find that
3 x 1016 tonnes, approximately, of such rocks have been taken
from the ocean. This accounts for one of the three classes of
material

48

into which the original dissolved matter has been divided.
Another of the three quantities is easily estimated: the amount
of matter still in solution in the ocean. The volume of the ocean
is 1,414 million cubic kilometres and its mass is 145 x 1016
tonnes. The dissolved salts in it constitute 3.4 per cent. of its
mass; or, rather more than 5 x 1016 tonnes. The limestones on the
land and the salts in the sea water together make up about 8 x
1016 tonnes. If we, now, deduct this from the total of 27 x 1016
tonnes, we find that about 19 x 1016 tonnes must exist as
precipitated matter on the floor of the ocean.

The area of the ocean is 367 x 1012 square metres, so that if the
precipitated sediment possesses an average specific gravity of
2.5, it would cover the entire floor to a uniform depth of 218
metres; that is 715 feet. This assumes that there was uniform
deposition of the abstracted matter over the floor of the ocean.
Of course, this assumption is not justifiable. It is certain that
the rate of deposition on the floor of the sea has varied
enormously with various conditions--principally with the depth.
Again, it must be remembered that this estimate takes no account
of solid materials otherwise brought into the oceanic deposits;
_e.g._, by wind-transported dust from the land or volcanic
ejectamenta in the ocean depths. It is not probable, however,
that any considerable addition to the estimated mean depth of
deposit from such sources would be allowable.

49

The greatness of the quantities involved in these determinations
is almost awe inspiring. Take the case of the dissolved salts in
the ocean. They are but a fraction, as we have seen, of the total
results of solvent denudation and represent the integration of
the minute traces contributed by the river water. Yet the common
salt (chloride of sodium) alone, contained in the ocean, would,
if abstracted and spread over the dry land as a layer of rock
salt having a specific gravity of 2.2, cover the whole to a depth
of 107 metres or 354 feet. The total salts in solution in the
ocean similarly spread over the land would increase the depth of
the layer to 460 feet. After considering what this means we have
to remember that this amount of matter now in solution in the
seas is, in point of fact, less than a fifth part of the total
dissolved from the rocks during geological time.

The transport by denudation of detrital and dissolved matter from
the land to the ocean has had a most important influence on the
events of geological history. The existing surface features of
the earth must have been largely conditioned by the dynamical
effects arising therefrom. In dealing with the subject of
mountain genesis we will, elsewhere, see that all the great
mountain ranges have originated in the accumulation of the
detrital sediments near the shore in areas which, in consequence
of the load, gradually became depressed and developed into
synclines of many thousands of feet in depth. The most impressive
surface features of the Globe originated

50

in this manner. We will see too that these events were of a
rhythmic character; the upraising of the mountains involving
intensified mechanical denudation over the elevated area and in
this way an accelerated transport of detritus to the sea; the
formation of fresh deposits; renewed synclinal sinking of the sea
floor, and, finally, the upheaval of a younger mountain range.
This extraordinary sequence of events has been determined by the
events of detrital denudation acting along with certain general
conditions which have all along involved the growth of
compressive stresses in the surface crust of the Earth.

The effects of purely solvent denudation are less easily traced,
but, very probably, they have been of not less importance. I
refer here to the transport from the land to the sea of matter in
solution.

Solvent denudation, as observed above, takes place mainly in the
soils and in this way over the more level continental areas. It
has resulted in the removal from the land and transfer to the
ocean of an amount of matter which represents a uniform layer of
one half a kilometre; that is of more than 1,600 feet of rock.
The continents have, during geological time, been lightened to
this extent. On the other hand all this matter has for the
greater part escaped the geosynclines and become uniformly
diffused throughout the ocean or precipitated over its floor
principally on the continental slopes before the great depths are
reached. Of this material the ocean

51

waters contain in solution an amount sufficient to increase their
specific gravity by 2.7 per cent.

Taking the last point first, it is interesting to note the
effects upon the bulk of the ocean which has resulted from the
matter dissolved in it. From the known density of average sea
water we find that 100 ccs. of it weigh just 102.7 grammes. Of
this 3.5 per cent. by weight are solids in solution. That is to
say, 3.594 grammes. Hence the weight of water present is 99.1
grammes, or a volume of 99.1 ccs. From this we see that the salts
present have increased the volume by 0.9 ccs. or 0.9 per cent.

The average depth of the ocean is 2,000 fathoms or 3,700 metres.
The increase of depth due to salts dissolved in the ocean has
been, therefore, 108 feet or 33.24 metres. This result assumes
that there has been no increased elastic compression due to the
increased pressure, and no change of compressional elastic
properties. We may be sure that the rise on the shore line of the
land has not been less than 100 feet.

We see then that as the result of solvent denudation we have to
do with a heavier and a deeper ocean, expanded in volume by
nearly one per cent. and the floor of which has become raised, on
an average, about 700 feet by precipitated sediment.

One of the first conceptions, which the student of geology has to
dismiss from his mind, is that of the immobility or rigidity of
the Earth's crust. The lane, we live on sways even to the gentle
rise and fail of ocean tides

52

around the coasts. It suffers its own tidal oscillations due to
the moon's attractions. Large tracts of semi-liquid matter
underlie it. There is every evidence that the raised features of
the Globe are sustained by such pressures acting over other and
adjacent areas as serve to keep them in equilibrium against the
force of gravity. This state of equilibrium, which was first
recognised by Pratt, as part of the dynamics of the Earth's
crust, has been named isostasy. The state of the crust is that of
"mobile equilibrium."

The transfer of matter from the exposed land surfaces to the
sub-oceanic slopes of the continents and the increase in the
density of the ocean, must all along have been attended by
isostatic readjustment. We cannot take any other view. On the one
hand the land was being lightened; on the other the sea was
increasing in mass and depth and the flanks of the continents
were being loaded with the matter removed from the land and borne
in solution to the ocean. How important the resulting movements
must have been may be gathered from the fact that the existing
land of the Globe stands at a mean elevation of no more than
2,000 feet above sea level. We have seen that solvent denudation
removed over 1,600 feet of rock. But we have no evidence that on
the whole the elevation of land in the past was ever very
different from what it now is.

We have, then, presented to our view the remarkable fact that
throughout the past, and acting with extreme

53

slowness, the land has steadily been melted down into the sea and
as steadily been upraised from the waters. It is possible that
the increased bulk of the ocean has led to a certain diminution
of the exposed land area. The point is a difficult one. One thing
we may without much risk assume. The sub-aereal current of
dissolved matter from the land to the ocean was accompanied by a
sub-crustal flux from the ocean areas to the land areas; the
heated viscous materials creeping from depths far beneath the
ocean floor to depths beneath the roots of the mountains which
arose around the oceans. Such movements took ages for their
accomplishment. Indeed, they have been, probably, continuous all
along and are still proceeding. A low degree of viscosity will
suffice to permit of movements so slow. Superimposed upon these
movements the rhythmic alternations of depression and elevation
of the geosynclines probably resulted in releasing the crust from
local accumulation of strains arising in the more rigid surface
materials. The whole sequence of movements presents an
extraordinary picture of pseudo-vitality--reminding us of the
circulatory and respiratory systems of a vast organism.

All great results in our universe are founded in motions and
forces the most minute. In contemplating the Cause or the Effect
we stand equally impressed with the spectacle presented to us. We
shall now turn from the great effects of denudation upon the
history and evolution of a world and consider for a moment
activities

54

so minute in detail that their operations will probably for ever
elude our bodily senses, but which nevertheless have necessarily
affected and modified the great results we have been
considering.

The ocean a little way from the land is generally so free from
suspended sediments that it has a blackness as of ink. This
blackness is due to its absolute freedom from particles
reflecting the sun's light. The beautiful blue of the Swiss and
Italian lakes is due to the presence of very fine particles
carried into them by the rivers; the finest flour of the
glaciers, which remain almost indefinitely suspended in the
water. But in the ocean it is only in those places where rapid
currents running over shallows stir continually the sediments or
where the fresh water of a great river is carried far from the
land, that the presence of silt is to be observed. The beautiful
phenomenon of the coal-black sea is familiar to every yachtsman
who has sailed to the west of our Islands.[1]

There is, in fact, a very remarkable difference in the manner of
settlement of fine sediments in salt and in fresh water. We are
here brought into contact with one of those subtle yet
influential natural actions the explanation of which involves
scientific advance along many apparently unconnected lines of
investigation.

[1] See Tyndall's Voyage to Algeria in _Fragments of Science._ The
cause of the blue colour of the lakes has been discussed by
various observers, not always with agreement.

55

It is easy to observe in the laboratory the fact of the different
behaviour of salt and fresh water towards finely divided
substances. The nature of the insoluble substance is not
important.

We place, in a good light, two glass vessels of equal dimensions;
the one filled with sea water, the other with fresh water. Into
each we stir the same weight of very finely powdered slate: just
so much as will produce a cloudiness. In a few hours we find the
sea water limpid. The fresh water is still cloudy, however; and,
indeed, may be hardly different in appearance from what it was at
starting. In itself this is a most extraordinary experiment. We
would have anticipated quite the opposite result owing to the
greater density of the sea water.

But a still more interesting experiment remains to be carried
out. In the sea water we have many different salts in solution.
Let us see if these salts are equally responsible for the result
we have obtained. For this purpose we measure out quantities of
sodium chloride and magnesium chloride in the proportion in which
they exist in sea water: that is about as seven to one. We add
such an equal amount of water to each as represents the dilution
of these salts in sea water. Then finally we stir a little of the
finely powdered slate into each. It will be found that the
magnesium chloride, although so much more dilute than the sodium
chloride, is considerably more active in clearing out the
suspension. We may now try such marine salts as magnesium
sulphate,

56

or calcium sulphate against sodium chloride; keeping the marine
proportions. Again we find that the magnesium and calcium salts
are the most effective, although so much more dilute than the
sodium salt.

There is no visible clue to the explanation of these results. But
we must conclude as most probable that some action is at work in
the sea water and in the salt solutions which clumps or
flocculates the sediment. For only by the gathering of the
particles together in little aggregates can we explain their
rapid fall to the bottom. It is not a question of viscosity
(_i.e._ of resistance to the motion of the particles), for the
salt solutions are rather more viscous than the fresh water.
Still more remarkable is the fact that every dissolved substance
will not bring about the result. Thus if we dissolve sugar in
water we find that, if anything, the silt settles more slowly in
the sugar solution than in fresh water.

Now there is one effect produced by the solution of such salts as
we have dealt with which is not produced by such bodies as sugar.
The water is rendered a conductor of electricity. Long ago
Faraday explained this as due to the presence of free atoms of
the dissolved salt in the solution, carrying electric charges. We
now speak of the salt as "ionised." That is it is partly split up
into ions or free electrified atoms of chlorine, sodium,
magnesium, etc., according to the particular salt in solution.
This fact leads us to think that these electrified

57

atoms moving about in the solution may be the cause of the
clumping or flocculation. Such electrified atoms are absent from
the sugar solution: sugar does not become "ionised" when it is
dissolved.

The suspicion that the free electrified atoms play a part in the
phenomenon is strengthened when we recall the remarkable
difference in the action of sodium chloride and magnesium
chloride. In each of the solutions of these substances there are
free chlorine atoms each of which carries a single charge of
negative electricity. As these atoms are alike in both solutions
the different behaviour of the solutions cannot be due to the
chlorine. But the metallic atom is very different in the two
cases. The ionised sodium atom is known to be _monad_ or carries
but _one_ positive charge; whereas the magnesium atom is _diad_ and
carries _two_ positive charges. If, then, we assume that the
metallic, positively electrified atom is in each case
responsible, we have something to go on. It may be now stated
that it has been found by experiment and supported by theory that
the clumping power of an ion rises very rapidly with its valency;
that is with the number of unit charges associated with it. Thus
diads such as magnesium, calcium, barium, etc., are very much
more efficient than monads such as sodium, potassium, etc., and
again, triads such as aluminium are, similarly, very much more
powerful than diad atoms. Here, in short, we have arrived at the
active cause of the phenomenon. Its inner mechanism

58

is, however, harder to fathom. A plausible explanation can be
offered, but a study of it would take us too far. Sufficient has
been said to show the very subtile nature of the forces at work.

We have here an effect due to the sea salts derived by denudation
from the land which has been slowly augmenting during geological
time. It is certain that the ocean was practically fresh water in
remote ages. During those times the silt from the great rivers
would have been carried very far from the land. A Mississippi of
those ages would have sent its finer suspensions far abroad on a
contemporary Gulf stream: not improbably right across the
Atlantic. The earlier sediments of argillaceous type were not
collected in the geosynclines and the genesis of the mountains
was delayed proportionately. But it was, probably, not for very
long that such conditions prevailed. For the accumulation of
calcium salts must have been rapid, and although the great
salinity due to sodium salts was of slow growth the salts of the
diad element calcium must have soon introduced the cooperation of
the ion in the work of building the mountain.

59

THE ABUNDANCE OF LIFE [1]

WE had reached the Pass of Tre Croci[2]and from a point a little
below the summit, looked eastward over the glorious Val Buona.
The pines which clothed the floor and lower slopes of the valley,
extended their multitudes into the furthest distance, among the
many recesses of the mountains, and into the confluent Val di
Misurina. In the sunshine the Alpine butterflies flitted from
stone to stone. The ground at our feet and everywhere throughout
the forests teamed with the countless millions of the small black
ants.

It was a magnificent display of vitality; of the aggressiveness
of vitality, assailing the barren heights of the limestone,
wringing a subsistence from dead things. And the question
suggested itself with new force: why the abundance of life and
its unending activity?

In trying to answer this question, the present sketch
originated.

I propose to refer for an answer to dynamic considerations. It is
apparent that natural selection can only be concerned in a
secondary way. Natural selection defines

[1] Proc. Roy. Dublin Soc., vol. vii., 1890.

[2] In the Dolomites of Southeast Tyrol; during the summer of
1890. Much of what follows was evolved in discussion with my
fellow-traveller, Henry H. Dixon. Much of it is his.

60

a certain course of development for the organism; but very
evidently some property of inherent progressiveness in the
organism must be involved. The mineral is not affected by natural
selection to enter on a course of continual variation and
multiplication. The dynamic relations of the organism with the
environment are evidently very different from those of inanimate
nature.

GENERAL DYNAMIC CONDITIONS ATTENDING INANIMATE ACTIONS

It is necessary, in the first place, to refer briefly to the
phenomena attending the transfer of energy within and into
inanimate material systems. It is not assumed here that these
phenomena are restricted in their sphere of action to inanimate
nature. It is, in fact, very certain that they are not; but while
they confer on dead nature its own dynamic tendencies, it will
appear that their effects are by various means evaded in living
nature. We, therefore, treat of them as characteristic of
inanimate actions. We accept as fundamental to all the
considerations which follow the truth of the principle of the
Conservation of Energy.[1]

[1] "The principle of the Conservation of Energy has acquired so
much scientific weight during the last twenty years that no
physiologist would feel any confidence in an experiment which
showed a considerable difference between the work done by the
animal and the balance of the account of Energy received and
spent."--Clerk Maxwell, _Nature_, vol. xix., p. 142. See also
Helmholtz _On the Conservation of Force._

61

Whatever speculations may be made as to the course of events very
distant from us in space, it appears certain that dissipation of
energy is at present actively progressing throughout our sphere
of observation in inanimate nature. It follows, in fact, from the
second law of thermodynamics, that whenever work is derived from
heat, a certain quantity of heat falls in potential without doing
work or, in short, is dissipated. On the other hand, work may be
entirely converted into heat. The result is the heat-tendency of
the universe. Heat, being an undirected form of energy, seeks, as
it were, its own level, so that the result of this heat-tendency
is continual approach to uniformity of potential.

The heat-tendency of the universe is also revealed in the
far-reaching "law of maximum work," which defines that chemical
change, accomplished without the intervention of external energy,
tends to the production of the body, or system of bodies, which
disengage the greatest quantity of heat.[1] And, again, vast
numbers of actions going on throughout nature are attended by
dissipatory thermal effects, as those arising from the motions of
proximate molecules (friction, viscosity), and from the fall of
electrical potential.

Thus, on all sides, the energy which was once most probably
existent in the form of gravitational potential, is being
dissipated into unavailable forms. We must

[1] Berthelot, _Essai de Mécanique Chimique._

62

recognize dissipation as an inevitable attendant on inanimate
transfer of energy.

But when we come to consider inanimate actions in relation to
time, or time-rate of change, we find a new feature in the
phenomena attending transfer of energy; a feature which is really
involved in general statements as to the laws of physical
interactions.[1] It is seen, that the attitude of inanimate
material systems is very generally, if not in all cases,
retardative of change--opposing it by effects generated by the
primary action, which may be called "secondary" for convenience.
Further, it will be seen that these secondary effects are those
concerned in bringing about the inevitable dissipation.

As example, let us endeavour to transfer gravitational potential
energy contained in a mass raised above the surface of the Earth
into an elastic body, which we can put into compression by
resting the weight upon it. In this way work is done against
elastic force and stored as elastic potential energy. We may deal
with a metal spring, or with a mass of gas contained in a
cylinder fitted with a piston upon which the weight may be
placed. In either case we find the effect of compression is to
raise the temperature of the substance, thus causing its

[1] Helmholtz, _Ice and Glaciers._ Atkinson's collection of his
Popular Lectures. First Series, p.120. Quoted by Tate, _Heat_,
p. 311.

63

expansion or increased resistance to the descent of the weight.
And this resistance continues, with diminishing intensity, till
all the heat generated is dissipated into the surrounding medium.
The secondary effect thus delays the final transfer of energy.

Again, if we suppose the gas in the cylinder replaced by a vapour
in a state of saturation, the effect of increased pressure, as of
a weight placed upon the piston, is to reduce the vapour to a
liquid, thereby bringing about a great diminution of volume and
proportional loss of gravitational potential by the weight. But
this change will by no means be brought about instantaneously.
When a little of the vapour is condensed, this portion parts with
latent heat of vaporisation, increasing the tension of the
remainder, or raising its point of saturation, so that before the
weight descends any further, this heat has to escape from the
cylinder.

Many more such cases might be cited. The heating of india-rubber
when expanded, its cooling when compressed, is a remarkable one;
for at first sight it appears as if this must render it
exceptional to the general law, most substances exhibiting the
opposite thermal effects when stressed. However, here, too, the
action of the stress is opposed by the secondary effects
developed in the substance; for it is found that this substance
contracts when heated, expands when cooled. Again, ice being a
substance which contracts in melting, the effect of pressure is
to facilitate melting, lowering its freezing point. But

64

so soon as a little melting occurs, the resulting liquid calls on
the residual ice for an amount of heat equivalent to the latent
heat of liquefaction, and so by cooling the whole, retards the
change.

Such particular cases illustrate a principle controlling the
interaction of matter and energy which seems universal in
application save when evaded, as we shall see, by the ingenuity
of life. This principle is not only revealed in the researches of
the laboratory; it is manifest in the history of worlds and solar
systems. Thus, consider the effects arising from the aggregation
of matter in space under the influence of the mutual attraction
of the particles. The tendency here is loss of gravitational
potential. The final approach is however retarded by the
temperature, or vis viva of the parts attending collision and
compression. From this cause the great suns of space radiate for
ages before the final loss of potential is attained.

Clerk Maxwell[1] observes on the general principle that less
force is required to produce a change in a body when the change
is unopposed by constraints than when it is subjected to such.
From this if we assume the external forces acting upon a system
not to rise above a certain potential (which is the order of
nature), the constraints of secondary actions may, under certain
circumstances, lead to final rejection of some of the energy, or,
in any

[1] _Theory of Heat_, p. 131.

65

case, to retardation of change in the system--dissipation of
energy being the result.[1]

As such constraints seem inherently present in the properties of
matter, we may summarise as follows:

_The transfer of energy into any inanimate material system is
attended by effects retardative to the transfer and conducive to
dissipation._

Was this the only possible dynamic order ruling in material
systems it is quite certain the myriads of ants and pines never
could have been, except all generated by creative act at vast
primary expenditure of energy. Growth and reproduction would have
been impossible in systems which retarded change at every step
and never proceeded in any direction but in that of dissipation.
Once created, indeed, it is conceivable that, as heat engines,
they might have dragged out an existence of alternate life and
death; life in the hours of sunshine, death in hours of darkness:
no final death, however, their lot, till their parts were simply
worn out by long use, never made good by repair. But the
sustained and increasing activity of organized nature is a fact;
therefore some other order of events must be possible.

[1] The law of Least Action, which has been applied, not alone in
optics, but in many mechanical systems, appears physically based
upon the restraint and retardation opposing the transfer of
energy in material systems.

66

GENERAL DYNAMIC CONDITIONS ATTENDING ANIMATE ACTIONS

What is the actual dynamic attitude of the primary organic
engine--the vegetable organism? We consider, here, in the first
place, not intervening, but resulting phenomena.

The young leaf exposed to solar radiation is small at first, and
the quantity of radiant energy it receives in unit of time cannot
exceed that which falls upon its surface. But what is the effect
of this energy? Not to produce a retardative reaction, but an
accelerative response: for, in the enlarging of the leaf by
growth, the plant opens for itself new channels of supply.

If we refer to "the living protoplasm which, with its unknown
molecular arrangement, is the only absolute test of the cell and
of the organism in general,[1] we find a similar attitude towards
external sources of available energy. In the act of growth
increased rate of assimilation is involved, so that there is an
acceleration of change till a bulk of maximum activity is
attained. The surface, finally, becomes too small for the
absorption of energy adequate to sustain further increase of mass
(Spencer[2]), and the acceleration ceases. The waste going on in
the central parts is then just balanced by the renewal at the
surface. By division, by spreading of the mass, by

[1] Claus, _Zoology_, p. 13.

[2] Geddes and Thomson, _The Evolution of Sex_, p. 220.

67

out-flowing processes, the normal activity of growth may be
restored. Till this moment nothing would be gained by any of
these changes. One or other of them is now conducive to
progressive absorption of energy by the organism, and one or
other occurs, most generally the best of them, subdivision. Two
units now exist; the total mass immediately on division is
unaltered, but paths for the more abundant absorption of energy
are laid open.

The encystment of the protoplasm (occurring under conditions upon
which naturalists do not seem agreed[1]) is to all appearance
protective from an unfavourable environment, but it is often a
period of internal change as well, resulting in a segregation
within the mass of numerous small units, followed by a breakup of
the whole into these units. It is thus an extension of the basis
of supply, and in an impoverished medium, where unit of surface
is less active, is evidently the best means of preserving a
condition of progress.

Thus, in the organism which forms the basis of all modes of life,
a definite law of action is obeyed under various circumstances of
reaction with the available energy of its environment.

Similarly, in the case of the more complex leaf, we see, not only
in the phenomenon of growth, but in its extension in a flattened
form, and in the orientation of greatest surface towards the
source of energy, an attitude towards

[1] However, "In no way comparable with death." Weismann,
_Biological Memoirs_, p. 158.

68

available energy causative of accelerated transfer. There is
seemingly a principle at work, leading to the increase of organic
activity.

Many other examples might be adduced. The gastrula stage in the
development of embryos, where by invagination such an arrangement
of the multiplying cells is secured as to offer the greatest
possible surface consistent with a first division of labour; the
provision of cilia for drawing upon the energy supplies of the
medium; and more generally the specialisation of organs in the
higher developments of life, may alike be regarded as efforts of
the organism directed to the absorption of energy. When any
particular organ becomes unavailing in the obtainment of
supplies, the organ in the course of time becomes aborted or
disappears.[1] On the other hand, when a too ready and liberal
supply renders exertion and specialisation unnecessary, a similar
abortion of functionless organs takes place. This is seen in the
degraded members of certain parasites.

During certain epochs of geological history, the vegetable world
developed enormously; in response probably to liberal supplies of
carbon dioxide. A structural adaptation to the rich atmosphere
occurred, such as was calculated to cooperate in rapidly
consuming the supplies, and to this obedience to a law of
progressive transfer of energy we owe the vast stores of energy
now accumulated

[1] Claus, _Zoology_, p. 157

69

in our coal fields. And when, further, we reflect that this store
of energy had long since been dissipated into space but for the
intervention of the organism, we see definitely another factor in
organic transfer of energy--a factor acting conservatively of
energy, or antagonistically to dissipation.

The tendency of organized nature in the presence of unlimited
supplies is to "run riot." This seems so universal a relation,
that we are safe in seeing here cause and effect, and in drawing
our conclusions as to the attitude of the organism towards
available energy. New species, when they come on the field of
geological history, armed with fresh adaptations, irresistible
till the slow defences of the subjected organisms are completed,
attain enormous sizes under the stimulus of abundant supply, till
finally, the environment, living and dead, reacts upon them with
restraining influence. The exuberance of the organism in presence
of energy is often so abundant as to lead by deprivation to its
self-destruction. Thus the growth of bacteria is often controlled
by their own waste products. A moment's consideration shows that
such progressive activity denotes an accelerative attitude on the
part of the organism towards the transfer of energy into the
organic material system. Finally, we are conscious in ourselves
how, by use, our faculties are developed; and it is apparent that
all such progressive developments must rest on actions which
respond to supplies with fresh demands. Possibly in the present
and ever-

70

increasing consumption of inanimate power by civilised races, we
see revealed the dynamic attitude of the organism working through
thought-processes.

Whether this be so or not, we find generally in organised nature
causes at work which in some way lead to a progressive transfer
of energy into the organic system. And we notice, too, that all
is not spent, but both immediately in the growth of the
individual, and ultimately in the multiplication of the species,
there are actions associated with vitality which retard the
dissipation of energy. We proceed to state the dynamical
principles involved in these manifestations, which appear
characteristic of the organism, as follows:--

_The transfer of energy into any animate material system is
attended by effects conducive to the transfer, and retardative of
dissipation._

This statement is, I think, perfectly general. It has been in
part advanced before, but from the organic more than the physical
point of view. Thus, "hunger is an essential characteristic of
living matter"; and again, "hunger is a dominant characteristic
of living matter,"[1] are, in part, expressions of the statement.
If it be objected against the generality of the statement, that
there are periods in the life of individuals when stagnation and
decay make their appearance, we may answer, that

[1] _Evolution of Sex._ Geddes and Thomson, chap. xvi. See also a
reference to Cope's theory of "Growth Force," in Wallace's
_Darwinism_, p. 425.

71

such phenomena arise in phases of life developed under conditions
of external constraint, as will be urged more fully further on,
and that in fact the special conditions of old age do not and
cannot express the true law and tendency of the dynamic relations
of life in the face of its evident advance upon the Earth. The
law of the unconstrained cell is growth on an ever increasing
scale; and although we assume the organic configuration, whether
somatic or reproductive, to be essentially unstable, so that
continual inflow of energy is required merely to keep it in
existence, this does not vitiate the fact that, when free of all
external constraint, growth gains on waste. Indeed, even in the
case of old age, the statement remains essentially true, for the
phenomena then displayed point to a breakdown of the functioning
power of the cell, an approximation to configurations incapable
of assimilation. It is not as if life showed in these phenomena
that its conditions could obtain in the midst of abundance, and
yet its law be suspended; but as if they represented a
degradation of the very conditions of life, a break up, under the
laws of the inanimate, of the animate contrivance; so that energy
is no longer available to it, or the primary condition, "the
transfer of energy into the animate system," is imperfectly
obeyed. It is to the perfect contrivance of life our statement
refers.

That the final end of all will be general non-availability there
seems little reason to doubt, and the organism, itself dependent
upon differences of potential, cannot

72

hope to carry on aggregation of energy beyond the period when
differences of potential are not. The organism is not accountable
for this. It is being affected by events external to it, by the
actions going on through inanimate agents. And although there be
only a part of the received energy preserved, there is a part
preserved, and this amount is continually on the increase. To see
this it is only necessary to reflect that the sum of animate
energy--capability of doing work in any way through animate
means--at present upon the Earth, is the result, although a small
one, of energy reaching the Earth since a remote period, and
which otherwise had been dissipated in space. In inanimate
actions throughout nature, as we know it, the availability is
continually diminishing. The change is all the one way. As,
however, the supply of available energy in the universe is
(probably) limited in amount, we must look upon the two as simply
effecting the final dissipation of potential in very different
ways. The animate system is aggressive on the energy available to
it, spends with economy, and invests at interest till death
finally deprives it of all. It has heirs, indeed, who inherit
some of its gains, but they, too, must die, and ultimately there
will be no successors, and the greater part must melt away as if
it had never been. The inanimate system responds to the forces
imposed upon it by sluggish changes; of that which is thrust upon
it, it squanders uselessly. The path of the energy is very
different in the two cases.

73

While it is true generally that both systems ultimately result in
the dissipation of energy to uniform potential, the organism can,
as we have seen, under particular circumstances evade the final
doom altogether. It can lay up a store of potential energy which
may be permanent. Thus, so long as there is free oxygen in the
universe, our coalfields might, at any time in the remote future,
generate light and heat in the universal grave.

It is necessary to observe on the fundamental distinction between
the growth of the protoplasm and the growth of the crystal. It is
common to draw comparison between the two, and to point to
metabolism as the chief distinction. But while this is the most
obvious distinction the more fundamental one remains in the
energy relations of the two with the environment.[1] The growth
of the crystal is the result of loss of energy; that of the
organism the result of gain of energy. The crystal represents a
last position of stable equilibrium assumed by molecules upon a
certain loss of kinetic energy, and the formation of the crystal
by evaporation and concentration of a liquid does not, in its
dynamic aspect, differ much from the precipitation of an
amorphous sediment. The organism, on the other hand, represents a
more or less unstable condition formed and maintained by inflow
of energy; its formation, indeed, often attended with a loss of
kinetic energy (fixation of carbon in plants), but, if so,
accompanied by

[1] It appears exceptional for the crystal line configuration to
stand higher in the scale of energy than the amorphous.

74

a more than compensatory increase of potential molecular energy.

Thus, between growth in the living world and growth in the dead
world, the energy relations with the environment reveal a marked
contrast. Again, in the phenomena of combustion, there are
certain superficial resemblances which have led to comparison
between the two. Here again, however, the attitudes towards the
energy of the environment stand very much as + and -. The life
absorbs, stores, and spends with economy. The flame only
recklessly spends. The property of storage by the organism calls
out a further distinction between the course of the two
processes. It secures that the chemical activity of the organism
can be propagated in a medium in which the supply of energy is
discontinuous or localised. The chemical activity of the
combustion can, strictly speaking, only be propagated among
contiguous particles. I need not dwell on the latter fact; an
example of the former is seen in the action of the roots of
plants, which will often traverse a barren place or circumvent an
obstacle in their search for energy. In this manner roots will
find out spots of rich nutriment.

Thus there is a dynamic distinction between the progress of the
organism and the progress of the combustion, or of the chemical
reaction generally. And although there be unstable chemical
systems which absorb energy during reaction, these are
(dynamically) no more than the expansion of the compressed gas.
There is a certain

75

initial capacity in the system for a given quantity of energy;
this satisfied, progress ceases. The progress of the organism in
time is continual, and goes on from less to greater so long as
its development is unconstrained and the supply of energy is
unlimited.

We must regard the organism as a configuration which is so
contrived as to evade the tendency of the universal laws of
nature. Except we are prepared to believe that a violation of the
second law of thermodynamics occurs in the organism, that a
"sorting demon" is at work within it, we must, I think, assume
that the interactions going on among its molecules are
accompanied by retardation and dissipation like the rest of
nature. That such conditions are not incompatible with the
definition of the dynamic attitude of the organism, can be shown
by analogy with our inanimate machines which, by aid of
hypotheses in keeping with the second law of thermodynamics, may
be supposed to fulfil the energy-functions of the plant or
animal, and, in fact, in all apparent respects conform to the
definition of the organism.

We may assume this accomplished by a contrivance of the nature of
a steam-engine, driven by solar energy. It has a boiler, which we
may suppose fed by the action of the engine. It has piston,
cranks, and other movable parts, all subject to resistance from
friction, etc. Now there is no reason why this engine should not
expend its surplus energy in shaping, fitting, and starting into
action other engines:--in fact, in reproductive sacrifice. All

76

these other engines represent a multiplied absorption of energy
as the effects of the energy received by the parent engine, and
may in time be supposed to reproduce themselves. Further, we may
suppose the parent engine to be small and capable of developing
very little power, but the whole series as increasing in power at
each generation. Thus the primary energy relations of the
vegetable organism are represented in these engines, and no
violation of the second law of thermodynamics involved.

We might extend the analogy, and assuming these engines to spend
a portion of their surplus energy in doing work against chemical
forces--as, for example, by decomposing water through the
intervention of a dynamo--suppose them to lay up in this way a
store of potential energy capable of heating the boilers of a
second order of engines, representing the graminivorous animal.
It is obvious without proceeding to a tertiary or carnivorous
order, that the condition of energy in the animal world may be
supposed fulfilled in these successive series of engines, and no
violation of the principles governing the actions going on in our
machines assumed. Organisms evolving on similar principles would
experience loss at every transfer. Thus only a portion of the
radiant energy absorbed by the leaf would be expended in actual
work, chemical and gravitational, etc. It is very certain that
this is, in fact, what takes place.

It is, perhaps, worth passing observation that, from the
nutritive dependence of the animal upon the vegetable,

77

and the fact that a conversion of the energy of the one to the
purposes of the other cannot occur without loss, the mean energy
absorbed daily by the vegetable for the purpose of growth must
greatly exceed that used in animal growth; so that the chemical
potential energy of vegetation upon the earth is much greater
than the energy of all kinds represented in the animal
configurations.[1] It appears, too, that in the power possessed
by the vegetable of remaining comparatively inactive, of
surviving hard times by the expenditure and absorption of but
little, the vegetable constitutes a veritable reservoir for the
uniform supply of the more unstable and active animal.

Finally, on the question of the manner of origin of organic
systems, it is to be observed that, while the life of the present
is very surely the survival of the fittest of the tendencies and
chances of the past, yet, in the initiation of the organised
world, a single chance may have decided a whole course of events:
for, once originated, its own law secures its increase, although
within the new order of actions, the law of the fittest must
assert itself. That such a progressive material system as an
organism was possible, and at some remote period was initiated,
is matter of knowledge; whether or not the initiatory living
configuration was rare and fortuitous, or the probable result of
the general action of physical laws acting among innumerable
chances, must remain matter of

[1] I find a similar conclusion arrived at in Semper's _Animal
Life_, p. 52.

78

speculation. In the event of the former being the truth, it is
evidently possible, in spite of a large finite number of
habitable worlds, that life is non-existent elsewhere. If the
latter is the truth, it is almost certain that there is life in
all, or many of those worlds.

EVOLUTION AND ACCELERATION OF ACTIVITY

The primary factor in evolution is the "struggle for existence."
This involves a "natural selection" among the many variations of
the organism. If we seek the underlying causes of the struggle,
we find that the necessity of food and (in a lesser degree) the
desire for a mate are the principal causes of contention. The
former is much the more important factor, and, accordingly, we
find the greater degree of specialisation based upon it.

The present view assumes a dynamic necessity for its demands
involved in the nature of the organism as such. This assumption
is based on observation of the outcome of its unconstrained
growth, reproduction, and life-acts. We have the same right to
assert this of the organism as we have to assert that retardation
and degradation attend the actions of inanimate machines, which
assertion, also, is based on observation of results. Thus we pass
from the superficial statements that organisms require food in
order to live, or that organisms desire food, to the more
fundamental one that:

_The organism is a configuration of matter which absorbs energy
acceleratively, without limit, when unconstrained._

79

This is the dynamic basis for a "struggle for existence." The
organism being a material system responding to accession of
energy with fresh demands, and energy being limited in amount,
the struggle follows as a necessity. Thus, evolution guiding' the
steps of the energy-seeking organism, must presuppose and find
its origin in that inherent property of the organism which
determines its attitude in presence of available energy.

Turning to the factor, "adaptation," we find that this also must
presuppose, in order to be explicable, some quality of
aggressiveness on the part of the organism. For adaptation in
this or that direction is the result of repulse or victory, and,
therefore, we must presuppose an attack. The attack is made by
the organism in obedience to its law of demand; we see in the
adaptation of the organism but the accumulated wisdom derived
from past defeats and victories.

Where the environment is active, that is living, adaptation
occurs on both sides. Improved means of defence or improved means
of attack, both presuppose activity. Thus the reactions to the
environment, animate and inanimate, are at once the outcome of
the eternal aggressiveness of the organism, and the source of
fresh aggressiveness upon the resources of the medium.

As concerns the "survival of the fittest" (or "natural
selection"), we can, I think, at once conclude that the organism
which best fulfils the organic law under the circumstances of
supply is the "fittest," _ipso facto._ In many

80

cases this is contained in the commonsense consideration, that to
be strong, consistent with concealment from enemies which are
stronger, is best, as giving the organism mastery over foes which
are weaker, and generally renders it better able to secure
supplies. Weismann points out that natural selection favours
early and abundant reproduction. But whether the qualifications
of the "fittest" be strength, fertility, cunning, fleetness,
imitation, or concealment, we are safe in concluding that growth
and reproduction must be the primary qualities which at once
determine selection and are fostered by it. Inherent in the
nature of the organism is accelerated absorption of energy, but
the qualifications of the "fittest" are various, for the supply
of energy is limited, and there are many competitors for it. To
secure that none be wasted is ultimately the object of natural
selection, deciding among the eager competitors what is best for
each.

In short, the facts and generalisations concerning evolution must
presuppose an organism endowed with the quality of progressive
absorption of energy, and retentive of it. The continuity of
organic activity in a world where supplies are intermittent is
evidently only possible upon the latter condition. Thus it
appears that the dynamic attitude of the organism, considered in
these pages, occupies a fundamental position regarding its
evolution.

We turn to the consideration of old age and death, endeavouring
to discover in what relation they stand to the innate
progressiveness of the organism.

81

THE PERIODICITY OF THE ORGANISM AND THE LAW OF PROGRESSIVE
ACTIVITY

The organic system is essentially unstable. Its aggressive
attitude is involved in the phenomenon of growth, and in
reproduction which is a form of growth. But the energy absorbed
is not only spent in growth. It partly goes, also, to make good
the decay which arises from the instability of the organic unit.
The cell is molecularly perishable. It possesses its entity much
as a top keeps erect, by the continual inflow of energy.
Metabolism is always taking place within it. Any other condition
would, probably, involve the difficulties of perpetual motion.

The phenomenon of old age is not evident in the case of the
unicellular organism reproducing by fission. At any stage of its
history all the individuals are of the same age: all contain a
like portion of the original cell, so far as this can be regarded
as persisting where there is continual flux of matter and energy.
In the higher organisms death is universally evident. Why is
this?

The question is one of great complexity. Considered from the more
fundamental molecular point of view we should perhaps look to
failure of the power of cell division as the condition of
mortality. For it is to this phenomenon--that of cell
division--that the continued life of the protozoon is to be
ascribed, as we have already seen. Reproduction is, in fact, the
saving factor here.

As we do not know the source or nature of the stimulus

82

responsible for cell division we cannot give a molecular account
of death in the higher organisms. However we shall now see that,
philosophically, we are entitled to consider reproduction as a
saving factor in this case also; and to regard the death of the
individual much as we regard the fall of the leaf from the tree:
_i.e._ as the cessation of an outgrowth from a development
extending from the past into the future. The phenomena of old age
and natural death are, in short, not at variance with the
progressive activity of the organism. We perceive this when we
come to consider death from the evolutionary point of view.

Professor Weismann, in his two essays, "The Duration of Life,"
and "Life and Death,"[1] adopts and defends the view that "death
is not a primary necessity but that it has been secondarily
acquired by adaptation." The cell was not inherently limited in
its number of cell-generations. The low unicellular organisms are
potentially immortal, the higher multicellular forms with
well-differentiated organs contain the germs of death within
themselves.

He finds the necessity of death in its utility to the species.
Long life is a useless luxury. Early and abundant reproduction is
best for the species. An immortal individual would gradually
become injured and would be valueless or even harmful to the
species by taking the place of those that are sound. Hence
natural selection will shorten life.

[1] See his _Biological Memoirs._ Oxford, 1888.

83

Weismann contends against the transmission of acquired characters
as being unproved.[1] He bases the appearance of death on
variations in the reproductive cells, encouraged by the ceaseless
action of natural selection, which led to a differentiation into
perishable somatic cells and immortal reproductive cells. The
time-limit of any particular organism ultimately depends upon the
number of somatic cell-generations and the duration of each
generation. These quantities are "predestined in the germ itself"
which gives rise to each individual. "The existence of immortal
metazoan organisms is conceivable," but their capacity for
existence is influenced by conditions of the external world; this
renders necessary the process of adaptation. In fact, in the
differentiation of somatic from reproductive cells, material was
provided upon which natural selection could operate to shorten or
to lengthen the life of the individual in accordance with the
needs of the species. The soma is in a sense "a secondary
appendage of the real bearer of life--the reproductive cells." The
somatic cells probably lost their immortal qualities, on this
immortality becoming useless to the species. Their mortality may
have been a mere consequence of their differentiation (loc. cit.,
p. 140), itself due to natural selection. "Natural death was
not," in fact, "introduced from absolute intrinsic necessity
inherent in the nature of living matter, but on grounds of
utility,

[1] Biological Memoirs, p. 142.

84

that is from necessities which sprang up, not from the general
conditions of life, but from those special conditions which
dominate the life of multicellular organisms."

On the inherent immortality of life, Weismann finally states:
"Reproduction is, in truth, an essential attribute of living
matter, just as the growth which gives rise to it.... Life is
continuous, and not periodically interrupted: ever since its
first appearance upon the Earth in the lowest organism, it has
continued without break; the forms in which it is manifest have
alone undergone change. Every individual alive today--even the
highest--is to be derived in an unbroken line from the first and
lowest forms." [1]

At the present day the view is very prevalent that the soma of
higher organisms is, in a sense, but the carrier for a period of
the immortal reproductive cells (Ray Lankester)[2]--an appendage
due to adaptation, concerned in their supply, protection, and
transmission. And whether we regard the time-limit of its
functions as due to external constraints, recurrently acting till
their effects become hereditary, or to variations more directly
of internal origin, encouraged by natural selection, we see in
old age and death phenomena ultimately brought about in obedience
to the action of an environment. These are not inherent in the
properties of living matter. But, in spite

[1] Loc. cit., p. 159

[2] Geddes and Thomson, The Evolution of Sex, chap. xviii.

85

of its mortality, the body remains a striking manifestation of
the progressiveness of the organism, for to this it must be
ascribed. To it energy is available which is denied to the
protozoon. Ingenious adaptations to environment are more
especially its privilege. A higher manifestation, however, was
possible, and was found in the development of mind. This, too, is
a servant of the cell, as the genii of the lamp. Through it
energy is available which is denied to the body. This is the
masterpiece of the cell. Its activity dates, as it were, but from
yesterday, and today it inherits the most diverse energies of the
Earth.

Taking this view of organic succession, we may liken the
individual to a particle vibrating for a moment and then coming
to rest, but sweeping out in its motion one wave in the
continuous organic vibration travelling from the past into the
future. But as this vibration is one spreading with increased
energy from each vibrating particle, its propagation involves a
continual accelerated inflow of energy from the surrounding
medium, a dynamic condition unknown in periodic effects
transmitted by inanimate actions, and, indeed, marking the
fundamental difference between the dynamic attitudes of the
animate and inanimate.

We can trace the periodic succession of individuals on a diagram
of activity with some advantage. Considering, first, the case of
the unicellular organism reproducing by subdivision and recalling
that conditions, definite and inevitable, oppose a limit to the
rate of growth, or, for our

86

present purpose, rate of consumption of energy, we proceed as
follows:

{Fig. 1}

Along a horizontal axis units of time are measured; along a
vertical axis units of energy. Then the life-history of the
amoeba, for example, appears as a line such as A in Fig. 1.
During the earlier stages of its growth the rate of absorption of
energy is small; so that in the unit interval of time, t, the
small quantity of energy, e1, is absorbed. As life advances, the
activity of the organism augments, till finally this rate attains
a maximum, when e2 units of energy are consumed in the unit of
time.[1]

[1] Reference to p. 76, where the organic system is treated as
purely mechanical, may help readers to understand what is
involved in this curve. The solar engine may, unquestionably,
have its activity defined by such a curve. The organism is,
indeed, more complex; but neither this fact nor our ignorance of
its mechanism, affects the principles which justify the diagram.

87

On this diagram reproduction, on the part of the organism, is
represented by a line which repeats the curvature of the parent
organism originating at such a point as P in the path of the
latter, when the rate of consumption of energy has become
constant. The organism A has now ceased to act as a unit. The
products of fission each carry on the vital development of

{Fig. 2}

the species along the curve B, which may be numbered (2), to
signify that it represents the activity of two individuals, and
so on, the numbering advancing in geometrical progression. The
particular curvature adopted in the diagram is, of course,
imaginary; but it is not of an indeterminate nature. Its course
for any species is a characteristic of fundamental physical
importance, regarding the part played in nature by the particular
organism.

88

In Fig. 2 is represented the path of a primitive multicellular
organism before the effects of competition produced or fostered
its mortality. The lettering of Fig. 1 applies; the successive
reproductive acts are marked P1, P2; Q1, Q2, etc., in the paths
of the successive individuals.

{Fig. 3}

The next figure (Fig. 3) diagrammatically illustrates death in
organic history. The path ever turns more and more from the axis
of energy, till at length the point is reached when no more
energy is available; a tangent to the curve at this point is at
right angles to the axis of energy and parallel to the time axis.
The death point is reached, and however great a length we measure
along the axis of time, no further consumption of energy is

89

indicated by the path of the organism. Drawing the line beyond
the death point is meaningless for our present purpose.

It is observable that while the progress of animate nature finds
its representation on this diagram by lines sloping _upwards_ from
left to right, the course of events in inanimate nature--for
example, the history of the organic configuration after death, or

{Fig. 4}

the changes progressing--let us say, in the solar system, or in
the process of a crystallisation, would appear as lines sloping
downwards from left to right.

Whatever our views on the origin of death may be, we have to
recognise a periodicity of functions in the life-history of the
successive individuals of the present day; and whether or not we
trace this directly or indirectly to

90

a sort of interference with the rising wave of life, imposed by
the activity of a series of derived units, each seeking energy,
and in virtue of its adaptation each being more fitted to obtain
it than its predecessor, or even leave the idea of interference
out of account altogether in the origination or perpetuation of
death, the truth of the diagram (Fig. 4) holds in so far as it
may be supposed to graphically represent the dynamic history of
the individual. The point chosen on the curve for the origination
of a derived unit is only applicable to certain organisms, many
reproducing at the very close of life. A chain of units are
supposed here represented.[1]

THE LENGTH OF LIFE

If we lay out waves as above to a common scale of time for
different species, the difference of longevity is shown in the
greater or less number of vibrations executed in a given time,
_i.e._ in greater or less "frequency." We cannot indeed draw the
curvature correctly, for this would necessitate a knowledge which
we have not of the activity of the organism at different periods
of its life-history, and so neither can we plot the direction of
the organic line of propagation with respect to the

[1] Projecting upon the axes of time and energy any one complete
vibration, as in Fig. 4, the total energy consumed by the
organism during life is the length E on the axis of energy, and
its period of life is the length T on the time-axis. The mean
activity is the quotient E/T.

91

axes of reference as this involves a knowledge of the mean
activity.[1]

The group of curves which follow, relating to typical animals
possessing very different activities (Fig. 5), are therefore
entirely diagrammatic, except in respect to the approximate

{Fig. 5}

longevity of the organisms. (1) might represent an animal of the
length of life and of the activity of Man; (2), on the same scale
of longevity,

[1] In the relative food-supply at various periods of life the
curvature is approximately determinable.

92

one of the smaller mammals; and (3), the life-history of a cold
blooded animal living to a great age; _e.g._ certain of the
reptilia.

It is probable, that to conditions of structural development,
under the influence of natural selection, the question of longer
or shorter life is in a great degree referable. Thus, development
along lines of large growth will tend to a slow rate of
reproduction from the simple fact that unlimited energy to supply
abundant reproduction is not procurable, whatever we may assume
as to the strength or cunning exerted by the individual in its
efforts to obtain its supplies. On the other hand, development
along lines of small growth, in that reproduction is less costly,
will probably lead to increased rate of reproduction. It is, in
fact, matter of general observation that in the case of larger
animals the rate of reproduction is generally slower than in the
case of smaller animals. But the rate of reproduction might be
expected to have an important influence in determining the
particular periodicity of the organism. Were we to depict in the
last diagram, on the same time-scale as Man, the vibrations of
the smaller and shorter-lived living things, we would see but a
straight line, save for secular variations in activity,
representing the progress of the species in time: the tiny
thrills of its units lost in comparison with the yet brief period
of Man.

The interdependence of the rate of reproduction and

93

the duration of the individual is, indeed, very probably revealed
in the fact that short-lived animals most generally reproduce
themselves rapidly and in great abundance, and vice versa. In
many cases where this appears contradicted, it will be found that
the young are exposed to such dangers that but few survive (_e.g._
many of the reptilia, etc.), and so the rate of reproduction is
actually slow.

Death through the periodic rigour of the inanimate environment
calls forth phenomena very different from death introduced or
favoured by competition. A multiplicity of effects simulative of
death occur. Organisms will, for example, learn to meet very
rigorous conditions if slowly introduced, and not permanent. A
transitory period of want can be tided over by contrivance. The
lily withdrawing its vital forces into the bulb, protected from
the greatest extremity of rigour by seclusion in the Earth; the
trance of the hibernating animal; are instances of such
contrivances.

But there are organisms whose life-wave truly takes up the
periodicity of the Earth in its orbit. Thus the smaller animals
and plants, possessing less resources in themselves, die at the
approach of winter, propagating themselves by units which,
whether egg or seed, undergo a period of quiescence during the
season of want. In these quiescent units the energy of the
organism is potential, and the time-energy function is in
abeyance. This condition is, perhaps, foreshadowed in the
encyst-

94

ment of the amoeba in resistance to drought. In most cases of
hibernation the time-energy function seems maintained at a loss
of potential by the organism, a diminished vital consumption of
energy being carried on at the expense of the stored energy of
the tissues. So, too, even among the largest organisms there will
be a diminution of activity periodically inspired by
climatological conditions. Thus, wholly or in part, the activity
of organisms is recurrently affected by the great energy--tides
set up by the Earth's orbital motion.

{Fig. 6}

Similarly in the phenomenon of sleep the organism responds to the
Earth's axial periodicity, for in the interval of night a period
of impoverishment has to be endured. Thus the diurnal waves of
energy also meet a response in the organism. These tides and
waves of activity would appear as larger and smaller ripples

95

on the life-curve of the organism. But in some, in which life and
death are encompassed in a day, this would not be so; and for the
annual among plants, the seed rest divides the waves with lines
of no activity (Fig. 6).

Thus, finally, we regard the organism as a dynamic phenomenon
passing through periodic variations of intensity. The material
systems concerned in the transfer of the energy rise, flourish,
and fall in endless succession, like cities of ancient dynasties.
At points of similar phase upon the waves the rate of consumption
of energy is approximately the same; the functions, too, which
demand and expend the energy are of similar nature.

That the rhythm of these events is ultimately based on harmony in
the configuration and motion of the molecules within the germ
seems an unavoidable conclusion. In the life of the individual
rhythmic dynamic phenomena reappear which in some cases have no
longer a parallel in the external world, or under conditions when
the individual is no longer influenced by these external
conditions.,, In many cases the periodic phenomena ultimately die
out under new influences, like the oscillations of a body in a
viscous medium; in others when they seem to be more deeply rooted
in physiological conditions they persist.

The "length of life is dependent upon the number

[1] The _Descent of Man._

96

of generations of somatic cells which can succeed one another in
the course of a single life, and furthermore the number as well
as the duration of each single cell-generation is predestined in
the germ itself."[1]

Only in the vague conception of a harmonising or formative
structural influence derived from the germ, perishing in each
cell from internal causes, but handed from cell to cell till the
formative influence itself degrades into molecular discords, does
it seem possible to form any physical representation of the
successive events of life. The degradation of the molecular
formative influence might be supposed involved in its frequent
transference according to some such dynamic actions as occur in
inanimate nature. Thus, ultimately, to the waste within the cell,
to the presence of a force retardative of its perpetual harmonic
motions, the death of the individual is to be ascribed. Perhaps
in protoplasmic waste the existence of a universal death should
be recognised. It is here we seem to touch inanimate nature; and
we are led back to a former conclusion that the organism in its
unconstrained state is to be regarded as a contrivance for
evading the dynamic tendencies of actions in which lifeless
matter participates.[2]

[1] Weismann, _Life and Death; Biological Memoirs_, p. 146.

[2] In connection with the predestinating power and possible
complexity of the germ, it is instructive to reflect on the very
great molecular population of even the smallest spores--giving
rise to very simple forms. Thus, the spores of the unicellular
Schizomycetes are estimated to dimensions as low as 1/10,000 of a
millimetre in diameter (Cornil et Babes, _Les Batteries_, 1. 37).
From Lord Kelvin's estimate of the number of molecules in water,
comprised within the length of a wave-length of yellow light
(_The Size of Atoms_, Proc. R. I., vol. x., p. 185) it is
probable that such spores contain some 500,000 molecules, while
one hundred molecules range along a diameter.

97

THE NUMERICAL ABUNDANCE OF LIFE

We began by seeking in various manifestations of life a dynamic
principle sufficiently comprehensive to embrace its very various
phenomena. This, to all appearance, found, we have been led to
regard life, to a great extent, as a periodic dynamic phenomenon.
Fundamentally, in that characteristic of the contrivance, which
leads it to respond favourably to transfer of energy, its
enormous extension is due. It is probable that to its instability
its numerical abundance is to be traced--for this, necessitating
the continual supply of all the parts already formed, renders
large, undifferentiated growth, incompatible with the limited
supplies of the environment. These are fundamental conditions of
abundant life upon the Earth.

Although we recognise in the instability of living systems the
underlying reason for their numerical abundance, secondary
evolutionary causes are at work. The most important of these is
the self-favouring nature of the phenomenon of reproduction. Thus
there is a tendency not only to favour reproductiveness, but
early reproductiveness, in the form of one prolific
reproductive.

98

act, after which the individual dies.[1] Hence the wavelength of
the species diminishes, reproduction is more frequent, and
correspondingly greater numbers come and go in an interval of
time.

Another cause of the numerical abundance of life exists, as
already stated, in the conditions of nourishment. Energy is more
readily conveyed to the various parts of the smaller mass, and
hence the lesser organisms will more actively functionate; and
this, as being the urging dynamic attitude, as well as that most
generally favourable in the struggle, will multiply and favour
such forms of life. On the other hand, however, these forms will
have less resource within themselves, and less power of
endurance, so that they are only suitable to fairly uniform
conditions of supply; they cannot survive the long continued want
of winter, and so we have the seasonal abundance of summer. Only
the larger and more resistant organisms, whether animal or
vegetable, will, in general, populate the Earth from year to
year. From this we may conclude that, but for the seasonal
energy-tides, the development of life upon the globe had gone
along very different lines from those actually followed. It is,
indeed, possible that the evolution of the larger organisms would
not have occurred; there would have been no vacant place for
their development, and a being so endowed as Man could hardly

[1] Weismann, _The Duration of Life._

99

have been evolved. We may, too, apply this reasoning elsewhere,
and regard as highly probable, that in worlds which are without
seasonal influences, the higher developments of life have not
appeared; except they have been evolved under other conditions,
when they might for a period persist. We have, indeed, only to
picture to ourselves what the consequence of a continuance of
summer would be on insect life to arrive at an idea of the
antagonistic influences obtaining in such worlds to the survival
of larger organisms.

It appears that to the dynamic attitude of life in the first
place, and secondarily to the environmental conditions limiting
undifferentiated growth, as well as to the action of heredity in
transmitting the reproductive qualities of the parent to the
offspring, the multitudes of the pines, and the hosts of ants,
are to be ascribed. Other causes are very certainly at work, but
these, I think, must remain primary causes.

We well know that the abundance of the ants and pines is not a
tithe of the abundance around us visible and invisible. It is a
vain endeavour to realise the countless numbers of our
fellow-citizens upon the Earth; but, for our purpose, the
restless ants, and the pines solemnly quiet in the sunshine, have
served as types of animate things. In the pine the gates of the
organic have been thrown open that the vivifying river of energy
may flow in. The ants and the butterflies sip for a brief moment
of its waters, and again vanish into the

100

inorganic: life, love and death encompassed in a day.

Whether the organism stands at rest and life comes to it on the
material currents of the winds and waters, or in the vibratory
energy of the æther; or, again, whether with restless craving it
hurries hither and thither in search of it, matters nothing. The
one principle--the accelerative law which is the law of the
organic--urges all alike onward to development, reproduction and
death. But although the individual dies death is not the end; for
life is a rhythmic phenomenon. Through the passing ages the waves
of life persist: waves which change in their form and in the
frequency to which they are attuned from one geologic period to
the next, but which still ever persist and still ever increase.
And in the end the organism outlasts the generations of the
hills.

101

THE BRIGHT COLOURS OF ALPINE FLOWERS [1]

IT is admitted by all observers that many species of flowering
plants growing on the higher alps of mountainous regions display
a more vivid and richer colour in their bloom than is displayed
in the same species growing in the valleys. That this is actually
the case, and not merely an effect produced upon the observer by
the scant foliage rendering the bloom more conspicuous, has been
shown by comparative microscopic examination of the petals of
species growing on the heights and in the valleys. Such
examination has revealed that in many cases pigment granules are
more numerous in the individuals growing at the higher altitudes.
The difference is specially marked in Myosotis sylvatica,
Campanula rotundifolia, Ranunculus sylvaticus, Galium cruciatum,
and others. It is less marked in the case of Thymus serpyllum and
Geranium sylvaticum; while in Rosa alpina and Erigeron alpinus no
difference is observable.[2]

In the following cases a difference of intensity of colour is,
according to Kerner ("Pflanzenleben," 11. 504), especially
noticeable:-- _Agrostemma githago, Campanula

[1] _Proc. Royal Dublin Society_, 1893.

[2] G. Bonnier, quoted by De Varigny, _Experimental Evolution_,
p. 55.

102

pusilla, Dianthus inodorus (silvestris), Gypsophila repens, Lotus
corniculatus, Saponaria ocymoides, Satureja hortensis, Taraxacumm
officinale, Vicia cracca, and Vicia sepium._

To my own observation this beautiful phenomenon has always
appeared most obvious and impressive. It appears to have struck
many unprofessional observers. Helmholtz offers the explanation
that the vivid colours are the result of the brighter sunlight of
the heights. It has been said, too, that they are the direct
chemical effects of a more highly ozonized atmosphere. The latter
explanation I am unable to refer to its author. The following
pages contain a suggestion on the matter, which occurred to me
while touring, along with Henry H. Dixon, in the Linthal district
of Switzerland last summer.[1]

If the bloom of these higher alpine flowers is especially
pleasing to our own æsthetic instincts, and markedly conspicuous
to us as observers, why not also especially attractive and
conspicuous to the insect whose mission it is to wander from
flower to flower over the pastures? The answer to this question
involves the hypothesis I would advance as accounting for the
bright colours of high-growing individuals. In short, I believe a
satisfactory explanation is to be found in the conditions of
insect life in the higher alps.

In the higher pastures the summer begins late and

[1] The summer of 1892.

103

closes early, and even in the middle of summer the day closes in
with extreme cold, and the cold of night is only dispelled when
the sun is well up. Again, clouds cover the heights when all is
clear below, and cold winds sweep over them when there is warmth
and shelter in the valleys. With these rigorous conditions the
pollinating insects have to contend in their search for food, and
that when the rival attractions of the valleys below are so many.
I believe it is these rigorous conditions which are indirectly
responsible for the bright colours of alpine flowers. For such
conditions will bring about a comparative scarcity of insect
activity on the heights; and a scarcity or uncertainty in the
action of insect agency in effecting fertilization will intensify
the competition to attract attention, and only the brightest
blooms will be fertilized.[1]

This will be a natural selection of the brightest, or the

[1] Grant Allen, I have recently learned, advances in _Science in
Arcady_ the theory that there is a natural selective cause
fostering the bright blooms of alpines. The selective cause is,
however, by him referred to the greater abundance of butterfly
relatively to bee fertilizers. The former, he says, display more
æsthetic instinct than bees. In the valley the bees secure the
fertilization of all. I may observe that upon the Fridolins Alp
all the fertilizers we observed were bees. I have always found
butterflies very scarce at altitudes of 7,000 to 8,000 feet. The
alpine bees are very light in body, like our hive bee, and I do
not think rarefaction of the atmosphere can operate to hinder its
ascent to the heights, as Grant Allen suggests. The observations
on the death-rate of bees and butterflies on the glacier, to be
referred to presently, seem to negative such a hypothesis, and to
show that a large preponderance of bees over butterflies make
their way to the heights.

104

brightest will be the fittest, and this condition, along with the
influence of heredity, will encourage a race of vivid flowers. On
the other hand, the more scant and uncertain root supply, and the
severe atmospheric conditions, will not encourage the grosser
struggle for existence which in the valleys is carried on so
eagerly between leaves and branches--the normal offensive and
defensive weapons of the plant--and so the struggle becomes
refined into the more æsthetic one of colour and brightness
between flower and flower. Hence the scant foliage and vivid
bloom would be at once the result of a necessary economy, and a
resort to the best method of securing reproduction under the
circumstances of insect fertilizing agency. Or, in other words,
while the luxuriant growth is forbidden by the conditions, and
thus methods of offence and defence, based upon vigorous
development, reduced in importance, it would appear that the
struggle is mainly referred to rivalry for insect preference. It
is probable that this is the more economical manner of carrying
on the contest.

In the valleys we see on every side the struggle between the
vegetative organs of the plant; the soundless battle among the
leaves and branches. The blossom here is carried aloft on a
slender stem, or else, taking but a secondary part in the
contest, it is relegated to obscurity (P1. XII.). Further up on
the mountains, where the conditions are more severe and the
supplies less abundant, the leaf and branch assume lesser
dimensions, for they are costly weapons to provide and the
elements are unfriendly

105

to their existence (Pl. XIII.). Still higher, approaching the
climatic limit of vegetable life, the struggle for existence is
mainly carried on by the æsthetic rivalry of lowly but
conspicuous blossoms.

As regards the conditions of insect life in the higher alps, it
came to my notice in a very striking manner that vast numbers of
such bees and butterflies as venture up perish in the cold of
night time. It appears as if at the approach of dusk these are
attracted by the gleam of the snow, and quitting the pastures,
lose themselves upon the glaciers and firns, there to die in
hundreds. Thus in an ascent of the Tödi from the Fridolinshüte we
counted in the early dawn sixty-seven frozen bees, twenty-nine
dead butterflies, and some half-dozen moths on the Biferten
Glacier and Firn. These numbers, it is to be remembered, only
included those lying to either side of our way over the snow, so
that the number must have mounted up to thousands when integrated
over the entire glacier and firn. Approaching the summit none
were found. The bees resembled our hive bee in appearance, the
butterflies resembled the small white variety common in our
gardens, which has yellow and black upon its wings. One large
moth, striped across the abdomen, and measuring nearly two inches
in length of body, was found. Upon our return, long after the
sun's rays had grown strong, we observed some of the butterflies
showed signs of reanimation. We descended so quickly to avoid the
inconvenience of the soft snow that we had time for no

106

close observation on the frozen bees. But dead bees are common
objects upon the snows of the alps.

These remarks I noted down roughly while at Linthal last summer,
but quite recently I read in Natural Science[1] the following
note:

"Late Flowering Plants.--While we write, the ivy is in flower, and
bees, wasps, and flies are jostling each other and struggling to
find standing-room on the sweet-smelling plant. How great must be
the advantage obtained by this plant through its exceptional
habit of flowering in the late autumn, and ripening its fruit in
the spring. To anyone who has watched the struggle to approach
the ivy-blossom at a time when nearly all other plants are bare,
it is evident that, as far as transport of pollen and
cross-fertilization go, the plant could not flower at a more
suitable time. The season is so late that most other plants are
out of flower, but yet it is not too late for many insects to be
brought out by each sunny day, and each insect, judging by its
behaviour, must be exceptionally hungry.

"Not only has the ivy the world to itself during its flowering
season, but it delays to ripen its seed till the spring, a time
when most other plants have shed their seed, and most edible
fruits have been picked by the birds. Thus birds wanting fruit in
the spring can obtain little but ivy, and how they appreciate the
ivy berry is evident

[1] For December, 1892, vol. i., p. 730.

107

by the purple stains everywhere visible within a short distance
of the bush."

These remarks suggest that the ivy adopts the converse attitude
towards its visitors to that forced upon the alpine flower. The
ivy bloom is small and inconspicuous, but then it has the season
to itself, and its inconspicuousness is no disadvantage, _i.e._
if one plant was more conspicuous than its neighbours, it would
not have any decided advantage where the pollinating insect is
abundant and otherwise unprovided for. Its dark-green berries in
spring, which I would describe as very inconspicuous, have a
similar advantage in relation to the necessities of bird life.

The experiments of M. C. Flahault must be noticed. This
naturalist grew seeds of coloured flowers which had ripened in
Paris, part in Upsala, and part in Paris; and seed which had
ripened in Upsala, part at Paris, and part at Upsala. The flowers
opening in the more northern city were in most cases the
brighter.[1] If this observation may be considered indisputable,
as appears to be the case, the question arises, Are we to regard
this as a direct effect of the more rigorous climate upon the
development of colouring matter on the blooms opening at Upsala?
If we suppose an affirmative answer, the theory of direct effect
by sun brightness must I think be abandoned. But I venture to
think that the explanation of the Upsala

[1] Quoted by De Varigny, _Experimental Evolution_, p. 56.

108

experiment is not to be found in direct climatic influence upon
the colour, but in causes which lie deeper, and involve some
factors deducible from biological theory.

The organism, as a result of the great facts of heredity and of
the survival of the fittest, is necessarily a system which
gathers experience with successive generations; and the principal
lesson ever being impressed upon it by external events is
economy. Its success depends upon the use it makes of its
opportunities for the reception of energy and the economy
attained in disposing of what is gained.

With regard to using the passing opportunity the entire seasonal
development of life is a manifestation of this attitude, and the
fleetness, agility, etc., of higher organisms are developments in
this direction. The higher vegetable organism is not locomotory,
save in the transferences of pollen and seed, for its food comes
to it, and the necessary relative motion between food and
organism is preserved in the quick motion of radiated energy from
the sun and the slower motion of the winds on the surface of the
earth. But, even so, the vegetable organism must stand ever ready
and waiting for its supplies. Its molecular parts must be ready
to seize the prey offered to it, somewhat as the waiting spider
the fly. Hence, the plant stands ready; and every cloud with
moving shadow crossing the fields handicaps the shaded to the
benefit of the unshaded plant in the adjoining field. The open
bloom

109

is a manifestation of the generally expectant attitude of the
plant, but in relation to reproduction.

As regards economy, any principle of maximum economy, where many
functions have to be fulfilled, will, we may very safely predict,
involve as far as possible mutual helpfulness in the processes
going on. Thus the process of the development towards meeting any
particular external conditions, A, suppose, will, if possible,
tend to forward the development towards meeting conditions B; so
that, in short, where circumstances of morphology and physiology
are favourable, the ideally economical system will be attained
when in place of two separate processes, a, ß, the one process y,
cheaper than a + ß, suffices to advance development
simultaneously in both the directions A and B. The economy is as
obvious as that involved in "killing two birds with the one
stone"--if so crude a simile is permissible--and it is to be
expected that to foster such economy will be the tendency of
evolution in all organic systems subjected to restraints as those
we are acquainted with invariably are.

Such economy might be simply illustrated by considering the case
of a reservoir of water elevated above two hydraulic motors, so
that the elevated mass of water possessed gravitational
potential. The available energy here represents the stored-up
energy in the organism. How best may the water be conveyed to the
two motors [the organic systems reacting towards conditions A and
B] so

110

that as little energy as possible is lost in transit? If the
motors are near together it is most economical to use the one
conduit, which will distribute the requisite supply of water to
both. If the motors are located far asunder it will be most
economical to lay separate conduits. There is greatest economy in
meeting a plurality of functions by the same train of
physiological processes where this is consistent with meeting
other demands necessitated by external or internal conditions.

But an important and obvious consequence arises in the supply of
the two motors from the one conduit. We cannot work one motor
without working the other. If we open a valve in the conduit both
motors start into motion and begin consuming the energy stored in
the tank. And although they may both under one set of conditions
be doing useful and necessary work, in some other set of
conditions it may be needless for both to be driven.

This last fact is an illustration of a consideration which must
enter into the phenomenon which an eminent biologist speaks of as
physiological or unconscious "memory,"[1] For the development of
the organism from the ovum is but the starting of a train of
interdependent events of a complexity depending upon the
experience of the past.

[1] Ewald Hering, quoted by Ray Lankaster, _The Advancement of
Science_, p. 283.

111

In short, we may suppose the entire development of the plant,
towards meeting certain groups of external conditions,
physiologically knit together according as Nature tends to
associate certain groups of conditions. Thus, in the case in
point, climatic rigour and scarcity of pollinating agency will
ever be associated; and in the long experience of the past the
most economical physiological attitude towards both is, we may
suppose, adopted; so that the presence of one condition excites
the apparent unconscious memory of the other. In reality the
process of meeting the one condition involves the process and
development for meeting the other.

And this consideration may be extended very generally to such
organisms as can survive under the same associated natural
conditions, for the history of evolution is so long, and the
power of locomotion so essential to the organism at some period
in its life history, that we cannot philosophically assume a
local history for members of a species even if widely severed
geographically at the present day. At some period in the past
then, it is very possible that the individuals today thriving at
Paris, acquired the experience called out at Upsala. The
perfection of physiological memory inspires no limit to the date
at which this may have occurred--possibly the result of a
succession of severe seasons at Paris; possibly the result of
migrations --and the seed of many flowering plants possess means
of migration only inferior to those possessed by the flying and
swimming animals. But, again, possibly the experi-

112

ence was acquired far back in the evolutionary history of the
flower.[1]

But a further consideration arises. Not only at each moment in
the life of the individual must maximum income and most judicious
expenditure be considered, but in its whole life history, and
even over the history of its race, the efficiency must tend to be
a maximum. This principle is even carried so far that when
necessary it leads to the death of the individual, as in the case
of those organisms which, having accomplished the reproductive
act, almost immediately expire. This view of nature may be
repellent, but it is, nevertheless, evident that we are parts of
a system which ruthlessly sacrifices the individual on general
grounds of economy. Thus, if the curve which defines the mean
rate of reception of energy of all kinds at different periods in
the life of the organism be opposed by a second curve, drawn
below the axis along which time is measured, representing the
mean rate of expenditure of energy on development, reproduction,
etc. (Fig. 7), this latter curve, which is, of course,

[1] The blooms of self-fertilising, and especially of
cleistogamic plants (_e.g._ Viola), are examples of unconscious
memory, or unconscious "association of ideas" leading to the
development of organs now functionless. The _Pontederia crassipes_
of the Amazon, which develops its floating bladders when grown in
water, but aborts them rapidly when grown on land, and seems to
retain this power of adaptation to the environment for an
indefinite period of time, must act in each case upon an
unconscious memory based upon past experience. Many other cases
might be cited.

113

physiologically dependent on the former, must be of such a nature
from its origin to its completion in death, that the condition is
realized of the most economical rate of expenditure at each
period of life.[1] The rate of expenditure of energy at any
period of life is, of course, in such a curve defined by the
slope of the curve towards the axis of time at the period in
question; but this particular slope _must be led to by a previous
part of the curve, and involves its past and future course to a
very great extent_.

{Fig. 7}

There will, therefore, be impressed upon the
organism by the factors of evolution a unified course of
economical expenditure completed only by its death, and which
will give to the developmental progress of the individual its
prophetic character.

In this way we look to the unified career of each organic unit,
from its commencement in the ovum to the day

[1] See _The Abundance of Life_.

114

when it is done with vitality, for that preparation for momentous
organic events which is in progress throughout the entire course
of development; and to the economy involved in the welding of
physiological processes for the phenomenon of physiological
memory, wherein we see reflected, as it were, in the development
of the organism, the association of inorganic restraints
occurring in nature which at some previous period impressed
itself upon the plastic organism. We may picture the seedling at
Upsala, swayed by organic memory and the inherited tendency to an
economical preparation for future events, gradually developing
towards the æsthetic climax of its career. In some such manner
only does it appear possible to account for the prophetic
development of organisms, not alone to be observed in the alpine
flowers, but throughout nature.

And thus, finally, to the effects of natural selection and to
actions defined by general principles involved in biology, I
would refer for explanation of the manner in which flowers on the
Alps develop their peculiar beauty.

115

MOUNTAIN GENESIS

OUR ancestors regarded mountainous regions with feelings of
horror, mingled with commiseration for those whom an unkindly
destiny had condemned to dwell therein. We, on the other hand,
find in the contemplation of the great alps of the Earth such
peaceful and elevated thoughts, and such rest to our souls, that
it is to those very solitudes we turn to heal the wounds of ife.
It is difficult to explain the cause of this very different point
of view. It is probably, in part, to be referred to that cloud of
superstitious horror which, throughout the Middle Ages, peopled
the solitudes with unknown terrors; and, in part, to the
asceticism which led the pious to regard the beauty and joy of
life as snares to the soul's well-being. In those eternal
solitudes where the overwhelming forces of Nature are most in
evidence, an evil principle must dwell or a dragon's dreadful
brood must find a home.

But while in our time the aesthetic aspect of the hills appeals
to all, there remains in the physical history of the mountains
much that is lost to those who have not shared in the scientific
studies of alpine structure and genesis. They lose a past history
which for interest com-

116

petes with anything science has to tell of the changes of the
Earth.

Great as are the physical features of the mountains compared with
the works of Man, and great as are the forces involved compared
with those we can originate or control, the loftiest ranges are
small contrasted with the dimensions of the Earth. It is well to
bear this in mind. I give here (Pl. XV.) a measured drawing
showing a sector cut from a sphere of 50 cms. radius; so much of
it as to exhibit the convergence of its radial boundaries which
if prolonged will meet at the centre. On the same scale as the
radius the diagram shows the highest mountains and the deepest
ocean. The average height of the land and the average depth of
the ocean are also exhibited. We see how small a movement of the
crust the loftiest elevation of the Himalaya represents and what
a little depression holds the ocean.

Nevertheless, it is not by any means easy to explain the genesis
of those small elevations and depressions. It would lead us far
from our immediate subject to discuss the various theoretical
views which have been advanced to account for the facts. The idea
that mountain folds, and the lesser rugosities of the Earth's
surface, arose in a wrinkling of the crust under the influence of
cooling and skrinkage of the subcrustal materials, is held by
many eminent geologists, but not without dissent from others.

The most striking observational fact connected with mountain
structure is that, without exception, the ranges

117

of the Earth are built essentially of sedimentary rocks: that is
of rocks which have been accumulated at some remote past time
beneath the surface of the ocean. A volcanic core there may
sometimes be--probably an attendant or consequence of the
uplifting--or a core of plutonic igneous rocks which has arisen
under the same compressive forces which have bowed and arched the
strata from their original horizontal position. It is not
uncommon to meet among unobservant people those who regard all
mountain ranges as volcanic in origin. Volcanoes, however, do not
build mountain ranges. They break out as more or less isolated
cones or hills. Compare the map of the Auvergne with that of
Switzerland; the volcanoes of South Italy with the Apennines.
Such great ranges as those which border with triple walls the
west coast of North America are in no sense volcanic: nor are the
Pyrenees, the Caucasus, or the Himalaya. Volcanic materials are
poured out from the summits of the Andes, but the range itself is
built up of folded sediments on the same architecture as the
other great ranges of the Earth.

Before attempting an explanation of the origin of the mountains
we must first become more closely acquainted with the phenomena
attending mountain elevation.

At the present day great accumulations of sediment are taking
place along the margins of the continents where the rivers reach
the ocean. Thus, the Gulf of Mexico receiving the sediment of the
Mississippi and Rio Grande;

118

the northeast coast of South America receiving the sediments of
the Amazons; the east coast of Asia receiving the detritus of the
Chinese rivers; are instances of such areas of deposition. Year
by year, century by century, the accumulation progresses, and as
it grows the floor of the sea sinks under the load. Of the
yielding of the crust under the burthen of the sediments we are
assured; for otherwise the many miles of vertically piled strata
which are uplifted to our view in the mountains, never could have
been deposited in the coastal seas of the past. The flexure and
sinking of the crust are undeniable realities.

Such vast subsiding areas are known as geosynclines. From the
accumulated sediments of the geosynclines the mountain ranges of
the past have in every case originated; and the mountains of the
future will assuredly arise and lofty ranges will stand where now
the ocean waters close over the collecting sediments. Every
mountain range upon the Earth enforces the certainty of this
prediction.

The mountain-forming movement takes place after a certain great
depth of sediment is collected. It is most intense where the
thickness of deposit is greatest. We see this when we examine the
structure of our existing mountain ranges. At either side where
the sediments thin out, the disturbance dies away, till we find
the comparatively shallow and undisturbed level sediments which
clothe the continental surface.

Whatever be the connection between the deposition and

119

the subsequent upheaval, _the element of great depth of
accumulation seems a necessary condition and must evidently enter
as a factor into the Physical Processes involved_. The mountain
range can only arise where the geosyncline is deeply filled by
long ages of sedimentation.

Dana's description of the events attending mountain building is
impressive:

"A mountain range of the common type, like that to which the
Appalachians belong, is made out of the sedimentary formations of
a long preceding era; beds that were laid down conformably, and
in succession, until they had reached the needed thickness; beds
spreading over a region tens of thousands of square miles in
area. The region over which sedimentary formations were in
progress in order to make, finally, the Appalachian range,
reached from New York to Alabama, and had a breadth of 100 to 200
miles, and the pile of horizontal beds along the middle was
40,000 feet in depth. The pile for the Wahsatch Mountains was
60,000 feet thick, according to King. The beds for the
Appalachians were not laid down in a deep ocean, but in shallow
waters, where a gradual subsidence was in progress; and they at
last, when ready for the genesis, lay in a trough 40,000 feet
deep, filling the trough to the brim. It thus appears that epochs
of mountain-making have occurred only after long intervals of
quiet in the history of a continent."[1]

[1] Dana, _Manual of Geology_, third edition, p. 794

120

On the western side of North America the work of
mountain-building was, indeed, on the grandest scale. For long
ages and through a succession of geological epochs, sedimentation
had proceeded so that the accumulations of Palaeozoic and
Mesozoic times had collected in the geosyncline formed by their
own ever increasing weight. The site of the future Laramide range
was in late Cretaceous times occupied by some 50,000 feet of
sedimentary deposits; but the limit had apparently been attained,
and at this time the Laramide range, as well as its southerly
continuation into the United States, the Rockies, had their
beginning. Chamberlin and Salisbury[1] estimate that the height
of the mountains developed in the Laramide range at this time was
20,000 feet, and that, owing to the further elevation which has
since taken place, from 32,000 to 35,000 feet would be their
present height if erosion had not reduced them. Thus on either
side of the American continent we have the same forces at work,
throwing up mountain ridges where the sediments had formerly been
shed into the ocean.

These great events are of a rhythmic character; the crust, as it
were, pulsating under the combined influences of sedimentation
and denudation. The first involves downward movements under a
gathering load, and ultimately a reversal of the movement to one
of upheaval; the second factor, which throughout has been in

[1] Chamberlin and Salisbury, _Geology_, 1906, iii., 163.

121

operation as creator of the sediments, then intervenes as an
assailant of the newly-raised mountains, transporting their
materials again to the ocean, when the rhythmic action is
restored to its first phase, and the age-long sequence of events
must begin all over again.

It has long been inferred that compressive stress in the crust
must be a primary condition of these movements. The wvork
required to effect the upheavals must be derived from some
preexisting source of energy. The phenomenon--intrinsically one of
folding of the crust--suggests the adjustment of the earth-crust
to a lessening radius; the fact that great mountain-building
movements have simultaneously affected the entire earth is
certainly in favour of the view that a generally prevailing cause
is at the basis of the phenomenon.

The compressive stresses must be confined to the upper few miles
of the crust, for, in fact, the downward increase of temperature
and pressure soon confers fluid properties on the medium, and
slow tangential compression results in hydrostatic pressure
rather than directed stresses. Thus the folding visible in the
mountain range, and the lateral compression arising therefrom,
are effects confined to the upper parts of the crust.

The energy which uplifts the mountain is probably a surviving
part of the original gravitational potential energy of the crust
itself. It must be assumed that the crust in following downwards
the shrinking subcrustal magma, develops immense compressive
stresses in

122

its materials, vast geographical areas being involved. When
folding at length takes place along the axis of the elongated
syncline of deposition, the stresses find relief probably for
some hundreds of miles, and the region of folding now becomes
compressed in a transverse direction. As an illustration, the
Laramide range, according to Dawson, represents the reduction of
a surface-belt 50 miles wide to one of 25 miles. The marvellous
translatory movements of crustal folds from south to north
arising in the genesis of the Swiss Alps, which recent research
has brought to light, is another example of these movements of
relief, which continue to take place perhaps for many millions of
years after they are initiated.

The result of this yielding of the crust is a buckling of the
surface which on the whole is directed upwards; but depression
also is an attendant, in many cases at least, on mountain
upheaval. Thus we find that the ocean floor is depressed into a
syncline along the western coast of South America; a trough
always parallel to the ranges of the Andes. The downward
deflection of the crust is of course an outcome of the same
compressive stresses which elevate the mountain.

The fact that the yielding of the crust is always situated where
the sediments have accumulated to the greatest depth, has led to
attempts from time to time of establishing a physical connexion
between the one and the other. The best-known of these theories
is that of Babbage and Herschel. This seeks the connexion in the
rise of the

123

geotherms into the sinking mass of sediment and the consequent
increase of temperature of the earth-crust beneath. It will be
understood that as these isogeotherms, or levels at which the
temperature is the same, lie at a uniform distance from the
surface all over the Earth, unless where special variations of
conductivity may disturb them, the introduction of material
pressed downwards from above must result in these materials
partaking of the temperature proper to the depth to which they
are depressed. In other words the geotherms rise into the sinking
sediments, always, however, preserving their former average
distance from the surface. The argument is that as this process
undoubtedly involves the heating up of that portion of the crust
which the sediments have displaced downwards, the result must be
a local enfeeblement of the crust, and hence these areas become
those of least resistance to the stresses in the crust.

When this theory is examined closely, we see that it only amounts
to saying that the bedded rocks, which have taken the place of
the igneous materials beneath, as a part of the rigid crust of
the Earth, must be less able to withstand compressive stress than
the average crust. For there has been no absolute rise of the
geotherms, the thermal conductivities of both classes of
materials differing but little. Sedimentary rock has merely taken
the place of average crust-rock, and is subjected to the same
average temperature and pressure prevailing in the surrounding
crust. But are there any grounds for the

124

assumption that the compressive resistance of a complex of
sedimentary rocks is inferior to one of igneous materials? The
metamorphosed siliceous sediments are among the strongest rocks
known as regards resistance to compressive stress; and if
limestones have indeed plastic qualities, it must be remembered
that their average amount is only some 5 per cent. of the whole.
Again, so far as rise of temperature in the upper crust may
affect the question, a temperature which will soften an average
igneous rock will not soften a sedimentary rock, for the reason
that the effect of solvent denudation has been to remove those
alkaline silicates which confer fusibility.

When, however, we take into account the radioactive content of
the sediments the matter assumes a different aspect.

The facts as to the general distribution of radioactive
substances at the surface, and in rocks which have come from
considerable depths in the crust, lead us to regard as certain
the widespread existence of heat-producing radioactive elements
in the exterior crust of the Earth. We find, indeed, in this fact
an explanation--at least in part--of the outflow of heat
continually taking place at the surface as revealed by the rising
temperature inwards. And we conclude that there must be a
thickness of crust amounting to some miles, containing the
radioactive elements.

Some of the most recent measurements of the quantities of radium
and thorium in the rocks of igneous origin--_e.g._ granites,
syenites, diorites, basalts, etc., show that the

125

radioactive heat continually given out by such rocks amounts to
about one millionth part of 0.6 calories per second per cubic
metre of average igneous rock. As we have to account for the
escape of about 0.0014 calorie[1] per square metre of the Earth's
surface per second (assuming the rise of temperature downwards,
_i.e._ the "gradient" of temperature, to be one degree centigrade
in 35 metres) the downward extension of such rocks might, _prima
facie_, be as much as 19 kilometres.

About this calculation we have to observe that we assume the
average radioactivity of the materials with which we have dealt
at the surface to extend uniformly all the way down, _i.e._ that
our experiments reveal the average radioactivity of a radioactive
crust. There is much to be said for this assumption. The rocks
which enter into the measurements come from all depths of the
crust. It is highly probable that the less silicious, _i.e._ the
more basic, rocks, mainly come from considerable depths; the more
acid or silica-rich rocks, from higher levels in the crust. The
radioactivity determined as the mean of the values for these two
classes of rock closely agrees with that found for intermediate
rocks, or rocks containing an intermediate amount of silica.
Clarke contends that this last class of material probably
represents the average composition of the Earth's crust so far as
it has been explored by us.

[1] The calorie referred to is the quantity of heat required to
heat one gram of water, _i.e._ one cubic centimetre of
water--through one degree centigrade.

126

It is therefore highly probable that the value found for the mean
radioactivity of acid and basic rocks, or that found for
intermediate rocks, truly represents the radioactive state of the
crust to a considerable depth. But it is easy to show that we
cannot with confidence speak of the thickness of this crust as
determinable by equating the heat outflow at the surface with the
heat production of this average rock.

This appears in the failure of a radioactive layer, taken at a
thickness of about 19-kilometres, to account for the deep-seated
high temperatures which we find to be indicated by volcanic
phenomena at many places on the surface. It is not hard to show
that the 19-kilometre layer would account for a temperature no
higher than about 270° >C. at its base.

It is true that this will be augmented beneath the sedimentary
deposits as we shall presently see; and that it is just in
association with these deposits that deep-seated temperatures are
most in evidence at the surface; but still the result that the
maximum temperature beneath the crust in general attains a value
no higher than 270° C. is hardly tenable. We conclude, then, that
some other source of heat exists beneath. This may be radioactive
in origin and may be easily accounted for if the radioactive
materials are more sparsely distributed at the base of the upper
crust. Or, again, the heat may be primeval or original heat,
still escaping from a cooling world. For our present purpose it
does not much matter which view

127

we adopt. But we must recognise that the calculated depth of 19
kilometres of crust, possessing the average radioactivity of the
surface, is excessive; for, in fact, we are compelled by the
facts to recognise that some other source of heat exists
beneath.

If the observed surface gradient of temperature persisted
uniformly downwards, at some 35 kilometres beneath the surface
there would exist temperatures (of about 1000° C.) adequate to
soften basic rocks. It is probable, however, that the gradient
diminishes downwards, and that the level at which such
temperatures exist lies rather deeper than this. It is,
doubtless, somewhat variable according to local conditions; nor
can we at all approximate closely to an estimate of the depth at
which the fusion temperatures will be reached, for, in fact, the
existence of the radioactive layer very much complicates our
estimates. In what follows we assume the depth of softening to
lie at about 40 kilometres beneath the surface of the normal
crust; that is 25 miles down. It is to be observed that Prestwich
and other eminent geologists, from a study of the facts of
crust-folding, etc., have arrived at similar estimates.[1] As a
further assumption we are probably not far wrong if we assign to
the radioactive part of this crust a thickness of about 10 or 12
kilometres; _i.e._ six or seven miles. This is necessarily a
rough approximation only; but the conclusions at which

[1] Prestwich, _Proc. Royal Soc._, xii., p. 158 _et seq._

128

we shall arrive are reached in their essential features allowing
a wide latitude in our choice of data. We shall speak of this
part of the crust as the normal radioactive layer.

An important fact is evolved from the mathematical investigation
of the temperature conditions arising from the presence of such a
radioactive layer. It is found that the greatest temperature, due
to the radioactive heat everywhere evolved in the layer--_i.e._
the temperature at its base--is proportional to the square of the
thickness of the layer. This fact has a direct bearing on the
influence of radioactivity upon mountain elevation; as we shall
now find.

The normal radioactive layer of the Earth is composed of rocks
extending--as we assume--approximately to a depth of 12 kilometres
(7.5 miles). The temperature at the base of this layer due to the
heat being continually evolved in it, is, say, t1°. Now, let us
suppose, in the trough of the geosyncline, and upon the top of
the normal layer, a deposit of, say, 10 kilometres (6.2 miles) of
sediments is formed during a long period of continental
denudation. What is the effect of this on the temperature at the
base of the normal layer depressed beneath this load? The total
thickness of radioactive rocks is now 22 kilometres. Accordingly
we find the new temperature t2°, by the proportion t1° : t2° ::
12° : 22° That is, as 144 to 484. In fact, the temperature is more
than trebled. It is true we here assume the radioactivity of the
sediments

129

and of the normal crust to be the same. The sediments are,
however, less radioactive in the proportion of 4 to 3.
Nevertheless the effects of the increased thickness will be
considerable.

Now this remarkable increase in the temperature arises entirely
from the condition attending the radioactive heating; and
involves something _additional_ to the temperature conditions
determined by the mere depression and thickening of the crust as
in the Babbage-Herschel theory. The latter theory only involves a
_shifting_ of the temperature levels (or geotherms) into the
deposited materials. The radioactive theory involves an actual
rise in the temperature at any distance from the surface; so that
_the level in the crust at which the rocks are softened is nearer
to the surface in the geosynclines than it is elsewhere in the
normal crust_ (Pl. XV, p. 118).

In this manner the rigid part of the crust is reduced in
thickness where the great sedimentary deposits have collected. A
ten-kilometre layer of sediment might result in reducing the
effective thickness of the crust by 30 per cent.; a
fourteen-kilometre layer might reduce it by nearly 50 per cent.
Even a four-kilometre deposit might reduce the effective
resistance of the crust to compressive forces, by 10 per cent.

Such results are, of course, approximate only. They show that as
the sediments grow in depth there is a rising of the geotherm of
plasticity--whatever its true temperature may be--gradually
reducing the thickness of that part

130

of the upper crust which is bearing the simultaneously increasing
compressive stresses. Below this geotherm long-continued stress
resolves itself into hydrostatic pressure; above it (there is, of
course, no sharp line of demarcation) the crust accumulates
elastic energy. The final yielding and flexure occur when the
resistant cross-section has been sufficiently diminished. It is
probable that there is also some outward hydrostaitic thrust over
the area of rising temperature, which assists in determining the
upward throw of the folds.

When yielding has begun in any geosyncline, and the materials are
faulted and overthrust, there results a considerably increased
thickness. As an instance, consider the piling up of sediments
over the existing materials of the Alps, which resulted from the
compressive force acting from south to north in the progress of
Alpine upheaval. Schmidt of Basel has estimated that from 15 to
20 kilometres of rock covered the materials of the Simplon as now
exposed, at the time when the orogenic forces were actively at
work folding and shearing the beds, and injecting into their
folds the plastic gneisses coming from beneath.[1] The lateral
compression of the area of deposition of the Laramide, already
referred to, resulted in a great thickening of the deposits. Many
other cases might be cited; the effect is always in some degree
necessarily produced.

[1] Schmidt, Ec. Geol. _Helvelix_, vol. ix., No. 4, p. 590

131

If time be given for the heat to accumulate in the lower depths
of the crushed-up sediments, here is an additional source of
increased temperature. The piled-up masses of the Simplon might
have occasioned a rise due to radioactive heating of one or two
hundred degrees, or even more; and if this be added to the
interior heat, a total of from 800° to 1000° might have prevailed
in the rocks now exposed at the surface of the mountain. Even a
lesser temperature, accompanied by the intense pressure
conditions, might well occasion the appearances of thermal
metamorphism described by Weinschenk, and for which, otherwise,
there is difficulty in accounting.[1]

This increase upon the primarily developed temperature conditions
takes place concurrently with the progress of compression; and
while it cannot be taken into account in estimating the
conditions of initial yielding of the crust, it adds an element
of instability, inasmuch as any progressive thickening by lateral
compression results in an accelerated rise of the goetherms. It
is probable that time sufficient for these effects to develop, if
not to their final, yet to a considerable extent, is often
available. The viscous movements of siliceous materials, and the
out-pouring of igneous rocks which often attend mountain
elevation, would find an explanation in such temperatures.

[1] Weinschenk, _Congrès Géol. Internat._, 1900, i., p. 332.

132

There is no more striking feature of the part here played by
radioactivity than the fact that the rhythmic occurrence of
depression and upheaval succeeding each other after great
intervals of time, and often shifting their position but little
from the first scene of sedimentation, becomes accounted for. The
source of thermal energy, as we have already remarked, is in fact
transported with the sediments--that energy which determines the
place of yielding and upheaval, and ordains that the mountain
ranges shall stand around the continental borders. Sedimentation
from this point of view is a convection of energy.

When the consolidated sediments are by these and by succeeding
movements forced upwards into mountains, they are exposed to
denudative effects greatly exceeding those which affect the
plains. Witness the removal during late Tertiary times of the
vast thickness of rock enveloping the Alps. Such great masses are
hurried away by ice, rivers, and rain. The ocean receives them;
and with infinite patience the world awaits the slow accumulation
of the radioactive energy beginning afresh upon its work. The
time for such events appears to us immense, for millions of years
are required for the sediments to grow in thickness, and the
geotherms to move upwards; but vast as it is, it is but a moment
in the life of the parent radioactive substances, whose atoms,
hardly diminished in numbers, pursue their changes while the
mountains come and go, and the

133

rudiments of life develop into its highest consummations.

To those unacquainted with the results of geological
investigation the history of the mountains as deciphered in the
rocks seems almost incredible.

The recently published sections of the Himalaya, due to H. H.
Hayden and the many distinguished men who have contributed to the
Geological Survey of India, show these great ranges to be
essentially formed of folded sediments penetrated by vast masses
of granite and other eruptives. Their geological history may be
summarised as follows

The Himalayan area in pre-Cambrian times was, in its southwestern
extension, part of the floor of a sea which covered much of what
is now the Indian Peninsula. In the northern shallows of this sea
were laid down beds of conglomerate, shale, sandstone and
limestone, derived from the denudation of Archæan rocks, which,
probably, rose as hills or mountains in parts of Peninsular India
and along the Tibetan edge of the Himalayan region. These beds
constitute the record of the long Purana Era[1] and are probably
coeval with the Algonkian of North America. Even in these early
times volcanic disturbances affected this area and the lower beds
of the Purana deposits were penetrated by volcanic outflows,
covered by sheets of lava, uplifted, denuded and again submerged

[1] See footnote, p. 139.

134

beneath the waters. Two such periods of instability have left
their records in the sediments of the Purana sea.

The succeeding era--the Dravidian Era--opens with Haimanta
(Cambrian) times. A shallow sea now extended over Kumaun, Garwal,
and Spiti, as well as Kashmir and ultimately over the Salt Range
region of the Punjab as is shown by deposits in these areas. This
sea was not, however, connected with the Cambrian sea of Europe.
The fossil faunas left by the two seas are distinct.

After an interval of disturbance during closing Haimanta times,
geographical changes attendant on further land movements
occurred. The central sea of Asia, the Tethys, extended westwards
and now joined with the European Paleozoic sea; and deposits of
Ordovician and Silurian age were laid down:--the Muth deposits.

The succeeding Devonian Period saw the whole Northern Himalayan
area under the waters of the Tethys which, eastward, extended to
Burma and China and, westward, covered Kashmir, the Hindu Kush
and part of Afghanistan. Deposits continued to be formed in this
area till middle Carboniferous times.

Near. the close of the Dravidian Era Kashmir became convulsed by
volcanic disturbance and the Penjal traps were ejected. It was a
time of worldwide disturbance and of redistribution of land and
water. Carboniferous times had begun, and the geographical
changes in

135

the southern limits of the Tethys are regarded as ushering in a
new and last era in Indian geological history the Aryan Bra.

India was now part of Gondwanaland; that vanished continent which
then reached westward to South Africa and eastward to Australia.
A boulder-bed of glacial origin, the Talchir Boulder-bed, occurs
in many surviving parts of this great land. It enters largely
into the Salt Range deposits. There is evidence that extensive
sheets of ice, wearing down the rocks of Rajputana, shoved their
moraines northward into the Salt Range Sea; then, probably, a
southern extension of the Tethys.

Subsequent to this ice age the Indian coalfields of the Gondwana
were laid down, with beds rich in the Glossopteris and
Gangamopteris flora. This remarkable carboniferous flora extends
to Southern Kashmir, so that it is to be inferred that this
region was also part of the main Gondwanaland. But its emergence
was but for a brief period. Upper Carboniferous marine deposits
succeeded; and, in fact, there was no important discontinuity in
the deposits in this area from Panjal times till the early
Tertiaries. During the whole of which vast period Kashmir was
covered with the waters of the Tethys.

The closing Dravidian disturbances of the Kashmir region did not,
apparently, extend to the eastern Himalayan area. But the
Carboniferous Period was, in this

136

eastern area, one of instability, culminating, at the close of
the Period, in a steady rise of the land and a northward retreat
of the Tethys. Nearly the entire Himalaya east of Kashmir became
a land surface and remained exposed to denudative forces for so
long a time that in places the whole of the Carboniferous,
Devonian, and a large part of the Silurian and Ordovician
deposits were removed--some thousands of feet in thickness--before
resubmergence in the Tethys occurred.

Towards the end of the Palaeozoic Age the Aryan Tethys receded
westwards, but still covered the Himalaya and was still connected
with the European Palæozoic sea. The Himalayan area (as well as
Kashmir) remained submerged in its waters throughout the entire
Mesozoic Age.

During Cretaceous times the Tethys became greatly extended,
indicating a considerable subsidence of northwestern India,
Afghanistan, Western Asia, and, probably, much of Tibet. The
shallow-water character of the deposits of the Tibetan Himalaya
indicates, however, a coast line near this region. Volcanic
materials, now poured out, foreshadow the incoming of the great
mountain-building epoch of the Tertiary Era. The enormous mass of
the Deccan traps, possessing a volume which has been estimated at
as much as 6,000 cubic miles, was probably extruded over the
Northern Peninsular region during late Cretaceous times. The sea
now began to retreat, and by the close of

137

the Eocene, it had moved westward to Sind and Baluchistan. The
movements of the Earth's crust were attended by intense volcanic
activity, and great volumes of granite were injected into the
sediments, followed by dykes and outflows of basic lavas.

The Tethys vanished to return no more. It survives in the
Mediterranean of today. The mountain-building movements continued
into Pliocene times. The Nummulite beds of the Eocene were, as
the result, ultimately uplifted 18,500 feet over sea level, a
total uplift of not less than 20,000 feet.

Thus with many vicissitudes, involving intervals of volcanic
activity, local uplifting, and extensive local denudation, the
Himalaya, which had originated in the sediments of the ancient
Purana sea, far back in pre-Cambrian times, and which had
developed potentially in a long sequence of deposits collecting
almost continuously throughout the whole of geological time,
finally took their place high in the heavens, where only the
winds--faint at such altitudes--and the lights of heaven can visit
their eternal snows.[1]

In this great history it is significant that the longest
continuous series of sedimentary deposits which the world has
known has become transfigured into the loftiest elevation upon
its surface.

[1] See A Sketch of the _Geography and Geology of the Himalaya
Mountains and Tibet_. By Colonel S. G. Burrard, R.E., F.R.S., and
H. H. Hayden, F.G.S., Part IV. Calcutta, 1908.

138

The diagrammatic sections of the Himalaya accompanying this brief
description arc taken from the monograph of Burrard and Hayden
(loc. cit.) on the Himalaya. Looking at the sections we see that
some of the loftiest summits are sculptured in granite and other
crystalline rocks. The appearance of these materials at the
surface indicates the removal by denudation and the extreme
metamorphism of much sedimentary deposit. The crystalline rocks,
indeed, penetrate some of the oldest rocks in the world. They
appear in contact with Archaean, Algonkian or early Palaeozoic
rocks. A study of the sections reveals not only the severe earth
movements, but also the immense amount of sedimentary deposits
involved in the genesis of these alps. It will be noted that the
vertical scale is not exaggerated relatively to the
horizontal.[1] Although there is no evidence of mountain
building

[1] To those unacquainted with the terminology of Indian geology
the following list of approximate equivalents in time will be of
use

Ngari Khorsum Beds - Pleistocene.
Siwalik Series -     Miocene and Pliocene.
Sirmur Series -      Oligocene.
Kampa System -       Eocene and Cretaceous.
Lilang System -      Triassic.
Kuling System -      Permian.
Gondwana System -    Carboniferous.
Kenawar System -     Carboniferous and Devonian
Muth System -        Silurian.
Haimanta System -    Mid. and Lower Cambrian.
Purana Group -       Algonkian.
Vaikrita System -    Archæan.
Daling Series -      Archæan.

139

on a large scale in the Himalayan area till the Tertiary
upheaval, it is, in the majority of cases, literally correct to
speak of the mountains as having their generations like organic
beings, and passing through all the stages of birth, life, death
and reproduction. The Alps, the Jura, the Pyrenees, the Andes,
have been remade more than once in the course of geological time,
the _débris_ of a worn-out range being again uplifted in succeeding
ages.

Thus to dwell for a moment on one case only: that of the
Pyrenees. The Pyrenees arose as a range of older Palmozoic rocks
in Devonian times. These early mountains, however, were
sufficiently worn out and depressed by Carboniferous times to
receive the deposits of that age laid down on the up-turned edges
of the older rocks. And to Carboniferous succeeded Permian,
Triassic, Jurassic and Lower Cretaceous sediments all laid down
in conformable sequence. There was then fresh disturbance and
upheaval followed by denudation, and these mountains, in turn,
became worn out and depressed beneath the ocean so that Upper
Greensand rocks were laid down unconforrnably on all beneath. To
these now succeeded Upper Chalk, sediments of Danian age, and so
on, till Eocene times, when the tale was completed and the
existing ranges rose from the sea. Today we find the folded
Nummulitic strata of Eocene age uplifted 11,000 feet, or within
200 feet of the greatest heights of the Pyrenees. And so they
stand awaiting

140

the time when once again they shall "fall into the portion of
outworn faces."[1]

Only mountains can beget mountains. Great accumulations of
sediment are a necessary condition for the localisation of
crust-flexure. The earliest mountains arose as purely igneous or
volcanic elevations, but the generations of the hills soon
originated in the collection of the _débris_, under the law of
gravity, in the hollow places. And if a foundered range is
exposed now to our view encumbered with thousands of feet of
overlying sediments we know that while the one range was sinking,
another, from which the sediments were derived, surely existed.
Through the "windows" in the deep-cut rocks of the Swiss valleys
we see the older Carboniferous Alps looking out, revisiting the
sun light, after scores of millions of years of imprisonment. We
know that just as surely as the Alps of today are founding by
their muddy torrents ranges yet to arise, so other primeval Alps
fed into the ocean the materials of these buried pre-Permian
rocks.

This succession of events only can cease when the rocks have been
sufficiently impoverished of the heat-producing substances, or
the forces of compression shall have died out in the surface
crust of the earth.

It seems impossible to escape the conclusion that in the great
development of ocean-encircling areas of

[1] See Prestwich, _Chemical and Physical Geology_, p. 302.

141

deposition and crustal folding, the heat of radioactivity has
been a determining factor. We recognise in the movements of the
sediments not only an influence localising and accelerating
crustal movements, but one which, in subservience to the primal
distribution of land and water, has determined some of the
greatest geographical features of the globe.

It is no more than a step to show that bound up with the
radioactive energy are most of the earthquake and volcanic
phenomena of the earth. The association of earthquakes with the
great geosynclines is well known. The work of De Montessus showed
that over 94 per cent. of all recorded shocks lie in the
geosynclinal belts. There can be no doubt that these
manifestations of instability are the results of the local
weakness and flexure which originated in the accumulation of
energy denuded from the continents. Similarly we may view in
volcanoes phenomena referable to the same fundamental cause. The
volcano was, in fact, long regarded as more intimately connected
with earthquakes than it, probably, actually is; the association
being regarded in a causative light, whereas the connexion is
more that of possessing a common origin. The girdle of volcanoes
around the Pacific and the earthquake belt coincide. Again, the
ancient and modern volcanoes and earthquakes of Europe are
associated with the geosyncline of the greater Mediterranean, the
Tethys of Mesozoic times. There is no difficulty in understanding
in a

142

general way the nature of the association. The earthquake is the
manifestation of rupture and slip, and, as Suess has shown, the
epicentres shift along that fault line where the crust has
yielded.[1] The volcano marks the spot where the zone of fusion
is brought so high in the fractured crust that the melted
materials are poured out upon the surface.

In a recent work on the subject of earthquakes Professor Hobbs
writes: "One of the most interesting of the generalisations which
De Montessus has reached as a result of his protracted studies,
is that the earthquake districts on the land correspond almost
exactly to those belts upon the globe which were the almost
continuous ocean basins of the long Secondary era of geological
history. Within these belts the sedimentary formations of the
crust were laid down in the greatest thickness, and the
formations follow each other in relatively complete succession.
For almost or quite the whole of this long era it is therefore
clear that the ocean covered these zones. About them the
formations are found interrupted, and the lacuna indicate that
the sea invaded the area only to recede from it, and again at
some later period to transgress upon it. For a long time,
therefore, these earthquake belts were the sea basins--the
geosynclines. They became later the rising mountains of the
Tertiary period, and mountains they

[1] Suess, _The Face of the Earth_, vol. ii., chap. ii.

143

are today. The earthquake belts are hence those portions of the
earth's crust which in recent times have suffered the greatest
movements in a vertical direction--they are the most mobile
portions of the earth's crust."[1] Whether the movements
attending mountain elevation and denudation are a connected and
integral part of those wide geographical changes which result in
submergence and elevation of large continental areas, is an
obscure and complex question. We seem, indeed, according to the
views of some authorities, hardly in a position to affirm with
certainty that such widespread movements of the land have
actually occurred, and that the phenomena are not the outcome of
fluctuations of oceanic level; that our observations go no
further than the recognition of positive and negative movements
of the strand. However this may be, the greater part of
mechanical denudation during geological time has been done on the
mountain ranges. It is, in short, indisputable that the orogenic
movements which uplift the hills have been at the basis of
geological history. To them the great accumulations of sediments
which now form so large a part of continental land are mainly
due. There can be no doubt of the fact that these movements have
swayed the entire history, both inorganic and organic, of the
world in which we live.

[1] Hobbs, _Earthquakes_, p. 58.

144

To sum the contents of this essay in the most general terms, we
find that in the conception of denudation as producing the
convection and accumulation of radiothermal energy the surface
features of the globe receive a new significance. The heat of the
earth is not internal only, but rather a heat-source exists at
the surface, which, as we have seen, cannot prevail to the same
degree within; and when the conditions become favourable for the
aggregation of the energy, the crust, heated both from beneath
and from above, assumes properties more akin to those of its
earlier stages of development, the secular heat-loss being
restored in the radioactive supplies. These causes of local
mobility have been in operation, shifting somewhat from place to
place, and defined geographically by the continental masses
undergoing denudation, since the earliest times.

145

ALPINE STRUCTURE

AN intelligent observer of the geological changes progressing in
southern Europe in Eocene times would have seen little to inspire
him with a premonition of the events then developing. The
Nummulitic limestones were being laid down in that enlarged
Mediterranean which at this period, save for a few islands,
covered most of south Europe. Of these stratified remains, as
well as of the great beds of Cretaceous, Jurassic, Triassic, and
Permian sediments beneath, our hypothetical observer would
probably have been regardless; just as today we observe, with an
indifference born of our transitoriness, the deposits rapidly
gathering wherever river discharge is distributing the sediments
over the sea-floor, or the lime-secreting organisms are actively
at work. And yet it took but a few millions of years to uplift
the deposits of the ancient Tethys; pile high its sediments in
fold upon fold in the Alps, the Carpathians, and the Himalayas;
and--exposing them to the rigours of denudation at altitudes where
glaciation, landslip, and torrent prevail--inaugurate a new epoch
of sedimentation and upheaval.

146

In the case of the Alps, to which we wish now specially to refer,
the chief upheaval appears to have been in Oligocene times,
although movement continued to the close of the Pliocene. There
was thus a period of some millions of years within which the
entire phenomena were comprised. Availing ourselves of Sollas'
computations,[1] we may sum the maximum depths of sedimentary
deposits of the geological periods concerned as follows:--

Pliocene - - - - - 3,950 m.

Miocene  - - - - - 4,250 m.

Oligocene  - - - - 3,660 m.

Eocene - - - - - - 6,100 m.

and assuming that the orogenic forces began their work in the
last quarter of the Eocene period, we have a total of 13,400 m.
as some measure of the time which elapsed. At the rate of io
centimetres in a century these deposits could not have collected
in less than 13.4 millions of years. It would appear that not
less than some ten millions of years were consumed in the genesis
of the Alps before constructive movements finally ceased.

The progress of the earth-movements was attended by the usual
volcanic phenomena. The Oligocene and Miocene volcanoes extended
in a band marked by the Auvergne, the Eiffel, the Bohemian, and
the eastern Carpathian eruptions; and, later, towards the close
of the movements in Pliocene times, the south border

[1] Sollas, Anniversary Address, Geol. Soc., London, 1909.

147

regions of the Alps became the scene of eruptions such as those
of Etna, Santorin, Somma (Vesuvius), etc.

We have referred to these well-known episodes with two objects in
view: to recall to mind the time-interval involved, and the
evidence of intense crustal disturbance, both dynamic and
thermal. According to views explained in a previous essay, the
energetic effects of radium in the sediments and upper crust were
a principal factor in localising and bringing about these
results. We propose now to inquire if, also, in the more intimate
structure of the Alps, the radioactive energy may not have borne
a part.

What we see today in the Alps is but a residue spared by
denudation. It is certain that vast thicknesses of material have
disappeared. Even while constructive effects were still in
progress, denudative forces were not idle. Of this fact the
shingle accumulations of the Molasse, where, on the northern
borders of the Alps, they stand piled into mountains, bear
eloquent testimony. In the sub-Apennine series of Italy, the
great beds of clays, marls, and limestones afford evidence of
these destructive processes continued into Pliocene times. We
have already referred to Schmidt's estimate that the sedimentary
covering must have in places amounted to from 15,000 to 20,000
metres. The evidence for this is mainly tectonic or structural;
but is partly forthcoming in the changes which the materials now
open to our inspection plainly reveal. Thus it is impos-

148

sible to suppose that gneissic rocks can become so far plastic as
to flow in and around the calcareous sediments, or be penetrated
by the latter--as we see in the Jungfrau and elsewhere--unless
great pressures and high temperatures prevailed. And, according
to some writers, the temperatures revealed by the intimate
structural changes of rock-forming minerals must have amounted to
those of fusion. The existence of such conditions is supported by
the observation that where the.crystallisation is now the most
perfect, the phenomena of folding and injection are best
developed.[1] These high temperatures would appear to be
unaccountable without the intervention of radiothermal effects;
and, indeed, have been regarded as enigmatic by observers of the
phenomena in question. A covering of 20,000 metres in thickness
would not occasion an earth-temperature exceeding 500° C. if the
gradients were such as obtain in mountain regions generally; and
600° is about the limit we could ascribe to the purely passive
effects of such a layer in elevating the geotherms.

Those who are still unacquainted with the recently published
observations on the structure of the Alps may find it difficult
to enter into what has now to be stated; for the facts are,
indeed, very different from the generally preconceived ideas of
mountain formation. Nor can we wonder that many geologists for
long held

[1] Weinschenk, C. R. _Congrès Géol._, 1900, p. 321, et seq.

149

back from admitting views which appeared so extreme. Receptivity
is the first virtue of the scientific mind; but, with every
desire to lay aside prejudice, many felt unequal to the
acceptance of structural features involving a folding of the
earth-crust in laps which lay for scores of miles from country to
country, and the carriage of mountainous materials from the south
of the Alps to the north, leaving them finally as Alpine ranges
of ancient sediments reposing on foundations of more recent date.
The historian of the subject will have to relate how some who
finally were most active in advancing the new views were at first
opposed to them. In the change of conviction of these eminent
geologists we have the strongest proof of the convincing nature
of the observations and the reality of the tectonic features upon
which the recent views are founded.

The lesser mountains which stand along the northern border of the
great limestone Alps, those known as the Préalpes, present the
strange characteristic of resting upon materials younger than
themselves. Such mountains as the remarkable-looking Mythen, near
Schwyz, for instance, are weathered from masses of Triassic and
Jurassic rock, and repose on the much more recent Flysch. In
sharp contrast to the Flysch scenery, they stand as abrupt and
gigantic erratics, which have been transported from the central
zone of the Alps lying far to the south. They are strangers
petrologically,

150

stratigraphically, and geographically,[1] to the locality in
which they now occur. The exotic materials may be dolomites,
limestones, schists, sandstones, or rocks of igneous origin. They
show in every case traces of the severe dynamic actions to which
they have been subjected in transit. The igneous, like the
sedimentary, klippen, can be traced to distant sources; to the
massif of Belladonne, to Mont Blanc, Lugano, and the Tyrol. The
Préalpes are, in fact, mountains without local roots.

In this last-named essential feature, the Préalpes do not differ
from the still greater limestone Alps which succeed them to the
south. These giants, _e.g._ the Jungfrau, Wetterhorn, Eiger, etc.,
are also without local foundations. They have been formed from
the overthrown and drawn-out anticlines of great crust-folds,
whose synclines or roots are traceable to the south side of the
Rhone Valley. The Bernese Oberland originated in the piling-up of
four great sheets or recumbent folds, one of which is continued
into the Préalpes. With Lugeon[2] we may see in the phenomenon of
the formation of the Préalpes a detail; regarding it as a normal
expression of that mechanism which has created the Swiss Alps.
For these limestone masses of the Oberland are not indications of
a merely local shift of the sedimentary covering of the Alps.
Almost the whole covering has

[1] De Lapparent, _Traité de Géologie_, p. 1,785.

[2] Lugeon, _Bulletin Soc. Géol. de France_, 1901, p. 772.

151

been pushed over and piled up to the north. Lugeon[l] concludes
that, before denudation had done its work and cut off the
Préalpes from their roots, there would have been found sheets, to
the number of eight, superimposed and extending between the Mont
Blanc massif and the massif of the Finsteraarhorn: these sheets
being the overthrown folds of the wrinkled sedimentary covering.
The general nature of the alpine structure

{Fig. 8}

will be understood from the presentation of it diagrammatically
after Schmidt of Basel (Fig. 8).[2] The section extends from
north to south, and brings out the relations of the several
recumbent folds. We must imagine almost the whole of these
superimposed folds now removed from the central regions of the
Alps by denudation,

[1] Lugeon, _loc. cit._

[2] Schmidt, _Ec. Geol. Helvetiae_, vol. ix., No. 4.

152

and leaving the underlying gneisses rising through the remains of
Permian, Triassic, and Jurassic sediments; while to the north the
great limestone mountains and further north still, the Préalpes,
carved from the remains of the recumbent folds, now stand with
almost as little resemblance to the vanished mountains as the
memories of the past have to its former intense reality.

These views as to the origin of the Alps, which are shared at the
present day by so many distinguished geologists, had their origin
in the labours of many now gone; dating back to Studer; finding
their inspiration in the work of Heim, Suess, and Marcel
Bertrand; and their consummation in that of Lugeon, Schardt,
Rothpletz, Schmidt, and many others. Nor must it be forgotten
that nearer home, somewhat similar phenomena, necessarily on a
smaller scale, were recognised by Lapworth, twenty-six years ago,
in his work on the structure of the Scottish Highlands.

An important tectonic principle underlies the development of the
phenomena we have just been reviewing. The uppermost of the
superimposed recumbent folds is more extended in its development
than those which lie beneath. Passing downwards from the highest
of the folds, they are found to be less and less extended both in
the northerly and in the southerly direction, speaking of the
special case--the Alps--now before us. This feature might be
described somewhat differently. We might say that those folds
which had their roots farther

153

to the south were the most drawn-out towards the north: or again
we might say that the synclinal or deep-seated part of the fold
has lagged behind the anticlinal or what was originally the
highest part of the fold, in the advance of the latter to the
north. The anticline has advanced relatively to the syncline. To
this law one exception only is observed in the Swiss Alps; the
sheet of the Brèche (_Byecciendecke_) falls short, in its northerly
extension, of the underlying fold, which extends to form the
Préalpes.

Contemplating such a generalised section as Professor Schmidt's,
or, indeed, more particular sections, such as those in the Mont
Blanc Massif by Marcel Bertrand,[1] of the Dent de Morcles,
Diablerets, Wildhorn, and Massif de la Brèche by Lugeon,[2] or
finally Termier's section of the Pelvoux Massif,[3] one is
reminded of the breaking of waves on a sloping beach. The wave,
retarded at its base, is carried forward above by its momentum,
and finally spreads far up on the strand; and if it could there
remain, the succeeding wave must necessarily find itself
superimposed upon the first. But no effects of inertia, no
kinetic effects, may be called to our aid in explaining the
formation of mountains. Some geologists have accordingly supposed
that in order to account for

[1] Marcel Bertrand, _Cong. Géol. Internat._, 1900, Guide Géol.,
xiii. a, p. 41.

[2] Lugeon, _loc. cit._, p. 773.

[3] De Lapparent, _Traite de Géol._, p. 1,773.

154

the recumbent folds and the peculiar phenomena of increasing
overlap, or _déferlement_, an obstacle, fixed and deep-seated, must
have arrested the roots or synclines of the folds, and held them
against translational motion, while a movement of the upper crust
drew out and carried forward the anticlines. Others have
contented themselves by recording the facts without advancing any
explanatory hypothesis beyond that embodied in the incontestable
statement that such phenomena must be referred to the effects of
tangential forces acting in the Earth's crust.

It would appear that the explanation of the phenomena of
recumbent folds and their _déferlement_ is to be obtained directly
from the temperature conditions prevailing throughout the
stressed pile of rocks; and here the subject of mountain
tectonics touches that with which we were elsewhere specially
concerned--the geological influence of accumulated radioactive
energy.

As already shown[1], a rise of temperature due to this source of
several hundred degrees might be added to such temperatures as
would arise from the mere blanketing of the Earth, and the
consequent upward movement of the geotherms. The time element is
here the most important consideration. The whole sequence of
events from the first orogenic movements to the final upheaval in
Pliocene times must probably have occupied not less than ten
million years.

[1] _Mountain Genesis_, p. 129, et seq.

155

Unfortunately the full investigation of the distribution of
temperature after any given time is beset with difficulties; the
conditions being extremely complex. If the radioactive heating
was strictly adiabatic--that is, if all the heat was conserved and
none entered from without--the time required for the attainment of
the equilibrium radioactive temperature would be just about six
million years. The conditions are not, indeed, adiabatic; but, on
the other hand, the rocks upraised by lateral pressure were by no
means at 0° C. to start with. They must be assumed to have
possessed such temperatures as the prior radiothermal effects,
and the conducted heat from the Earth's interior, may have
established.

It would from this appear probable that if a duration of ten
million years was involved, the equilibrium radioactive
temperatures must nearly have been attained. The effects of heat
conducted from the underlying earthcrust have to be added,
leading to a further rise in temperature of not less than 500° or
600° . In such considerations the observed indications of high
temperatures in materials now laid bare by denudation, probably
find their explanation (P1. XIX).

The first fact that we infer from the former existence of such a
temperature distribution is the improbability, indeed the
impossibility, that anything resembling a rigid obstacle, or
deep-seated "horst," can have existed beneath the present
surface-level, and opposed the northerly movement of the
deep-lying synclines. For

156

such a horst can only have been constituted of some siliceous
rock-material such as we find everywhere rising through the
worn-down sediments of the Alps; and the idea that this could
retain rigidity under the prevailing temperature conditions, must
be dismissed. There is no need to labour this question; the horst
cannot have existed. To what, then, is the retardation of the
lower parts of the folds, their overthrow, above, to the north,
and their _déferlement_, to be ascribed?

A little consideration shows that the very conditions of high
temperature and viscosity, which render untenable the hypothesis
of a rigid obstacle, suffice to afford a full explanation of the
retardation of the roots of the folds. For directed translatory
movements cannot be transmitted through a fluid, pressure in
which is necessarily hydrostatic, and must be exerted equally in
every direction. And this applies, not only to a fluid, but to a
body which will yield viscously to an impressed force. There will
be a gradation, according as viscosity gives place to rigidity,
between the states in which the applied force resolves itself
into a purely hydrostatic pressure, and in which it is
transmitted through the material as a directed thrust. The nature
of the force, in the most general case, of course, has to be
considered; whether it is suddenly applied and of brief duration,
or steady and long-continued. The latter conditions alone apply
to the present case.

It follows from this that, although a tangential force

157

or pressure be engendered by a crustal movement occurring to the
south, and the resultant effects be transmitted northwards, these
stresses can only mechanically affect the rigid parts of the
crust into which they are carried. That is to say, they may
result in folding and crushing, or horizontally transporting, the
upper layers of the Earth's crust; but in the deeper-lying
viscous materials they must be resolved into hydrostatic pressure
which may act to upheave the overlying covering, but must refuse
to transmit the horizontal translatory movements affecting the
rigid materials above.

Between the regions in which these two opposing conditions
prevail there will be no hard and fast line; but with the
downward increase of fluidity there will be a gradual failure of
the mechanical conditions and an increase of the hydrostatic.
Thus while the uppermost layers of the crust may be transported
to the full amount of the crustal displacement acting from the
south (speaking still of the Alps) deeper down there will be a
lesser horizontal movement, and still deeper there is no
influence to urge the viscous rock-materials in a northerly
direction. The consequences of these conditions must be the
recumbence of the folds formed under the crust-stress, and their
_déferlement_ towards the north. To see this, we must follow the
several stages of development.

The earliest movements, we may suppose, result in flexures of the
Jura-Mountain type--that is, in a

158

succession of undulations more or less symmetrical. As the
orogenic force continues and develops, these undulations give
place to folds, the limbs of which are approximately vertical,
and the synclinal parts of which become ever more and more
depressed into the deeper, and necessarily hotter, underlying
materials; the anticlines being probably correspondingly
elevated. These events are slowly developed, and the temperature
beneath is steadily rising in consequence of the conducted
interior heat, and the steady accumulation of radioactive energy
in the sedimentary rocks and in the buried radioactive layer of
the Earth. The work expended on the crushed and sheared rock also
contributes to the developing temperature. Thus the geotherms
must move upwards, and the viscous conditions extend from below;
continually diminishing the downward range of the translatory
movements progressing in the higher parts. While above the folded
sediments are being carried northward, beneath they are becoming
anchored in the growing viscosity of the medium. The anticlines
will bend over, and the most southerly of the folds will
gradually become pushed or bent over those lying to the north.
Finally, the whole upper part of the sheaf will become
horizontally recumbent; and as the uppermost folds will be those
experiencing the greatest effects of the continued displacement,
the _déferlement_ or overlap must necessarily arise.

We may follow these stages of mountain evolution

159

in a diagram (Fig. 9) in which we eliminate intermediate
conditions, and regard the early and final stages of development
only. In the upper sketch we suppose the lateral compression much
developed and the upward movement of the geotherms in progress.
The dotted line may be assumed to be a geotherm having a
temperature of viscosity. If the conditions here shown persist

{Fig. 9}

indefinitely, there is no doubt that the only further
developments possible are the continued crushing of the sediments
and the bodily displacement of the whole mass to the north. The
second figure is intended to show in what manner these results
are evaded. The geotherm of viscosity has risen. All above it is
affected mechanically by the continuing stress, and borne
northwards in varying

160

degree depending upon the rigidity. The folds have been
overthrown and drawn out; those which lay originally most to the
south have become the uppermost; and, experiencing the maximum
amount of displacement, overlap those lying beneath. There has
also been a certain amount of upthrow owing to the hydrostatic
pressure. This last-mentioned element of the phenomena is of
highly indeterminate character, for we know not the limits to
which the hydrostatic pressure may be transmitted, and where it
may most readily find relief. While, according to some of the
published sections, the uplifting force would seem to have
influenced the final results of the orogenic movements, a
discussion of its effects would not be profitable.

161

OTHER MINDS THAN OURS?

IN the year 1610 Galileo, looking through his telescope then
newly perfected by his own hands, discovered that the planet
Jupiter was attended by a train of tiny stars which went round
and round him just as the moon goes round the Earth.

It was a revelation too great to be credited by mankind. It was
opposed to the doctrine of the centrality of the Earth, for it
suggested that other worlds constituted like ours might exist in
the heavens.

Some said it was a mere optic illusion; others that he who looked
through such a tube did it at the peril of his soul--it was but a
delusion of Satan. Galileo converted a few of the unbelievers who
had the courage to look through his telescope. To the others he
said, he hoped they would see those moons on their way to heaven.
Old as this story is it has never lost its pathos or its
teaching.

The spirit which assailed Galileo's discoveries and which finally
was potent to overshadow his declining years, closed in former
days the mouths of those who asked the question written at the
head of this lecture: "Are we to believe that there are other
minds than ours?"

162

Today we consider the question in a very different spirit. Few
would regard it as either foolish or improper. Its intense
interest would be admitted by all, and but for the limitations
closing our way on every side it would, doubtless, attract the
most earnest investigation. Even on the mere balance of judgment
between the probable and the improbable, we have little to go on.
We know nothing definitely as to the conditions under which life
may originate: whether these are such as to be rare almost to
impossibility, or common almost to certainty. Only within narrow
limits of temperature and in presence of certain of the elements,
can life like ours exist, and outside these conditions life, if
such there be, must be different from ours. Once originated it is
so constituted as to assail the energies around it and to advance
from less to greater. Do we know more than these vague facts?
Yes, we have in our experience one other fact and one involving
much.

We know that our world is very old; that life has been for many
millions of years upon it; and that Man as a thinking being is
but of yesterday. Here is then a condition to be fulfilled. To
every world is physically assigned a limit to the period during
which it is habitable according to our knowledge of life and its
necessities. This limit passed and rationality missed, the chance
for that world is gone for ever, and other minds than ours
assuredly will not from it contemplate the universe. Looking at
our own world we see that the tree of life has,

163

indeed, branched, leaved and, possibly, budded many times; it
never bloomed but once.

All difficulties dissolve and speculations become needless under
one condition only: that in which rationality may be inferred
directly or indirectly by our observations on some sister world
in space, This is just the evidence which in recent years has
been claimed as derived from a study of the surface of Mars. To
that planet our hope of such evidence is restricted. Our survey
in all other directions is barred by insurmountable difficulties.
Unless some meteoric record reached our Earth, revelationary of
intelligence on a perished world, our only hope of obtaining such
evidence rests on the observation of Mars' surface features. To
this subject we confine our attention in what follows.

The observations made during recent years upon the surface
features of Mars have, excusably enough, given rise to
sensational reports. We must consider under what circumstances
these observations have been made.

Mars comes into particularly favourable conditions for
observation every fifteen years. It is true that every two years
and two months we overtake him in his orbit and he is then in
"opposition." That is, the Earth is between him and the sun: he
is therefore in the opposite part of the heavens to the sun. Now
Mars' orbit is very excentric, sometimes he is 139 million miles
from the sun, and sometimes he as as much as 154 million miles
from the sun. The Earth's orbit is, by comparison, almost

164

a circle. Evidently if we pass him when he is nearest to the sun
we see him at his best; not only because he is then nearest to
us, but because he is then also most brightly lit. In such
favourable oppositions we are within 35 million miles of him; if
Mars was in aphelion we would pass him at a distance of 61
million miles. Opposition occurs under the most favourable
circumstances every fifteen years. There was one in 1862, another
in 1877, one in 1892, and so on.

When Mars is 35 million miles off and we apply a telescope
magnifying 1,000 diameters, we see him as if placed 35,000 miles
off. This would be seven times nearer than we see the moon with
the naked eye. As Mars has a diameter about twice as great as
that of the moon, at such a distance he would look fourteen times
the diameter of the moon. Granting favourable conditions of
atmosphere much should be seen.

But these are just the conditions of atmosphere of which most of
the European observatories cannot boast. It is to the honour of
Schiaparelli, of Milan, that under comparatively unfavourable
conditions and with a small instrument, he so far outstripped his
contemporaries in the observation of the features of Mars that
those contemporaries received much of his early discoveries with
scepticism. Light and dark outlines and patches on the planet's
surface had indeed been mapped by others, and even a couple of
the canals sighted; but at the opposition of 1877 Schiaparelli
first mapped any considerable

165

number of the celebrated "canals" and showed that these
constituted an extraordinary and characteristic feature of the
planet's geography. He called them "canali," meaning thereby
"channels." It is remarkable indeed that a mistranslation appears
really responsible for the initiation of the idea that these
features are canals.

In 1882 Schiaparelli startled the astronomical world by declaring
that he saw some of the canals double--that is appearing as two
parallel lines. As these lines span the planet's surface for
distances of many thousands of miles the announcement naturally
gave rise to much surprise and, as I have said, to much
scepticism. But he resolutely stuck to his statement. Here is his
map of 1882. It is sufficiently startling.

In 1892 he drew a new map. It adds a little to the former map,
but the doubling was not so well seen. It is just the strangest
feature about this doubling that at times it is conspicuous, at
times invisible. A line which is distinctly seen as a single line
at one time, a few weeks later will appear distinctly to consist
of two parallel lines; like railway tracks, but tracks perhaps
200 miles apart and up to 3,000 or even 4,000 miles in length.

Many speculations were, of course, made to account for the origin
of such features. No known surface peculiarity on the Earth or
moon at all resembles these features. The moon's surface as you
know is cracked and

166

streaked. But the cracks are what we generally find cracks to
be--either aimless, wandering lines, or, if radiating from a
centre, then lines which contract in width as they leave the
point of rupture. Where will we find cracks accurately parallel
to one another sweeping round a planet's face with steady
curvature for, 4,000 miles, and crossing each other as if quite
unhampered by one another's presence? If the phenomenon on Mars
be due to cracks they imply a uniformity in thickness and
strength of crust, a homogeneity, quite beyond all anticipation.
We will afterwards see that the course of the lines is itself
further opposed to the theory that haphazard cracking of the
crust of the planet is responsible for the lines. It was also
suggested that the surface of the planet was covered with ice and
that these were cracks in the ice. This theory has even greater
difficulties than the last to contend with. Rivers have been
suggested. A glance at our own maps at once disposes of this
hypothesis. Rivers wander just as cracks do and parallel rivers
like parallel cracks are unknown.

In time the many suggestions were put aside. One only remained.
That the lines are actually the work of intelligence; actually
are canals, artificially made, constructed for irrigation
purposes on a scale of which we can hardly form any conception
based on our own earthly engineering structures.

During the opposition of 1894, Percival Lowell, along with A. E.
Douglass, and W. H. Pickering,

167

observed the planet from the summit of a mountain in Arizona,
using an 18-inch refracting telescope and every resource of
delicate measurement and spectroscopy. So superb a climate
favoured them that for ten months the planet was kept under
continual observation. Over 900 drawings were made and not only
were Schiaparelli's channels confirmed, but they added 116 to his
79, on that portion of the planet visible at that opposition.
They made the further important discovery that the lines do not
stop short at the dark regions of the planet's surface, as
hitherto believed, but go right on in many cases; the curvature
of the lines being unaltered.

Lowell is an uncompromising advocate of the "canal" theory. If
his arguments are correct we have at once an answer to our
question, "Are there other minds than ours?"

We must consider a moment Lowell's arguments; not that it is my
intention to combat them. You must form your own conclusions. I
shall lay before you another and, as I venture to think, more
adequate hypothesis in explanation of the channels of
Schiaparelli. We learn, however, much from Lowell's book--it is
full of interest.[1]

Lowell lays a deep foundation. He begins by showing that Mars has
an atmosphere. This must be granted him till some counter
observations are made.

[1] _Mars_, by Percival Lowell (Longmans, Green & Co.), 1896,

168

It is generally accepted. What that atmosphere is, is another
matter. He certainly has made out a good case for the presence of
water as one of its constituents,

It was long known that Mars possessed white regions at his poles,
just as our Earth does. The waning of these polar snows--if indeed
they are such--with the advance of the Martian summer, had often
been observed. Lowell plots day by day this waning. It is evident
from his observations that the snowfall must be light indeed. We
see in his map the south pole turned towards us. Mars in
perihelion always turns his south pole towards the sun and
therefore towards the Earth. We see that between the dates June
3rd to August 3rd--or in two months--the polar snow had almost
completely vanished. This denotes a very scanty covering. It must
be remembered that Mars even when nearest to the sun receives but
half our supply of solar heat and light.

But other evidence exists to show that Mars probably possesses
but little water upon his surface. The dark places are not
water-covered, although they have been named as if they were,
indeed, seas and lakes. Various phenomena show this. The canals
show it. It would never do to imagine canals crossing the seas.
No great rivers are visible. There is a striking absence of
clouds. The atmosphere of Mars seems as serene as that of Venus
appears to be cloudy. Mists and clouds, however, sometime appear
to veil his face and add to the difficulty of

169

making observations near the limb of the planet. Lowell concludes
it must be a calm and serene atmosphere; probably only
one-seventh of our own in density. The normal height of the
barometer in Mars would then be but four and a half inches. This
is a pressure far less than exists on the top of the highest
terrestrial mountain. A mountain here must have an altitude of
about ten miles to possess so low a pressure on its summit. Drops
of water big enough to form rain can hardly collect in such a
rarefied atmosphere. Moisture will fall as dew or frost upon the
ground. The days will be hot owing to the unimpeded solar
radiation; the nights bitterly cold owing to the free radiation
into space.

We may add that in such a climate the frost will descend
principally upon the high ground at night time and in the
advancing day it will melt. The freer radiation brings about this
phenomenon among our own mountains in clear and calm weather.

With the progressive melting of the snow upon the pole Lowell
connected many phenomena upon the planet's surface of much
interest. The dark spaces appear to grow darker and more
greenish. The canals begin to show themselves and reveal their
double nature. All this suggests that the moisture liberated by
the melting of the polar snow with the advancing year, is
carrying vitality and springtime over the surface of the planet.
But how is the water conveyed?

Lowell believes principally by the canals. These are

170

constructed triangulating the surface of the planet in all
directions. What we see, according to Lowell, is not the canal
itself, but the broad band of vegetation which springs up on the
arrival of the water. This band is perhaps thirty or forty miles
wide, but perhaps much less, for Lowell reports that the better
the conditions of observation the finer the lines appeared, so
that they may be as narrow, possibly, as fifteen miles. It is to
be remarked that a just visible dot on the surface of Mars must
possess a diameter of 30 miles. But a chain of much smaller dots
will be visible, just as we can see such fine objects as spiders'
webs. The widening of the canals is then accounted for, according
to Lowell, by the growth of a band of vegetation, similar to that
which springs into existence when the floods of the Nile irrigate
the plains of Egypt.

If no other explanation of the lines is forthcoming than that
they are the work of intelligence, all this must be remembered.
If all other theories fail us, much must be granted Lowell. We
must not reason like fishes--as Lowell puts it--and deny that
intelligent beings can thrive in an atmospheric pressure of four
and half inches of mercury. Zurbriggen has recently got to the
top of Aconcagua, a height of 24,000 feet. On the summit of such
a mountain the barometer must stand at about ten inches. Why
should not beings be developed by evolution with a lung capacity
capable of living at two and a half times this altitude. Those
steadily

171

curved parallel lines are, indeed, very unlike anything we have
experience of. It would be rather to be expected that another
civilisation than our own would present many wide differences in
its development.

What then is the picture we have before us according to Lowell?
It is a sufficiently dramatic one.

Mars is a world whose water supply, never probably very abundant,
has through countless years been drying up, sinking into his
surface. But the inhabitants are making a brave fight for it,
They have constructed canals right round their world so that the
water, which otherwise would run to waste over the vast deserts,
is led from oasis to oasis. Here the great centres of
civilisation are placed: their Londons, Viennas, New Yorks. These
gigantic works are the works of despair. A great and civilised
world finds death staring it in the face. They have had to triple
their canals so that when the central canal has done its work the
water is turned into the side canals, in order to utilise it as
far as possible. Through their splendid telescopes they must view
our seas and ample rivers; and must die like travellers in the
desert seeing in a mirage the cool waters of a distant lake.

Perhaps that lonely signal reported to have been seen in the
twilight limb of Mars was the outcome of pride in their splendid
and perishing civilisation. They would leave some memory of it:
they would have us witness how great was that civilisation before
they perish!

I close this dramatic picture with the poor comfort

172

that several philanthropic people have suggested signalling to
them as a mark of sympathy. It is said that a fortune was
bequeathed to the French Academy for the purpose of communicating
with the Martians. It has been suggested that we could flash
signals to them by means of gigantic mirrors reflecting the light
of our Sun. Or, again, that we might light bonfires on a
sufficiently large scale. They would have to be about ten miles
in diameter! A writer in the Pall Mall Gazette suggested that
there need really be no difficulty in the matter. With the kind
cooperation of the London Gas Companies (this was before the days
of electric lighting) a signal might be sent without any
additional expense if the gas companies would consent to
simultaneously turn off the gas at intervals of five minutes over
the whole of London, a signal which would be visible to the
astronomers in Mars would result. He adds, naively: "If only
tried for an hour each night some results might be obtained."

II

We have reviewed the theory of the artificial construction of the
Martian lines. The amount of consideration we are disposed to
give to the supposition that there are upon Mars other minds than
ours will--as I have stated--necessarily depend upon whether or not
we can assign a probable explanation of the lines upon purely
physical grounds. If it is apparent that such

173

lines would be formed with great probability under certain
conditions, which conditions are themselves probable, then the
argument by exclusion for the existence of civilisation on Mars,
at once breaks down.

{Fig. 10}

As a romance writer is sometimes under the necessity of
transporting his readers to other scenes, so I must now ask you
to consent to be transported some millions

174

of miles into the region of the heavens which lies outside Mars'
orbit.

Between Mars and Jupiter is a chasm of 341 millions of miles.
This gap in the sequence of planets was long known to be quite
out of keeping with the orderly succession of worlds outward from
the Sun. A society was formed at the close of the last century
for the detection of the missing world. On the first day of the
last century, Piazzi--who, by the way, was not a member of the
society--discovered a tiny world in the vacant gap. Although
eagerly welcomed, as better than nothing, it was a disappointing
find. The new world was a mere rock. A speck of about 160 miles
in diameter. It was obviously never intended that such a body
should have all this space to itself. And, sure enough, shortly
after, another small world was discovered. Then another was
found, and another, and so on; and now more than 400 of these
strange little worlds are known.

But whence came such bodies? The generally accepted belief is
that these really represent a misbegotten world. When the Sun was
younger he shed off the several worlds of our system as so many
rings. Each ring then coalesced into a world. Neptune being the
first born; Mercury the youngest born.

After Jupiter was thrown off, and the Sun had shrunk away inwards
some 20o million miles, he shed off another ring. Meaning that
this offspring of his should grow up like the rest, develop into
a stable world with the

175

potentiality even, it may be, of becoming the abode of rational
beings. But something went wrong. It broke up into a ring of
little bodies, circulating around him.

It is probable on this hypothesis that the number we are
acquainted with does not nearly represent the actual number of
past and present asteroids. It would take 125,000 of the biggest
of them to make up a globe as big as our world. They, so far as
they are known, vary in size from 10 miles to 160 miles in
diameter. It is probable then--on the assumption that this failure
of a world was intended to be about the mass of our Earth--that
they numbered, and possibly number, many hundreds of thousands.

Some of these little bodies are very peculiar in respect to the
orbits they move in. This peculiarity is sometimes in the
eccentricity of their orbits, sometimes in the manner in which
their orbits are tilted to the general plane of the ecliptic, in
which all the other planets move.

The eccentricity, according to Proctor, in some cases may attain
such extremes as to bring the little world inside Mars' mean
distance from the sun. This, as you will remember, is very much
less than his greatest distance from the sun. The entire belt of
asteroids--as known--lie much nearer to Mars than to Jupiter.

As regards the tilt of their orbits, some are actually as much as
34 degrees inclined to the ecliptic, so that in fact they are
seen from the Earth among our polar constellations.

176

From all this you see that Mars occupies a rather hot comer in
the solar system. Is it not possible that more than once in the
remote past Mars may have encountered one of these wanderers? If
he came within a certain distance of the small body his great
mass would sway it from its orbit, and under certain conditions
he would pick up a satellite in this manner. That his present
satellites were actually so acquired is the suggestion of Newton,
of Yale College.

Mars' satellites are indeed suspiciously and most abnormally
small. I have not time to prove this to you by comparison with
the other worlds of the solar system. In fact, they were not
discovered till 1877--although they were predicted in a most
curious manner, with the most uncannily accurate details, by
Swift.

One of these bodies is about 36 miles in diameter. This is
Phobos. Phobos is only 3.700 miles from the surface of Mars. The
other is smaller and further off. He is named Deimos, and his
diameter is only 10 miles. He is 12,500 miles from Mars' surface.
With the exception of Phobos the next smallest satellite known in
the solar system is one of Saturn's--Hyperion; almost 800 miles in
diameter. The inner one goes all round Mars in 7½ hours. This is
Phobos' month. Mars turns on his axis in 24 hours and 40 minutes,
so that people in Mars would see the rise of Phobos twice in the
course of a day and night; lie would apparently cross the sky

177

going against the other satellite; that is, he would move
apparently from west to east.

We may at least assume as probable that other satellites have
been gathered by Mars in the past from the army of asteroids.

Some of the satellites so picked up would be direct: that is,
would move round the planet in the direction of his axial
rotation. Others, on the chances, would be retrograde: that is,
would move against his axial rotation. They would describe orbits
making the same various angles with the ecliptic as do the
asteroids; and we may be sure they would be of the same varying
dimensions.

We go on to inquire what would be the consequence to Mars of such
captures.

A satellite captured in this manner is very likely to be pulled
into the Planet. This is a probable end of a satellite in any
case. It will probably be the end of our satellite too. The
satellite Phobos is indeed believed to be about to take this very
plunge into his planet. But in the case when the satellite picked
up happens to be rotating round the planet in the opposite
direction to the axial rotation of the planet, it is pretty
certain that its career as a satellite will be a brief one. The
reasons for this I cannot now give. If, then, Mars picked up
satellites he is very sure to have absorbed them sooner or later.
Sooner if they happened to be retrograde satellites, later if
direct satellites. His present satellites are recent additions.
They are direct.

178

The path of an expiring satellite will be a slow spiral described
round the planet. The spiral will at last, after many years,
bring the satellite down upon the surface of the primary. Its
final approach will be accelerated if the planet possesses an
atmosphere, as Mars probably does. A satellite of the dimensions
of Phobos--that is 36 miles in diameter--would hardly survive more
than 30 to 60 years within seventy miles of Mars' surface. It
will then be rotating round Mars in an hour and forty minutes,
moving, in fact, at the rate of 2.2 miles per second. In the
course of this 30 or 60 years it will, therefore, get round
perhaps 200,000 times, before it finally crashes down upon the
Martians. During this closing history of the satellite there is
reason to believe, however, that it would by no means pursue
continually the same path over the surface of the planet. There
are many disturbing factors to be considered. Being so small any
large surface features of Mars would probably act to perturb the
orbit of the satellite.

The explanation of Mars' lines which I suggest, is that they were
formed by the approach of such satellites in former times. I do
not mean that they are lines cut into his surface by the actual
infall of a satellite. The final end of the satellite would be
too rapid for this, I think. But I hope to be able to show you
that there is reason to believe that the mere passage of the
satellite, say at 70 miles above the surface of the planet, will,
in itself, give rise to effects on the crust of the planet
capable

179

of accounting for just such single or parallel lines as we see.

In the first place we have to consider the stability of the
satellite. Even in the case of a small satellite we cannot
overlook the fact that the half of the satellite near the planet
is pulled towards the planet by a gravitational force greater
than that attracting the outer half, and that the centrifugal
force is less on the inner than on the outer hemisphere. Hence
there exists a force tending to tear the satellite asunder on the
equatorial section tangential

{Fig. 11}

to the planet's surface. If in a fluid or plastic state, Phobos,
for instance, could not possibly exist near the planet's surface.
The forces referred to would decide its fate. It may be shown by
calculation, however, that if Phobos has the strength of basalt
or glass there would remain a considerable coefficient of safety
in favour of the satellite's stability; even when the surfaces of
planet and satellite were separated by only five miles.

We have now to consider some things which we expect will happen
before the satellite takes its final plunge into the planet.

180

This diagram (Fig. 11) shows you the satellite travelling above
the surface of the planet. The satellite is advancing towards, or
away from, the spectator. The planet is supposed to show its
solid crust in cross section, which may be a few miles in
thickness. Below this is such a hot plastic magma as we have
reason to believe underlies much of the solid crust of our own
Earth. Now there is an attraction between the satellite and the
crust of the planet; the same gravitational attraction which
exists between every particle of matter in the universe. Let us
consider how this attraction will affect the planet's crust. I
have drawn little arrows to show how we may consider the
attraction of the satellite pulling the crust of the planet not
only upwards, but also pulling it inwards beneath the satellite.
I have made these arrows longer where calculation shows the
stress is greater. You see that the greatest lifting stress is
just beneath the satellite, whereas the greatest stress pulling
the crust in under the satellite is at a point which lies out
from under the satellite, at a considerable distance. At each
side of the satellite there is a point where the stress pulling
on the crust is the greatest. Of the two stresses the lifting
stress will tend to raise the crust a little; the pulling stress
may in certain cases actually tear the crust across; as at A and
B.

It is possible to calculate the amount of the stress at the point
at each side of the satellite where the stress is at its
greatest. We must assume the satellite to be a certain size and
density; we must also assume the crust of

181

Mars to be of some certain density. To fix our ideas on these
points I take the case of the present satellite Phobos. What
amount of stress will he exert upon the crust of Mars when he
approaches within, say, 40 miles of the planet's surface? We know
his size approximately--he is about 36 miles in diameter. We can
guess his density to be between four times that of water and
eight times that of water. We may assume the density of Mars'
surface to be about the same as that of our Earth's surface, that
is three times as dense as water. We now find that the greatest
stress tending to rend open the surface crust of Mars will be
between 4,000 and 8,000 pounds to the square foot according to
the density we assign to Phobos.

Will such a stress actually tear open the crust? We are not able
to answer this question with any certainty. Much will depend upon
the nature and condition of the crust. Thus, suppose that we are
here (Fig. 12) looking down upon the satellite which is moving
along slowly relatively to Mars' surface, in the direction of the
arrow. The satellite has just passed over a weak and cracked part
of the planet's crust. Here the stress has been sufficient to
start two cracks. Now you know how easy it is to tear a piece of
cloth when you go to the edge of it in order to make a beginning.
Here the stress from the satellite has got to the edge of the
crust. It is greatly concentrated just at the extremities of the
cracks. It will, unler such circumstances probably carry on the

182

tear. If it does not do so this time, remember the satellite will
some hours later be coming over the same place again, and then
again for, at least, many hundreds of times. Then also we are not
limited to the assumption that the

{Fig. 12}

satellite is as small as Phobos. Suppose we consider the case of
a satellite approaching Mars which has a diameter double that of
Phobos; a diameter still much less than that of the larger class
of asteroids. Even at the distance

183

of 65 miles the stress will now amount to as much as from 15 to
30 tons per square foot. It is almost certain that such a stress
repeated a comparatively few times over the same parts of the
planet's surface would so rend the crust as to set up lines along
which plutonic action would find a vent. That is, we might expect
along these lines all the phenomena of upheaval and volcanic
eruption which give rise to surface elevations.

The probable effect of a satellite of this dimension travelling
slowly relatively to the surface of Mars is, then, to leave a
very conspicuous memorial of his presence behind him. You see
from the diagram that this memorial will consist o: two parallel
lines of disturbance.

The linear character of the gravitational effects of the
satellite is due entirely to the motion of the satellite
relatively to the surface of the planet. If the satellite stood
still above the surface the gravitational stress in the crust
would, of course, be exerted radially outwards from the centre of
the satellite. It would attain at the central point beneath the
satellite its maximum vertical effect, and at some radial
distance measured outwards from this point, which distance we can
calculate, its maximum horizontal tearing effect. When the
satellite moves relatively to the planet's crust, the horizontal
tearing force acts differently according to whether it is
directed in the line of motion or at right angles to this line.

In the direction of motion we see that the satellite

184

creates as it passes over the crust a wave of rarefaction or
tension as at D, followed by compression just beneath the
satellite and by a reversed direction of gravitational pull as
the satellite passes onwards. These stresses rapidly replace one
another as the satellite travels along. They are resisted by the
inertia of the crust, and are taken up by its elasticity. The
nature of this succession of alternate compressions and
rarefactions in the crust possess some resemblance to those
arising in an earthquake shock.

If we consider the effects taking place laterally to the line of
motion we see that there are no such changes in the nature of the
forces in the crust. At each passage of the satellite the
horizontal tearing stress increases to a maximum, when it is
exerted laterally, along the line passing through the horizontal
projection of the satellite and at right angles to the line of
motion, and again dies away. It is always a tearing stress,
renewed again and again.

This effect is at its maximum along two particular parallel lines
which are tangents to the circle of maximum horizontal stress and
which run parallel with the path of the satellite. The distance
separating these lines depend upon the elevation of the satellite
above the planet's surface. Such lines mark out the theoretical
axes of the "double canals" which future crustal movements will
more fully develop.

It is interesting to consider what the effect of such

185

conditions would be if they arose at the surface of our own
planet. We assume a horizontal force in the crust adequate to set
up tensile stresses of the order, say, of fifteen tons to the
square foot and these stresses to be repeated every few hours;
our world being also subject to the dynamic effects we recognise
in and beneath its crust.

It is easy to see that the areas over which the satellite exerted
its gravitational stresses must become the foci --foci of linear
form--of tectonic developments or crust movements. The relief of
stresses, from whatever cause arising, in and beneath the crust
must surely take place in these regions of disturbance and along
these linear areas. Here must become concentrated the folding
movements, which are under existing conditions brought into the
geosynclines, along with their attendant volcanic phenomena. In
the case of Mars such a concentration of tectonic events would
not, owing to the absence of extensive subaerial denudation and
great oceans, be complicated by the existence of such synclinal
accumulations as have controlled terrestrial surface development.
With the passage of time the linear features would probably
develop; the energetic substratum continually asserting its
influence along such lines of weakness. It is in the highest
degree probable that radioactivity plays no less a part in
Martian history than in terrestrial. The fact of radioactive
heating allows us to assume the thin surface crust and continued
sub-crustal energy throughout the entire period of the planet's
history.

186

How far willl these effects resemble the double canals of Mars?
In this figure and in the calculations I have given you I have
supposed the satellite engaged in marking the planet's surface
with two lines separated by about the interval separating the
wider double canals of Mars--that is about 220 miles apart. What
the distance between the lines will be, as already stated, will
depend upon the height of the satellite above the surface when it
comes upon a part of the crust in a condition to be affected by
the stresses it sets up in it. If the satellite does its work at
a point lower down above the surface the canal produced will be
narrower. The stresses, too, will then be much greater. I must
also observe that once the crust has yielded to the pulling
stress, there is great probability that in future revolutions of
the satellite a central fracture will result. For then all the
pulling force adds itself to the lifting force and tends to crush
the crust inwards on the central line beneath the satellite. It
is thus quite possible that the passage of a satellite may give
rise to triple lines. There is reason to believe that the canals
on Mars are in some cases triple.

I have spoken all along of the satellite moving slowly over the
surface of Mars. I have done so as I cannot at all pronounce so
readily on what will happen when the satellite's velocity over
the surface of Mars is very great. To account for all the lines
mapped by Lowell some of them must have been produced by
satellities moving relatively to the surface of Mars at
velocities so great

187

as three miles a second or even rather more. The stresses set up
are, in such cases, very difficult to estimate. It has not yet
been done. Parallel lines of greatest stress or impulse ought to
be formed as in the other case.

I now ask your attention to another kind of evidence that the
lines are due in some way to the motion of satellites passing
over the surface of Mars.

I may put the fresh evidence to which I refer, in this way: In
Lowell's map (P1. XXII, p. 192), and in a less degree in
Schiaparelli's map (ante p. 166), we are given the course of the
lines as fragments of incomplete curves. Now these curves might
have been anything at all. We must take them as they are,
however, when we apply them as a test of the theory that the
motion of a satellite round Mars can strike such lines. If it can
be shown that satellites revolving round Mars might strike just
such curves then we assume this as an added confirmation of the
hypothesis.

We must begin by realising what sort of curves a satellite which
disturbs the surface of a planet would leave behind it after its
demise. You might think that the satellite revolving round and
round the planet must simply describe a circle upon the spherical
surface of the planet: a "great circle" as it is called; that is
the greatest circle which can be described upon a sphere. This
great circle can, however, only be struck, as you will see, when
the planet is not turning upon its axis: a condition not likely
to be realised.

This diagram (PI. XXI) shows the surface of a globe

188

covered with the usual imaginary lines of latitude and longitude.
The orbit of a supposed satellite is shown by a line crossing the
sphere at some assumed angle with the equator. Along this line
the satellite always moves at uniform velocity, passing across
and round the back of the sphere and again across. If the sphere
is not turning on its polar axis then this satellite, which we
will suppose armed with a pencil which draws a line upon the
sphere, will strike a great circle right round the sphere. But
the sphere is rotating. And it is to be expected that at
different times in a planet's history the rate of rotation varies
very much indeed. There is reason to believe that our own day was
once only 2½ hours long, or thereabouts. After a preliminary rise
in velocity of axial rotation, due to shrinkage attending rapid
cooling, a planet as it advances in years rotates slower and
slower. This phenomenon is due to tidal influences of the sun or
of satellites. On the assumption that satellites fell into Mars
there would in his case be a further action tending to shorten
his day as time went on.

The effect of the rotation of the planet will be, of course, that
as the satellite advances with its pencil it finds the surface of
the sphere being displaced from under it. The line struck ceases
to be the great circle but wanders off in another curve--which is
in fact not a circle at all.

You will readily see how we find this curve. Suppose the sphere
to be rotating at such a speed that while the satellite is
advancing the distance _Oa_, the point _b_ on the

189

sphere will be carried into the path of the satellite. The pencil
will mark this point. Similarly we find that all the points along
this full curved line are points which will just find themselves
under the satellite as it passes with its pencil. This curve is
then the track marked out by the revolving satellite. You see it
dotted round the back of the sphere to where it cuts the equator
at a certain point. The course of the curve and the point where
it cuts the equator, before proceeding on its way, entirely
depend upon the rate at which we suppose the sphere to be
rotating and the satellite to be describing the orbit. We may
call the distance measured round the planet's equator separating
the starting point of the curve from the point at which it again
meets the equator, the "span" of the curve. The span then depends
entirely upon the rate of rotation of the planet on its axis and
of the satellite in its orbit round the planet.

But the nature of events might have been somewhat different. The
satellite is, in the figure, supposed to be rotating round the
sphere in the same direction as that in which the sphere is
turning. It might have been that Mars had picked up a satellite
travelling in the opposite direction to that in which he was
turning. With the velocity of planet on its axis and of satellite
in its orbit the same as before, a different curve would have
been described. The span of the curve due to a retrograde
satellite will be greater than that due to a direct satellite.
The retrograde satellite will have a span more than half

190

way round the planet, the direct satellite will describe a curve
which will be less than half way round the planet: that is a span
due to a retrograde satellite will be more than 180 degrees,
while the span due to a direct satellite will be less than 180
degrees upon the planet's equator.

I would draw your attention to the fact that what the span will
be does not depend upon how much the orbit of the satellite is
inclined to the equator. This only decides how far the curve
marked out by the satellite will recede from the equator.

We find then, so far, that it is easy to distinguish between the
direct and the retrograde curves. The span of one is less, of the
other greater, than 180 degrees. The number of degrees which
either sort of curve subtends upon the equator entirely depends
upon the velocity of the satellite and the axial velocity of the
planet.

But of these two velocities that of the satellite may be taken as
sensibly invariable, when close enough to use his pencil. This
depends upon the law of centrifugal force, which teaches us that
the mass of the planet alone decides the velocity of a satellite
in its orbit at any fixed distance from the planet's centre. The
other velocity--that of the planet upon its axis--was, as we have
seen, not in the past what it is now. If then Mars, at various
times in his past history, picked up satellites, these satellites
will describe curves round him having different spans which will
depend upon the velocity of axial rotation of Mars at the time
and upon this only.

191

In what way now can we apply this knowledge of the curves
described by a satellite as a test of the lunar origin of the
lines on Mars?

To do this we must apply to Lowell's map. We pick out preferably,
of course, the most complete and definite curves. The chain of
canals of which Acheron and Erebus are members mark out a fairly
definite curve. We produce it by eye, preserving the curvature as
far as possible, till it cuts the equator. Reading the span on
the equator we find' it to be 255 degrees. In the first place we
say then that this curve is due to a retrograde satellite. We
also note on Lowell's map that the greatest rise of the curve is
to a point about 32 degrees north of the equator. This gives the
inclination of the satellite's orbit to the plane of Mars'
equator.

With these data we calculate the velocity which the planet must
have possessed at the time the canal was formed on the hypothesis
that the curve was indeed the work of a satellite. The final
question now remains If we determine the curve due to this
velocity of Mars on its axis, will this curve fit that one which
appears on Lowell's map, and of which we have really availed
ourselves of only three points? To answer this question we plot
upon a sphere, the curve of a satellite, in the manner I have
described, assigning to this sphere the velocity derived from the
span of 255 degrees. Having plotted the curve on the sphere it
only remains to transfer it to Lowell's map. This is easily
done.

192

This map (Pl. XXII) shows you the result of treating this, as
well as other curves, in the manner just described. You see that
whether the fragmentary curves are steep and receding far from
the equator; or whether they are flat and lying close along the
equator; whether they span less or more than 180 degrees; the
curves determined on the supposition that they are the work of
satellites revolving round Mars agree with the mapped curves;
following them with wonderful accuracy; possessing their
properties, and, indeed, in some cases, actually coinciding with
them.

I may add that the inadmissible span of 180 degrees and spans
very near this value, which are not well admissible, are so far
as I can find, absent. The curves are not great circles.

You will require of me that I should explain the centres of
radiation so conspicuous here and there on Lowell's map. The
meeting of more than two lines at the oases is a phenomenon
possibly of the same nature and also requiring explanation.

In the first place the curves to which I have but briefly
referred actually give rise in most cases to nodal, or crossing
points; sometimes on the equator, sometimes off the equator;
through which the path of the satellite returns again and again.
These nodal points will not, however, afford a general
explanation of the many-branched radiants.

It is probable that we should refer such an appearance

193

as is shown at the Sinus Titanum to the perturbations of the
satellite's path due to the surface features on Mars. Observe
that the principal radiants are situated upon the boundary of the
dark regions or at the oases. Higher surface levels may be
involved in both cases. Some marked difference in topography must
characterise both these features. The latter may possibly
originate in the destruction of satellites. Or again, they may
arise in crustal disturbance of a volcanic nature, primarily
induced or localised by the crossing of two canals. Whatever the
origin of these features it is only necessary to assume that they
represent elevated features of some magnitude to explain the
multiplication of crossing lines. We must here recall what
observers say of the multiplicity of the canals. According to
Lowell, "What their number maybe lies quite beyond the
possibility of count at present; for the better our own air, the
more of them are visible."

Such innumerable canals are just what the present theory
requires. An in-falling satellite will, in the course of the last
60 or 80 years of its career, circulate some 100,000 times over
Mars' surface. Now what will determine the more conspicuous
development of a particular canal? The mass of the satellite; the
state of the surface crust; the proximity of the satellite; and
the amount of repetition over the same ground. The after effects
may be taken as proportional to the primary disturbance.

194

It is probable that elevated surface features will influence two
of these conditions: the number of repetitions and the proximity
to the surface. A tract 100 miles in diameter and elevated 5,000
or 10,000 feet would seriously perturb the orbit of such a body as
Phobos. It is to be expected that not only would it be effective
in swaying the orbit of the satellite in the horizontal direction
but also would draw it down closer to the surface. It is even to
be considered if such a mass might not become nodal to the
satellite's orbit, so that this passed through or above this
point at various inclinations with its primary direction. If
acting to bring down the orbit then this will quicken the speed
and cause the satellite further on its path to attain a somewhat
higher elevation above the surface. The lines most conspicuous in
the telescope are, in short, those which have been favoured by a
combination of circumstances as reviewed above, among which
crustal features have, in some cases, played a part.

I must briefly refer to what is one of the most interesting
features of the Martian lines: the manner in which they appear to
come and go like visions.

Something going on in Mars determines the phenomenon. On a
particular night a certain line looks single. A few nights later
signs of doubling are perceived, and later still, when the seeing
is particularly good, not one but two lines are seen. Thus, as an
example, we may take the case of Phison and Euphrates. Faint
glimpses of the dual state were detected in the summer

195

and autumn, but not till November did they appear as distinctly
double. Observe that by this time the Antarctic snows had melted,
and there was in addition, sufficient time for the moisture so
liberated to become diffused in the planet's atmosphere.

This increase in the definition and conspicuousness of certain
details on Mars' surface is further brought into connection with
the liberation of the polar snows and the diffusion of this water
through the atmosphere, by the fact that the definition appeared
progressively better from the south pole upwards as the snow
disappeared. Lowell thinks this points to vegetation springing up
under the influence of moisture; he considers, however, as we
have seen, that the canals convey the moisture. He has to assume
the construction of triple canals to explain the doubling of the
lines.

If we once admit the canals to be elevated ranges--not necessarily
of great height--the difficulty of accounting for increased
definition with increase of moisture vanishes. We need not
necessarily even suppose vegetation concerned. With respect to
this last possibility we may remark that the colour observations,
upon which the idea of vegetation is based, are likely to be
uncertain owing to possible fatigue effects where a dark object
is seen against a reddish background.

However this may be we have to consider what the effects of
moisture increasing in the atmosphere of Mars will be with regard
to the visibility of elevated ranges,

196

We assume a serene and rare atmosphere: the nights intensely
cold, the days hot with the unveiled solar radiation. On the hill
tops the cold of night will be still more intense and so, also,
will the solar radiation by day. The result of this state of
things will be that the moisture will be precipitated mainly on
the mountains during the cold of night--in the form of frost--and
during the day this covering of frost will melt; and, just as we
see a heavy dew-fall darken the ground in summer, so the melting
ice will set off the elevated land against the arid plains below.
Our valleys are more moist than our mountains only because our
moisture is so abundant that it drains off the mountains into the
valleys. If moisture was scarce it would distil from the plains
to the colder elevations of the hills. On this view the
accentuation of a canal is the result of meteorological effects
such as would arise in the Martian climate; effects which must be
influenced by conditions of mountain elevation, atmospheric
currents, etc. We, thus, follow Lowell in ascribing the
accentuation of the canals to the circulation of water in Mars;
but we assume a simple and natural mode of conveyance and do not
postulate artificial structures of all but impossible magnitude.
That vegetation may take part in the darkening of the elevated
tracts is not improbable. Indeed we would expect that in the
Martian climate these tracts would be the only fertile parts of
the surface.

Clouds also there certainly are. More recent observations

197

appear to have set this beyond doubt. Their presence obviously
brings in other possible explanations of the coming and going of
elevated surface features.

Finally, we may ask what about the reliability of the maps? About
this it is to be said that the most recent map--that by Lowell--has
been confirmed by numerous drawings by different observers, and
that it is,itself the result of over 900 drawings. It has become
a standard chart of Mars, and while it would be rash to contend
for absence of errors it appears certain that the trend of the
principal canals may be relied on, as, also, the general features
of the planet's surface.

The question of the possibility of illusion has frequently been
raised. What I have said above to a great extent answers such
objections. The close agreement between the drawings of different
observers ought really to set the matter at rest. Recently,
however, photography has left no further room for scepticism.
First photographed in 1905, the planet has since been
photographed many thousands of times from various observatories.
A majority of the canals have been so mapped. The doubling of the
canals is stated to have been also so recorded.[1]

The hypothesis which I have ventured to put before you involves
no organic intervention to account for the

[1] E. C. Slipher's paper in _Popular Astronomy_ for March, 1914,
gives a good account of the recent work.

198

details on Mars' surface. They are physical surface features.
Mars presents his history written upon his face in the scars of
former encounters--like the shield of Sir Launcelot. Some of the
most interesting inferences of mathematical and physical
astronomy find a confirmation in his history. The slowing down in
the rate of axial rotation of the primary; the final inevitable
destruction of the satellite; the existence in the past of a far
larger number of asteroids than we at present are acquainted
with; all these great facts are involved in the theory now
advanced. If justifiably, then is Mars' face a veritable
Principia.

To fully answer the question which heads these lectures, we
should go out into the populous solitudes (if the term be
permitted) which lie beyond our system. It is well that there is
now no time left to do so; for, in fact, there we can only dream
dreams wherein the limits of the possible and the impossible
become lost.

The marvel of the infinite number of stars is not so marvellous
as the rationality that fain would comprehend them. In seeking
other minds than ours we seek for what is almost infinitely
complex and coordinated in a material universe relatively simple
and heterogeneous. In our mental attitude towards the great
question, this fact must be regarded as fundamental.

I can only fitly close a discourse which has throughout weighed
the question of the living thought against the unthinking laws of
matter, by a paraphrase of the words

199

of a great poet when he, in higher and, perhaps, more philosophic
language, also sought to place the one in comparison with the
other.[1]

Richter thought that he was--with his human heart
unstrengthened--taken by an angel among the universe of stars.
Then, as they journeyed, our solar system was sunken like a faint
star in the abyss, and they travelled yet further, on the wings
of thought, through mightier systems: through all the countless
numbers of our galaxy. But at length these also were left behind,
and faded like a mist into the past. But this was not all. The
dawn of other galaxies appeared in the void. Stars more countless
still with insufferable light emerged. And these also were
passed. And so they went through galaxies without number till at
length they stood in the great Cathedral of the Universe. Endless
were the starry aisles; endless the starry columns; infinite the
arches and the architraves of stars. And the poet saw the mighty
galaxies as steps descending to infinity, and as steps going up
to infinity.

Then his human heart fainted and he longed for some narrow cell;
longed to lie down in the grave that he might hide from infinity.
And he said to the angel:

"Angel, I can go with thee no farther. Is there, then, no end to
the universe of stars?"

[1] De Quincy in his _System of the Heavens_ gives a fine
paraphrase of "Richter's Dream."

200

Then the angel flung up his glorious hands to the heaven of
heavens, saying "End is there none to the universe of God? Lo!
also there is no beginning."

201

THE LATENT IMAGE [1]

My inclination has led me, in spite of a lively dread of
incurring a charge of presumption, to address you principally on
that profound and most subtle question, the nature and mode of
formation of the photographic image. I am impelled to do so, not
only because the subject is full of fascination and hopefulness,
but because the wide topics of photographic methods or
photographic applications would be quite unfittingly handled by
the president you have chosen.

I would first direct your attention to Sir James Dewar's
remarkable result that the photographic plate retains
considerable power of forming the latent image at temperatures
approaching the absolute zero--a result which, as I submit,
compels us to regard the fundamental effects progressing in the
film under the stimulus of light undulations as other than those
of a purely chemical nature. But few, if any, instances of
chemical combination or decomposition are known at so low a
temperature. Purely chemical actions cease, indeed, at far higher
temperatures, fluorine being among the few bodies which still
show

[1] Presidential address to the Photographic Convention of the
United Kingdom, July, 1905. _Nature_, Vol. 72, p. 308.

202

chemical activity at the comparatively elevated temperature of
-180° C. In short, this result of Sir James Dewar's suggests that
we must seek for the foundations of photographic action in some
physical or intra-atomic effect which, as in the case of
radioactivity or fluorescence, is not restricted to intervals of
temperature over which active molecular vis viva prevails. It
compels us to regard with doubt the role of oxidation or other
chemical action as essential, but rather points to the view that
such effects must be secondary or subsidiary. We feel, in a word,
that we must turn for guidance to some purely photo-physical
effect.

Here, in the first place, we naturally recall the views of Bose.
This physicist would refer the formation of the image to a strain
of the bromide of silver molecule under the electric force in the
light wave, converting it into what might be regarded as an
allotropic modification of the normal bromide which subsequently
responds specially to the attack of the developer. The function
of the sensitiser, according to this view, is to retard the
recovery from strain. Bose obtained many suggestive parallels
between the strain phenomena he was able to observe in silver and
other substances under electromagnetic radiation and the
behaviour of the photographic plate when subjected to
long-continued exposure to light.

This theory, whatever it may have to recommend it, can hardly be
regarded as offering a fundamental explanation. In the first
place, we are left in the dark as to what

203

the strain may be. It may mean many and various things. We know
nothing as to the inner mechanism of its effects upon subsequent
chemical actions--or at least we cannot correlate it with what is
known of the physics of chemical activity. Finally, as will be
seen later, it is hardly adequate to account for the varying
degrees of stability which may apparently characterise the latent
image. Still, there is much in Bose's work deserving of careful
consideration. He has by no means exhausted the line of
investigation he has originated.

Another theory has doubtless been in the minds of many. I have
said we must seek guidance in some photo-physical phenomenon.
There is one such which preeminently connects light and chemical
phenomena through the intermediary of the effects of the former
upon a component part of the atom. I refer to the phenomena of
photo-electricity.

It was ascertained by Hertz and his immediate successors that
light has a remarkable power of discharging negative
electrification from the surface of bodies--especially from
certain substances. For long no explanation of the cause of this
appeared. But the electron--the ubiquitous electron--is now known
with considerable certainty to be responsible. The effect of the
electric force in the light wave is to direct or assist the
electrons contained in the substance to escape from the surface
of the body. Each electron carries away a very small charge of
negative electrification. If, then, a body is

204

originally charged negatively, it will be gradually discharged by
this convective process. If it is not charged to start with, the
electrons will still be liberated at the surface of the body, and
this will acquire a positive charge. If the body is positively
charged at first, we cannot discharge it by illumination.

It would be superfluous for me to speak here of the nature of
electrons or of the various modes in which their presence may be
detected. Suffice it to say, in further connection with the Hertz
effect, that when projected among gaseous molecules the electron
soon attaches itself to one of these. In other words, it ionises
a molecule of the gas or confers its electric charge upon it. The
gaseous molecule may even be itself disrupted by impact of the
electron, if this is moving fast enough, and left bereft of an
electron.

We must note that such ionisation may be regarded as conferring
potential chemical properties upon the molecules of the gas and
upon the substance whence the electrons are derived. Similar
ionisation under electric forces enters, as we now believe, into
all the chemical effects progressing in the galvanic cell, and,
indeed, generally in ionised solutes.

An experiment will best illustrate the principles I wish to
remind you of. A clean aluminium plate, carefully insulated by a
sulphur support, is faced by a sheet of copper-wire-gauze placed
a couple of centimetres away from it. The gauze is maintained at
a high positive

205

potential by this dry pile. A sensitive gold-leaf electroscope is
attached to the aluminium plate, and its image thrown upon the
screen. I now turn the light from this arc lamp upon the wire
gauze, through which it in part passes and shines upon the
aluminium plate. The electroscope at once charges up rapidly.
There is a liberation of negative electrons at the surface of the
aluminium; these, under the attraction of the positive body, are
rapidly removed as ions, and the electroscope charges up
positively.

Again, if I simply electrify negatively this aluminium plate so
that the leaves of the attached electroscope diverge widely, and
now expose it to the rays from the arc lamp, the charge, as you
see, is very rapidly dissipated. With positive electrification of
the aluminium there is no effect attendant on the illumination.

Thus from the work of Hertz and his successors we know that
light, and more particularly what we call actinic light, is an
effective means of setting free electrons from certain
substances. In short, our photographic agent, light, has the
power of expelling from certain substances the electron which is
so potent a factor in most, if not in all, chemical effects. I
have not time here to refer to the work of Elster and Geitel
whereby they have shown that this action is to be traced to the
electric force in the light wave, but must turn to the probable
bearing of this phenomenon on the familiar facts of photography.
I assume that the experiment I have shown you is the most

206

fundamental photographic experiment which it is now in our power
to make.

We must first ask from what substances can light liberate
electrons. There are many--metals as well as non-metals and
liquids. It is a very general phenomenon and must operate widely
throughout nature. But what chiefly concerns the present
consideration is the fact that the haloid salts of silver are
vigorously photo-electric, and, it is suggestive, possess,
according to Schmidt, an activity in the descending order
bromide, chloride, iodide. This is, in other words, their order
of activity as ionisers (under the proper conditions) when
exposed to ultra-violet light. Photographers will recognise that
this is also the order of their photographic sensitiveness.

Another class of bodies also concerns our subject: the special
sensitisers used by the photographer to modify the spectral
distribution of sensibility of the haloid salts, _e.g._ eosine,
fuchsine, cyanine. These again are electron-producers under light
stimulus. Now it has been shown by Stoletow, Hallwachs, and
Elster and Geitel that there is an intimate connection between
photo-electric activity and the absorption of light by the
substance, and, indeed, that the particular wave-lengths absorbed
by the substance are those which are effective in liberating the
electrons. Thus we have strong reason for believing that the
vigorous photo-electric activity displayed by the special
sensitisers must be dependent upon their colour absorption. You
will recognise that this is just

207

the connection between their photographic effects and their
behaviour towards light.

There is yet another suggestive parallel. I referred to the
observation of Sir James Dewar as to the continued sensitiveness
of the photographic film at the lowest attained extreme of
temperature, and drew the inference that the fundamental
photographic action must be of intra-atomic nature, and not
dependent upon the vis viva of the molecule or atom. In then
seeking the origin of photographic action in photo-electric
phenomena we naturally ask, Are these latter phenomena also
traceable at low temperatures? If they are, we are entitled to
look upon this fact as a qualifying characteristic or as another
link in the chain of evidence connecting photographic with
photo-electric activity.

I have quite recently, with the aid of liquid air supplied to me
from the laboratory of the Royal Dublin Society, tested the
photo-sensibility of aluminium and also of silver bromide down to
temperatures approaching that of the liquid air. The mode of
observation is essentially that of Schmidt--what he terms his
static method. The substance undergoing observation is, however,
contained at the bottom of a thin copper tube, 5 cm. in diameter,
which is immersed to a depth of about 10 cm in liquid air. The
tube is closed above by a paraffin stopper which carries a thin
quartz window as well as the sulphur tubes through which the
connections pass. The air within is very carefully dried by
phosphorus

208

pentoxide before the experiment. The arc light is used as source
of illumination. It is found that a vigorous photo-electric
effect continues in the case of the clean aluminium. In the case
of the silver bromide a distinct photo-electric effect is still
observed. I have not had leisure to make, as yet, any trustworthy
estimate of the percentage effect at this temperature in the case
of either substance. Nor have I determined the temperature
accurately. The latter may be taken as roughly about -150° C,

Sir James Dewar's actual measilrements afforded twenty per cent.
of the normal photographic effect at -180° C. and ten per cent.
at the temperature of -252.5° C.

With this much to go upon, and the important additional fact that
the electronic discharge--as from the X-ray tube or from
radium--generates the latent image, I think we are fully entitled
to suggest, as a legitimate lead to experiment, the hypothesis
that the beginnings of photographic action involve an electronic
discharge from the light-sensitive molecule; in other words that
the latent image is built up of ionised atoms or molecules the
result of the photo-electric effect on the illuminated silver
haloid, and it is upon these ionised atoms that the chemical
effects of the developer are subsequently directed. It may be
that the liberated electrons ionise molecules not directly
affected, or it may be that in their liberation they disrupt
complex molecules built up in the ripening of the

209

emulsion. With the amount we have to go upon we cannot venture to
particularise. It will be said that such an action must be in
part of the nature of a chemical effect. This must be admitted,
and, in so far as the rearrangement of molecular fabrics is
involved, the result will doubtless be controlled by temperature
conditions. The facts observed by Sir James Dewar support this.
But there is involved a fundamental process--the liberation of the
electron by the electric force in the light wave, which is a
physical effect, and which, upon the hypothesis of its reality as
a factor in forming the latent image, appears to explain
completely the outstanding photographic sensitiveness of the film
at temperatures far below those at which chemical actions in
general cease.

Again, we may assume that the electron--producing power of the
special sensitiser or dye for the particular ray it absorbs is
responsible, or responsible in part, for the special
sensitiveness it confers upon the film. Sir Wm. Abney has shown
that these sensitisers are active even if laid on as a varnish on
the sensitive surface and removed before development. It must be
remembered, however, that at temperatures of about -50° these
sensitisers lose much of their influence on the film; as I have
pointed out in a paper read before the Photographic Convention of
1894.

It. appears to me that on these views the curious phenomenon of
recurrent reversals does not present a problem hopeless of
explanation. The process of photo-

210

ionisation constituting the latent image, where the ion is
probably not immediately neutralised by chemical combination,
presents features akin to the charging of a capacity--say a Leyden
jar. There may be a rising potential between the groups of ions
until ultimately a point is attained when there is a spontaneous
neutralisation. I may observe that the phenomena of reversal
appear to indicate that the change in the silver bromide
molecule, whatever be its nature, is one of gradually increasing
intensity, and finally attains a maximum when a return to the
original condition occurs. The maximum is the point of most
intense developable image. It is probable that the sensitiser--in
this case the gelatin in which the bromide of silver is
immersed--plays a part in the conditions of stability which are
involved.

Of great interest in all our considerations and theories is the
recent work of Wood on photographic reversal. The result of this
work is--as I take it--to show that the stability of the latent
image may be very various according to the mode of its formation.
Thus it appears that the sort of latent effect which is produced
by pressure or friction is the least stable of any. This may be
reversed or wiped out by the application of any other known form
of photographic stimulus. Thus an exposure to X-rays will
obliterate it, or a very brief exposure to light. The latent
image arising from X-rays is next in order of increasing
stability. Light action will remove this. Third in order is a
very brief light-shock or sudden flash. This

211

cannot be reversed by any of the foregoing modes of stimulation,
but a long-continued undulatory stimulus, as from lamp-light,
will reverse it. Last and most stable of all is the gradually
built-up configuration due to long-continued light exposure. This
can only be reversed by overdoing it according to the known facts
of recurrent reversal. Wood takes occasion to remark that these
phenomena are in bad agreement with the strain theory of Bose. We
have, in fact, but the one resource--the allotropic modification
of the haloid--whereby to explain all these orders of stability.
It appears to me that the elasticity of the electronic theory is
greater. The state of the ionised system may be very various
according as it arises from continued rhythmic effects or from
unorganised shocks. The ionisation due to X-rays or to friction
will probably be quite unorganised, that due to light more or
less stable according to the gradual and gentle nature of the
forces at work. I think we are entitled to conclude that on the
whole there is nothing in Wood's beautiful experiments opposed to
the photo-electric origin of photographic effects, but that they
rather fall in with what might be anticipated according to that
theory.

When we look for further support to the views I have laid before
you we are confronted with many difficulties. I have not as yet
detected any electronic discharge from the film under light
stimulus. This may be due to my defective experiments, or to a
fact noted by Elster and Geitel concerning the photo-electric
properties of gelatin.

212

They obtained a vigorous effect from Balmain's luminous paint,
but when this was mixed in gelatin there was no external effect.
Schmidt's results as to the continuance of photo-electric
activity when bodies in general are dissolved in each other lead
us to believe that an actual conservative property of the medium
and not an effect of this on the luminous paint is here involved.
This conservative effect of the gelatin may be concerned with its
efficacy as a sensitiser.

In the views I have laid before you I have endeavoured to show
that the recent addition to our knowledge of the electron as an
entity taking part in many physical and chemical effects should
be kept in sight in seeking an explanation of the mode of origin
of the latent image.[1]

[1] For a more detailed account of the subject, and some
ingenious extensions of the views expressed above, see
_Photo-Electricity_, by H. Stanley Allen: Longmans, Green & Ca.,
1913.

213

PLEOCHROIC HALOES [1]

IT is now well established that a helium atom is expelled from
certain of the radioactive elements at the moment of
transformation. The helium atom or alpha ray leaves the
transforming atom with a velocity which varies in the different
radioactive elements, but which is always very great, attaining
as much as 2 x 109 cms. per second; a velocity which, if
unchecked, would carry the atom round the earth in less than two
seconds. The alpha ray carries a positive charge of double the
ionic amount.

When an alpha ray is discharged from the transforming element
into a gaseous medium its velocity is rapidly checked and its
energy absorbed. A certain amount of energy is thus transferred
from the transforming atom to the gas. We recognise this energy
in the gas by the altered properties of the latter; chiefly by
the fact that it becomes a conductor of electricity. The
mechanism by which this change is effected is in part known. The
atoms of the gas, which appear to be freely penetrated by the
alpha ray, are so far dismembered as to yield charged electrons
or ions; the atoms remaining charged with an equal and opposite
charge. Such a medium of

[1] Being the Huxley Lecture, delivered at the University of
Birmingham on October 30th, 1912. Bedrock, Jan., 1913.

214

free electric charges becomes a conductor of electricity by
convection when an electromotive force is applied. The gas also
acquires other properties in virtue of its ionisation. Under
certain conditions it may acquire chemical activity and new
combinations may be formed or existing ones broken up. When its
initial velocity is expended the helium atom gives up its
properties as an alpha ray and thenceforth remains possessed of
the ordinary varying velocity of thermal agitation. Bragg and
Kleeman and others have investigated the career of the alpha ray
when its path or range lies in a gas at ordinary or obtainable
conditions of pressure and temperature. We will review some of
the facts ascertained.

The range or distance traversed in a gas at ordinary pressures is
a few centimetres. The following table, compiled by Geiger, gives
the range in air at the temperature of 15° C.:

               cms.                   cms.                   cms.
Uranium 1 -   2.50    Thorium -     2.72     Radioactinium  4.60
Uranium 2 -   2.90    Radiothorium  3.87     Actinium X -   4.40
Ionium -      3.00    Thorium X -   4.30     Act Emanation  5.70
Radium -      3.30    Th Emanation  5.00     Actinium A -   6.50
Ra Emanation  4.16    Thorium A -   5.70     Actinium C -   5.40
Radium A -    4.75    Thorium C1 -  4.80
Radium C -    6.94    Thorium C2 -  8.60
Radium F -    3.77

It will be seen that the ray of greatest range is that proceeding
from thorium C2, which reaches a distance of 8.6 cms. In the
uranium family the fastest ray is

215

that of radium C. It attains 6.94 cms. There is thus an
appreciable difference between the ultimate distances traversed
by the most energetic rays of the two families. The shortest
ranges are those of uranium 1 and 2.

The ionisation effected by these rays is by no means uniform
along the path of the ray. By examining the conductivity of the
gas at different points along the path of the ray, the ionisation
at these points may be determined. At the limits of the range the
ionisation

{Fig. 13}

ceases. In this manner the range is, in fact, determined. The
dotted curve (Fig. 13) depicts the recent investigation of the
ionisation effected by a sheaf of parallel rays of radium C in
air, as determined by Geiger. The range is laid out horizontally
in centimetres. The numbers of ions are laid out vertically. The
remarkable nature of the results will be at once apparent. We
should have expected that the ray at the beginning of its path,
when its velocity and kinetic energy were greatest, would have
been more effective than towards the end of its range

216

when its energy had almost run out. But the curve shows that it
is just the other way. The lagging ray, about to resign its
ionising properties, becomes a much more efficient ioniser than
it was at first. The maximum efficiency is, however, in the case
of a bundle of parallel rays, not quite at the end of the range,
but about half a centimetre from it. The increase to the maximum
is rapid, the fall from the maximum to nothing is much more
rapid.

It can be shown that the ionisation effected anywhere along the
path of the ray is inversely proportional to the velocity of the
ray at that point. But this evidently does not apply to the last
5 or 10 mms. of the range where the rate of ionisation and of the
speed of the ray change most rapidly. To what are the changing
properties of the rays near the end of their path to be ascribed?
It is only recently that this matter has been elucidated.

When the alpha ray has sufficiently slowed down, its power of
passing right through atoms, without appreciably experiencing any
effects from them, diminishes. The opposing atoms begin to exert
an influence on the path of the ray, deflecting it a little. The
heavier atoms will deflect it most. This effect has been very
successfully investigated by Geiger. It is known as "scattering."
The angle of scattering increases rapidly with the decrease of
velocity. Now the effect of the scattering will be to cause some
of the rays to complete their ranges

217

or, more accurately, to leave their direct line of advance a
little sooner than others. In the beautiful experiments of C. T.
R. Wilson we are enabled to obtain ocular demonstration of the
scattering. The photograph (Fig. 14.), which I owe to the
kindness of Mr. Wilson, shows the deflection of the ray towards
the end of its path. In

{Fig. 14}

this case the path of the ray has been rendered visible by the
condensation of water particles under the influence of the
ionisation; the atmosphere in which the ray travels being in a
state of supersaturation with water vapour at the instant of the
passage of the ray. It is evident that if we were observing the
ionisation along a sheaf of parallel rays, all starting with
equal velocity,

218

the effect of the bending of some of the rays near the end of
their range must be to cause a decrease in the aggregate
ionisation near the very end of the ultimate range. For, in fact,
some of the rays complete their work of ionising at points in the
gas before the end is reached. This is the cause, or at least an
important contributory cause, of the decline in the ionisation
near the end of the range, when the effects of a bundle of rays
are being observed. The explanation does not suggest that the
ionising power of any one ray is actually diminished before it
finally ceases to be an alpha ray.

The full line in Fig. 13 gives the ionisation curve which it may
be expected would be struck out by a single alpha ray. In it the
ionisation goes on increasing till it abruptly ceases altogether,
with the entire loss of the initial kinetic energy of the
particle.

A highly remarkable fact was found out by Bragg. The effect of
the atom traversed by the ray in checking the velocity of the ray
is independent of the physical and chemical condition of the
atom. He measured the "stopping power" of a medium by the
distance the ray can penetrate into it compared with the distance
to which it can penetrate in air. The less the ratio the greater
is the stopping power. The stopping power of a substance is
proportional to the square root of its atomic weight. The
stopping power of an atom is not altered if it is in chemical
union with another atom. The atomic weight is the one quality of
importance. The physical

219

state, whether the element is in the solid, liquid or gaseous
state, is unimportant. And when we deal with molecules the
stopping power is simply proportional to the sum of the square
roots of the atomic weights of the atoms entering into the
molecule. This is the "additive law," and it obviously enables us
to calculate what the range in any substance of known chemical
composition and density will be, compared with the range in air.

This is of special importance in connection with phenomena we
have presently to consider. It means that, knowing the chemical
composition and density of any medium whatsoever, solid, liquid
or gaseous, we can calculate accurately the distance to which any
particular alpha ray will penetrate. Nor have the temperature and
pressure to which the medium is subjected any influence save in
so far as they may affect the proximity of one atom to another.
The retardation of the alpha ray in the atom is not affected.

This valuable additive law, however, cannot be applied in
strictness to the amount of ionisation attending the ray. The
form of the molecule, or more generally its volume, may have an
influence upon this. Bragg draws the conclusion, from this fact
as well as from the notable increase of ionisation with loss of
speed, that the ionisation is dependent upon the time the ray
spends in the molecule. The energy of the ray is, indeed, found
to be less efficient in producing ionisation in the smaller
atomm.

220

Before leaving our review of the general laws governing the
passage of alpha rays through matter, another point of interest
must be referred to. We have hitherto spoken in general terms of
the fact that ionisation attends the passage of the ray. We have
said nothing as to the nature of the ionisation so produced. But
in point of fact the ionisation due to an alpha ray is sui
generis. A glance at one of Wilson's photographs (Fig. 14.)
illustrates this. The white streak of water particles marks the
path of the ray. The ions produced are evidently closely crowded
along the track of the ray. They have been called into existence
in a very minute instant of time. Now we know that ions of
opposite sign if left to themselves recombine. The rate of
recombination depends upon the product of the number of each sign
present in unit volume. Here the numbers are very great and the
volume very small. The ionic density is therefore high, and
recombination very rapidly removes the ions after they are
formed. We see here a peculiarity of the ionisation effected by
alpha rays. It is linear in distribution and very local. Much of
the ionisation in gases is again undone by recombination before
diffusion leads to the separation of the ions. This "initial
recombination" is greatest towards the end of the path of the ray
where the ionisation is a maximum. Here it may be so effective
that the form of the curve is completely lost unless a very large
electromotive force is used to separate the ions when the
ionisation is being investigated.

221

We have now reviewed recent work at sufficient length to
understand something of the nature of the most important advance
ever made in our knowledge of the atom. Let us glance briefly at
what we have learned. The radioactive atom in sinking to a lower
atomic weight casts out with enormous velocity an atom of helium.
It thus loses a definite portion of its mass and of its energy.
Helium which is chemically one of the most inert of the elements,
is, when possessed of such great kinetic energy, able to
penetrate and ionise the atoms which it meets in its path. It
spends its energy in the act of ionising them, coming to rest,
when it moves in air, in a few centimetres. Its initial velocity
depends upon the particular radioactive element which has given
rise to it. The length of its path is therefore different
according to the radioactive element from which it proceeds. The
retardation which it experiences in its path depends entirely
upon the atomic weight of the atoms which it traverses. As it
advances in its path its effectiveness in ionising the atom
rapidly increases and attains a very marked maximum. In a gas the
ions produced being much crowded together recombine rapidly; so
rapidly that the actual ionisation may be quite concealed unless
a sufficiently strong electric force is applied to separate them.
Such is a brief summary of the climax of radioactive
discovery:--the birth, life and death of the alpha ray. Its advent
into Science has altered fundamentally our conception of

222

matter. It is fraught with momentous bearings upon Geological
Science. How the work of the alpha ray is sometimes recorded
visibly in the rocks and what we may learn from that record, I
propose now to bring before you.

In certain minerals, notably the brown variety of mica known as
biotite, the microscope reveals minute circular marks occurring
here and there, quite irregularly. The most usual appearance is
that of a circular area darker in colour than the surrounding
mineral. The radii of these little disc-shaped marks when well
defined are found to be remarkably uniform, in some cases four
hundredths of a millimetre and in others three hundredths, about.
These are the measurements in biotite. In other minerals the
measurements are not quite the same as in biotite. Such minute
objects are quite invisible to the naked eye. In some rocks they
are very abundant, indeed they may be crowded together in such
numbers as to darken the colour of the mineral containing them.
They have long been a mystery to petrologists.

Close examination shows that there is always a small speck of a
foreign body at the centre of the circle, and it is often
possible to identify the nature of this central substance, small
though it be. Most generally it is found to be the mineral
zircon. Now this mineral was shown by Strutt to contain radium in
quantities much exceeding those found in ordinary rock
substances.

223

Some other mineral may occasionally form the nucleus, but we
never find any which is not known to be specially likely to
contain a radioactive substance. Another circumstance we notice.
The smaller this central nucleus the more perfect in form is the
darkened circular area surrounding it. When the circle is very
perfect and the central mineral clearly defined at its centre we
find by measurement that the radius of the darkened area is
generally 0.033 mm. It may sometimes be 0.040 mm. These are
always the measurements in biotite. In other minerals the radii
are a little different.

We see in the photograph (Pl. XXIII, lower figure), much
magnified, a halo contained in biotite. We are looking at a
region in a rock-section, the rock being ground down to such a
thickness that light freely passes through it. The biotite is in
the centre of the field. Quartz and felspar surround it. The rock
is a granite. The biotite is not all one crystal. Two crystals,
mutually inclined, are cut across. The halo extends across both
crystals, but owing to the fact that polarised light is used in
taking the photograph it appears darker in one crystal than in
the other. We see the zircon which composes the nucleus. The fine
striated appearance of the biotite is due to the cleavage of that
mineral, which is cut across in the section.

The question arises whether the darkened area surrounding the
zircon may not be due to the influence of the radioactive
substances contained in the zircon. The

224

extraordinary uniformity of the radial measurements of perfectly
formed haloes (to use the name by which they have long been
known) suggests that they may be the result of alpha radiation.
For in that case, as we have seen, we can at once account for the
definite radius as simply representing the range of the ray in
biotite. The furthest-reaching ray will define the radius of the
halo. In the case of the uranium family this will be radium C,
and in the case of thorium it will be thorium C. Now here we
possess a means of at once confirming or rejecting the view that
the halo is a radioactive phenomenon and occasioned by alpha
radiation; for we can calculate what the range of these rays will
be in biotite, availing ourselves of Bragg's additive law,
already referred to. When we make this calculation we find that
radium C just penetrates 0.033 mm. and thorium C 0.040 mm. The
proof is complete that we are dealing with the effects of alpha
rays. Observe now that not only is the coincidence of measurement
and calculation a proof of the view that alpha radiation has
occasioned the halo, but it is a very complete verification of
the important fact stated by Bragg, that the stopping power
depends solely on the atomic weight of the atoms traversed by the
ray.

We have seen that our examination of the rocks reveals only the
two sorts of halo: the radium halo and the thorium halo. This is
not without teaching. For why not find an actinium halo? Now
Rutherford long ago suggested that this element and its
derivatives were

225

probably an offspring of the uranium family; a side branch, as it
were, in the formation of which relatively few transforming atoms
took part. On Rutherford's theory then, actinium should always
accompany uranium and radium, but in very subordinate amount. The
absence of actinium haloes clearly supports this view. For if
actinium was an independent element we would be sure to find
actinium haloes. The difference in radius should be noticeable.
If, on the other hand, actinium

was always associated with uranium and radium, then its effects
would be submerged in those of the much more potent effects of
the uranium series of elements.

It will have occurred to you already that if the radioactive
origin of the halo is assured the shape of a halo is not really
circular, but spherical. This is so. There is no such thing as a
disc-shaped halo. The halo is a spherical volume containing the
radioactive nucleus at its centre. The true radius of the halo
may, therefore, only be measured on sections passing through the
nucleus.

226

In order to understand the mode of formation of a halo we may
profitably study on a diagram the events which go on within the
halo-sphere. Such a diagram is seen in Fig. 15. It shows to
relatively correct scale the limiting range of all the alpha-ray
producing members of the uranium and thorium families. We know
that each member of a family will exist in equilibrium amount
within the nucleus possessing the parent element. Each alpha ray
leaving the nucleus will just attain its range and then cease to
affect the mica. Within the halosphere, there must be, therefore,
the accumulated effects of the influences of all the rays. Each
has its own sphere of influence, and the spheres are all
concentric.

The radii in biotite of the several spheres are given in the
following table

URANIUM FAMILY.
Radium C -       0.0330 mm.
Radium A -       0.0224 mm.
Ra Emanation -   0.0196 mm.
Radium F -       0.0177 mm.
Radium -         0.0156 mm.
Ionium -         0.0141 mm.
Uranium 1 -      0.0137 mm.
Uranium 2 -      0.0118 mm.

THORIUM FAMILY.
Thorium CE -     0.040 mm.
Thorium A -      0.026 mm.
Th Emanation -   0.023 mm.
Thorium Ci -     0.022 mm.
Thorium X -      0.020 mm.
Radiothorium -   0.119 mm.
Thorium -        0.013 mm.

In the photograph (Pl. XXIV, lower figure), we see a uranium and
a thorium halo in the same crystal of mica. The mica is contained
in a rock-section and is cut across the cleavage. The effects of
thorium Ca are clearly shown

227

as a lighter border surrounding the accumulated inner darkening
due to the other thorium rays. The uranium halo (to the right)
similarly shows the effects of radium C, but less distinctly.

Haloes which are uniformly dark all over as described above are,
in point of fact, "over-exposed"; to borrow a familiar
photographic term. Haloes are found which show much beautiful
internal detail. Too vigorous action obscures this detail just as
detail is lost in an over-exposed photograph. We may again have
"under-exposed" haloes in which the action of the several rays is
incomplete or in which the action of certain of the rays has left
little if any trace. Beginning at the most under-exposed haloes
we find circular dark marks having the radius 0.012 or 0.013 mm.
These haloes are due to uranium, although their inner darkening
is doubtless aided by the passage of rays which were too few to
extend the darkening beyond the vigorous effects of the two
uranium rays. Then we find haloes carried out to the radii 0.016,
0.018 and 0.019 mm. The last sometimes show very beautiful outer
rings having radial dimensions such as would be produced by
radium A and radium C. Finally we may have haloes in which
interior detail is lost so far out as the radius due to emanation
or radium A, while outside this floats the ring due to radium C.
Certain variations of these effects may occur, marking,
apparently, different stages of exposure. Plates XXIII and XXIV
(upper figure) illustrate some of these stages;

228

the latter photograph being greatly enlarged to show clearly the
halo-sphere of radium A.

In most of the cases mentioned above the structure evidently
shows the existence of concentric spherical shells of darkened
biotite. This is a very interesting fact. For it proves that in
the mineral the alpha ray gives rise to the same increased
ionisation towards the end of its range, as Bragg determined in
the case of gases. And we must conclude that the halo in every
case grows in this manner. A spherical shell of darkened biotite
is first produced and the inner colouration is only effected as
the more feeble ionisation along the track of the ray in course
of ages gives rise to sufficient alteration of the mineral. This
more feeble ionisation is, near the nucleus, enhanced in its
effects by the fact that there all the rays combine to increase
the ionisation and, moreover, the several tracks are there
crowded by the convergency to the centre. Hence the most
elementary haloes seldom show definite rings due to uranium,
etc., but appear as embryonic disc-like markings. The photographs
illustrate many of the phases of halo development.

Rutherford succeeded in making a halo artificially by compressing
into a capillary glass tube a quantity of the emanation of
radium. As the emanation decayed the various derived products
came into existence and all the several alpha rays penetrated the
glass, darkening the walls of the capillary out to the limit of
the range of radium C in glass. Plate XXV shows a magnified
section of the

229

tube. The dark central part is the capillary. The tubular halo
surrounds it. This experiment has, however, been anticipated by
some scores of millions of years, for here is the same effect in
a biotite crystal (Pl. XXV). Along what are apparently tubular
passages or cracks in the mica, a solution, rich in radioactive
substances, has moved; probably during the final consolidation of
the granite in which the mica occurs. A continuous and very
regular halo has developed along these conduits. A string of
halo-spheres may lie along such passages. We must infer that
solutions or gases able to establish the radioactive nuclei moved
along these conduits, and we are entitled to ask if all the
haloes in this biotite are not, in this sense, of secondary
origin. There is, I may add, much to support such a conclusion.

The widespread distribution of radioactive substances is most
readily appreciated by examination of sections of rocks cut thin
enough for microscopic investigation. It is, indeed, difficult to
find, in the older rocks of granitic type, mica which does not
show haloes, or traces of haloes. Often we find that every one of
the inclusions in the mica--that is, every one of the earlier
formed substances--contain radioactive elements, as indicated by
the presence of darkened borders. As will be seen presently the
quantities involved are generally vanishingly small. For example
it was found by direct determination that in one gram of the
halo-rich mica of Co. Carlow there was rather less than twelve
billionths of a gram of radium, We are

230

entitled to infer that other rare elements are similarly widely
distributed but remain undetectable because of their more stable
properties.

It must not be thought that the under-exposed halo is a recent
creation. By no means. All are old, appallingly old; and in the
same rock all are, probably, of the same, or neatly the same,
age. The under-exposure is simply due to a lesser quantity of the
radioactive elements in the nucleus. They are under-exposed, in
short, not because of lesser duration of exposure, but because of
insufficient action; as when in taking a photograph the stop is
not open enough for the time of the exposure.

The halo has, so far, told us that the additive law is obeyed in
solid media, and that the increased ionisation attending the
slowing down of the ray obtaining in gases, also obtains in
solids; for, otherwise, the halo would not commence its
development as a spherical shell or envelope. But here we learn
that there is probably a certain difference in the course of
events attending the immediate passage of the ray in the gas and
in the solid. In the former, initial recombination may obscure
the intense ionisation near the end of the range. We can only
detect the true end-effects by artificially separating the ions
by a strong electric force. If this recombination happened in the
mineral we should not have the concentric spheres so well defined
as we see them to be. What, then, hinders the initial
recombination in the solid? The answer probably is that the newly
formed

231

ion is instantly used up in a fresh chemical combination. Nor is
it free to change its place as in the gas. There is simply a new
equilibrium brought about by its sudden production. In this
manner the conditions in the complex molecule of biotite,
tourmaline, etc., may be quite as effective in preventing initial
recombination as the most effective electric force we could
apply. The final result is that we find the Bragg curve
reproduced most accurately in the delicate shading of the rings
making up the perfectly exposed halo.

That the shading of the rings reproduces the form of the Bragg
curve, projected, as it were, upon the line of advance of the ray
and reproduced in depth of shading, shows that in yet another
particular the alpha ray behaves much the same in the solid as in
the gas. A careful examination of the outer edge of the circles
always reveals a steep but not abrupt cessation of the action of
the ray. Now Geiger has investigated and proved the existence of
scattering of the alpha ray by solids. We may, therefore, suppose
with much probability that there is the same scattering within
the mineral near the end of the range. The heavy iron atom of the
biotite is, doubtless, chiefly responsible for this in biotite
haloes. I may observe that this shading of the outer bounding
surface of the sphere of action is found however minute the
central nucleus. In the case of a nucleus of considerable size
another effect comes in which tends to produce an enhanced
shading. This will

232

result if rays proceed from different depths in the nucleus. If
the nucleus were of the same density and atomic weight as the
surrounding mica, there would be little effect. But its density
and molecular weight are generally greater, hence the retardation
is greater, and rays proceeding from deep in the nucleus
experience more retardation than those which proceed from points
near to the surface. The distances reached by the rays in the
mica will vary accordingly, and so there will be a gradual
cessation of the effects of the rays.

The result of our study of the halo may be summed up in the
statement that in nearly every particular we have the phenomena,
which have been measured and observed in the gas, reproduced on a
minute scale in the halo. Initial recombination seems, however,
to be absent or diminished in effectiveness; probably because of
the new stability instantly assumed by the ionised atoms.

One of the most interesting points about the halo remains to be
referred to. The halo is always uniformly darkened all round its
circumference and is perfectly spherical. Sections, whether taken
in the plane of cleavage of the mica or across it, show the same
exactly circular form, and the same radius. Of course, if there
was any appreciable increase of range along or across the
cleavage the form of the halo on the section across the cleavage
should be elliptical. The fact that there is no measurable
ellipticity is, I think, one which on first consideration would
not be expected.

233

For what are the conditions attending the passage of the ray in a
medium such as mica? According to crystallographic conceptions we
have here an orderly arrangement of molecules, the units
composing the crystal being alike in mass, geometrically spaced,
and polarised as regards the attractions they exert one upon
another. Mica, more especially, has the cleavage phenomenon
developed to a degree which transcends its development in any
other known substance. We can cleave it and again cleave it till
its flakes float in the air, and we may yet go on cleaving it by
special means till the flakes no longer reflect visible light.
And not less remarkable is the uniplanar nature of its cleavage.
There is little cleavage in any plane but the one, although it is
easy to show that the molecules in the plane of the flake are in
orderly arrangement and are more easily parted in some directions
than in others. In such a medium beyond all others we must look
with surprise upon the perfect sphere struck out by the alpha
rays, because it seems certain that the cleavage is due to lesser
attraction, and, probably, further spacing of the molecules, in a
direction perpendicular to the cleavage.

It may turn out that the spacing of the molecules will influence
but little the average number per unit distance encountered by
rays moving in divergent paths. If this is so, we seem left to
conclude that, in spite of its unequal and polarised attractions,
there is equal retardation and equal ionisation in the molecule
in whatever

234

direction it is approached. Or, again, if the encounters indeed
differ in number, then some compensating effect must exist
whereby a direction of lesser linear density involves greater
stopping power in the molecule encountered, and vice versa.

The nature of the change produced by the alpha rays is unknown.
But the formation of the halo is not, at least in its earlier
stages, attended by destruction of the crystallographic and
optical properties of the medium. The optical properties are
unaltered in nature but are increased in intensity. This applies
till the halo has become so darkened that light is no longer
transmitted under the conditions of thickness obtaining in rock
sections. It is well known that there is in biotite a maximum
absorption of a plane-polarised light ray, when the plane of
vibration coincides with the plane of cleavage. A section across
the cleavage then shows a maximum amount of absorption. A halo
seen on this section simply produces this effect in a more
intense degree. This is well shown in Plate XXIII (lower figure),
on a portion of the halo-sphere. The descriptive name "Pleochroic
Halo" has originated from this fact. We must conclude that the
effect of the ionisation due to the alpha ray has not been to
alter fundamentally the conditions which give rise to the optical
properties of the medium. The increased absorption is probably
associated with some change in the chemical state of the iron
present. Haloes are, I believe, not found in minerals from which
this

235

element is absent. One thing is quite certain. The colouration is
not due to an accumulation of helium atoms, _i.e._ of spent alpha
rays. The evidence for this is conclusive. If helium was
responsible we should have haloes produced in all sorts of
colourless minerals. Now we sometimes see zircons in felspars and
in quartz, etc., but in no such case is a halo produced. And
halo-spheres formed within and sufficiently close to the edge of
a crystal of mica are abruptly truncated by neighbouring areas of
fclspar or quartz, although we know that the rays must pass
freely across the boundary. Again it is easy to show that even in
the oldest haloes the quantity of helium involved is so small
that one might say the halo-sphere was a tolerably good vacuum as
regards helium. There is, finally, no reason to suppose that the
imprisoned helium would exhibit such a colouration, or, indeed,
any at all.

I have already referred to the great age of the halo. Haloes are
not found in the younger igneous rocks. It is probable that a
halo less than a million years old has never been seen. This,
primâ facie, indicates an extremely slow rate of formation. And
our calculations quite support the conclusions that the growth of
a halo, if this has been uniform, proceeds at a rate of almost
unimaginable slowness.

Let us calculate the number of alpha rays which may have gone to
form a halo in the Devonian granite of Leinster.

236

It is common to find haloes developed perfectly in this granite,
and having a nucleus of zircon less than 5 x 10-4 cms. in
diameter. The volume of zircon is 65 x 10-12 c.cs. and the mass
3 x 10-10 grm.; and if there was in this zircon 10-8 grm. radium
per gram (a quantity about five times the greatest amount
measured by Strutt), the mass of radium involved is 3 x 10-18
grm. From this and from the fact ascertained by Rutherford that
the number of alpha rays expelled by a gram of radium in one
second is 3.4 x 1010, we find that three rays are shot from the
nucleus in a year. If, now, geological time since the Devonian is
50 millions of years, then 150 millions of rays built up the
halo. If geological time since the Devonian is 400 millions of
years, then 1,200 millions of alpha rays are concerned in its
genesis. The number of ions involved, of course, greatly exceeds
these numbers. A single alpha ray fired from radium C will
produce 2.37 x 105 ions in air.

But haloes may be found quite clearly defined and fairly dark out
to the range of the emanation ray and derived from much less
quantities of radioactive materials. Thus a zircon nucleus with a
diameter of but 3.4 x 10-4 cms. formed a halo strongly darkened
within, and showing radium A and radium C as clear smoky rings.
Such a nucleus, on the assumption made above as to its radium
content, expels one ray in a year. But, again, haloes are
observed with less blackened pupils and with faint ring due to
radium C, formed round nuclei

237

of rather less than 2 x 10-4 cms. diameter. Such nuclei would
expel one ray in five years. And even lesser nuclei will generate
in these old rocks haloes with their earlier characteristic
features clearly developed. In the case of the most minute
nuclei, if my assumption as to the uranium content is correct, an
alpha ray is expelled, probably, no oftener than once in a
century; and possibly at still longer intervals.

The equilibrium amount of radium contained in some nuclei may
amount to only a few atoms. Even in the case of the larger nuclei
and more perfectly developed haloes the quantity of radium
involved is many millions of times less than the least amount we
can recognise by any other means. But the delicacy of the
observation is not adequately set forth in this statement. We can
not only tell the nature of the radioactive family with which we
are dealing; but we can recognise the presence of some of its
constituent members. I may say that it is not probable the
zircons are richer in radium than I have assumed. My assumption
involves about 3 per cent. of uranium. I know of no analyses
ascribing so great an amount of uranium to zircon. The variety
cyrtolite has been found to contain half this amount, about. But
even if we doubled our estimate of radium content, the remarkable
nature of our conclusions is hardly lessened.

It may appear strange that the ever-interesting question of the
Earth's age should find elucidation from the

238

study of haloes. Nevertheless the subjects are closely connected.
The circumstances are as follows. Geologists have estimated the
age of the Earth since denudation began, by measurements of the
integral effects of denudation. These methods agree in showing an
age of about rob years. On the other hand, measurements have been
made of the accumulation in minerals of radioactive _débris_--the
helium and lead--and results obtained which, although they do not
agree very well among themselves, are concordant in assigning a
very much greater age to the rocks. If the radioactive estimate
is correct, then we are now living in a time when the denudative
forces of the Earth are about eight or nine times as active as
they have been on the average over the past. Such a state of
things is absolutely unaccountable. And all the more
unaccountable because from all we know we would expect a somewhat
_lesser_ rate of solvent denudation as the world gets older and the
land gets more and more loaded with the washed-out materials of
the rocks.

Both the methods referred to of finding the age assume the
principle of uniformity. The geologist contends for uniformity
throughout the past physical history of the Earth. The physicist
claims the like for the change-rates of the radioactive elements.
Now the study of the rocks enables us to infer something as to
the past history of our Globe. Nothing is, on the other hand,
known respecting the origin of uranium or thorium--the parent
radioactive bodies. And while not questioning the law

239

and regularity which undoubtedly prevail in the periods of the
members of the radioactive families, it appears to me that it is
allowable to ask if the change rate of uranium has been always
what we now believe it to be. This comes to much the same thing
as supposing that atoms possessing a faster change rate once were
associated with it which were capable of yielding both helium and
lead to the rocks. Such atoms might have been collateral in
origin with uranium from some antecedent element. Like helium,
lead may be a derivative from more than one sequence of
radioactive changes. In the present state of our knowledge the
possibilities are many. The rate of change is known to be
connected with the range of the alpha ray expelled by the
transforming element; and the conformity of the halo with our
existing knowledge of the ranges is reason for assuming that,
whatever the origin of the more active associate of uranium, this
passed through similar elemental changes in the progress of its
disintegration. There may, however, have been differences in the
ranges which the halo would not reveal. It is remarkable that
uranium at the present time is apparently responsible for two
alpha rays of very different ranges. If these proceed from
different elements, one should be faster in its change rate than
the other. Some guidance may yet be forthcoming from the study of
the more obscure problems of radioactivity.

Now it is not improbable that the halo may contribute directly to
this discussion. We can evidently attack

240

the biotite with a known number of alpha rays and determine how
many are required to produce a certain intensity of darkening,
corresponding to that of a halo with a nucleus of measurable
dimensions. On certain assumptions, which are correct within
defined limits, we can calculate, as I have done above, the
number of rays concerned in forming the halo. In doing so we
assume some value for the age of the halo. Let us take the
maximum radioactive value. A halo originating in Devonian times
may attain a certain central blackening from the effects of, say,
rob rays. But now suppose we find that we cannot produce the same
degree of blackening with this number of rays applied in the
laboratory. What are we to conclude? I think there is only the
one conclusion open to us; that some other source of alpha rays,
or a faster rate of supply, existed in the past. And this
conclusion would explain the absence of haloes from the younger
rocks; which, in view of the vast range of effects possible in
the development of haloes, is, otherwise, not easy to account
for. It is apparent that the experiment on the biotite has a
direct bearing on the validity of the radioactive method of
estimating the age of the rocks. It is now being carried out by
Professor Rutherford under reliable conditions.

Finally, there is one very certain and valuable fact to be
learned from the halo. The halo has established the extreme
rarity of radioactivity as an atomic phenomenon. One and all of
the speculations as to

241

the slow breakdown of the commoner elements may be dismissed. The
halo shows that the mica of the rocks is radioactively sensitive.
The fundamental criterion of radioactive change is the expulsion
of the alpha ray. The molecular system of the mica and of many
other minerals is unstable in presence of these rays, just as a
photographic plate is unstable in presence of light. Moreover,
the mineral integrates the radioactive effects in the same way as
a photographic salt integrates the effects of light. In both
cases the feeblest activities become ultimately apparent to our
inspection. We have seen that one ray in each year since the
Devonian period will build the fully formed halo: an object
unlike any other appearance in the rocks. And we have been able
to allocate all the haloes so far investigated to one or the
other of the known radioactive families. We are evidently
justified in the belief that had other elements been radioactive
we must either find characteristic haloes produced by them, or
else find a complete darkening of the mica. The feeblest alpha
rays emitted by the relatively enormous quantities of the
prevailing elements, acting over the whole duration of geological
time--and it must be remembered that the haloes we have been
studying are comparatively young--must have registered their
effects on the sensitive minerals. And thus we are safe in
concluding that the common elements, and, indeed, many which
would be called rare, are possessed of a degree of stability
which has preserved them un

242

changed since the beginning of geological time. Each unaffected
flake of mica is, thus, unassailable proof of a fact which but
for the halo would, probably, have been for ever beyond our
cognisance.

THE USE OF RADIUM IN MEDICINE [1]

IT has been unfortunate for the progress of the radioactive
treatment of disease that its methods and claims involve much of
the marvellous. Up till recently, indeed, a large part of
radioactive therapeutics could only be described as bordering on
the occult. It is not surprising that when, in addition to its
occult and marvellous characters, claims were made on its behalf
which in many cases could not be supported, many medical men came
to regard it with a certain amount of suspicion.

Today, I believe, we are in a better position. I think it is
possible to ascribe a rational scientific basis to its legitimate
claims, and to show, in fact, that in radioactive treatment we
are pursuing methods which have been already tried extensively
and found to be of definite value; and that new methods differ
from the old mainly in their power and availability, and little,
or not at all, in kind.

Let us briefly review the basis of the science. Radium is a
metallic element chemically resembling barium. It

[1] A Lecture to Postgraduate Students of Medicine in connection
with the founding of the Dublin Radium Institute, delivered in
the School of Physic in Ireland, Trinity College, on October 2nd,
1914

244

possesses, however, a remarkable property which barium does not.
Its atoms are not equally stable. In a given quantity of radium a
certain very small percentage of the total number of atoms
present break up per second. By "breaking up" we mean their
transmutation to another element. Radium, which is a solid
element under ordinary conditions, gives rise by transmutation to
a gaseous element--the emanation of radium. The new element is a
heavy gas at ordinary temperatures and, like other gases, can be
liquified by extreme cold. The extraordinary property of
transmutation is entirely automatic. No influence which chemist
or physicist can apply can affect the rate of transformation.

The emanation inherits the property of instability, but in its
case the instability is more pronounced. A relatively large
fraction of its atoms transmute per second to a solid element
designated Radium A. In turn this new generation of atoms breaks
up--even faster than the emanation--becoming yet another element
with specific chemical properties. And so on for a whole sequence
of transmutations, till finally a stable substance is formed,
identical with ordinary lead in chemical and physical properties,
but possessing a slightly lower atomic weight.

The genealogy of the radium series of elements shows that radium
is not the starting point. It possesses ancestors which have been
traced back to the element uranium.

Now what bearing has this series of transmutations

245

upon medical science? Radium or emanation, &c., are not in the
Pharmacopoeia as are, say, arsenic or bismuth. The whole
medicinal value of these elements resides in the very wonderful
phenomena of their radiations. They radiate in the process of
transmuting.

The changing atom may radiate a part of its own mass. The
"alpha"-ray (a-ray) is such a material ray. It is an electrified
helium atom cast out of the parent atom with enormous
velocity--such a velocity as would carry it, if not impeded, all
round the earth in two seconds. All alpha-rays are positively
electrified atoms of the element helium, which thereby is shown
to be an integral constituent of many elements. The alpha-ray is
not of much value to medical science, for, in spite of its great
velocity, it is soon stopped by encounter with other atoms. It
can penetrate only a minute fraction of a millimetre into
ordinary soft tissues. We shall not further consider it.

Transmuting atoms give out also material rays of another kind:
the ß-rays. The ß-ray is in mass but a very small fraction of,
even, a hydrogen atom. Its speed may approach that of light. As
cast out by radioactive elements it starts with speeds which vary
with the element, and may be from one-third to nine-tenths the
velocity of light. The ß-ray is negatively electrified. It has
long been known to science as the electron. It is also identical
with the cathode ray of the vacuum tube.

246

Another and quite different kind of radiation is given out by
many of the transmuting elements:--the y-ray. This is not
material, it is ethereal. It is known now with certainty that the
y-ray is in kind identical with light, but of very much shorter
wave length than even the extreme ultraviolet light of the solar
spectrum. The y-ray is flashed from the transmuting atom along
with the ß-ray. It is identical in character with the x-ray but
of even shorter wave length.

There is a very interesting connection between the y-ray and the
ß-ray which it is important for the medical man to understand--as
far as it is practicable on our present knowledge.

When y-rays or x-rays fall on matter they give rise to ß-rays.
The mechanism involved is not known but it is possibly a result
of the resonance of the atom, or of parts of it, to the short
light waves. And it is remarkable that the y-rays which, as we
have seen, are shorter and more penetrating waves than the
x-rays, give rise to ß-rays possessed of greater velocity and
penetration than ß-rays excited by the x-rays. Indeed the ß-rays
originated by y-rays may attain a velocity nearly approaching
that of light and as great as that of any ß-rays emitted by
transmuting atoms. Again there is demonstrable evidence that
ß-rays impinging on matter may give rise to y-rays. The most
remarkable demonstration of this is seen in the x-ray tube. Here
the x-rays originate where the stream of ß- or cathode-rays

247

are arrested on the anode. But the first relation is at present
of most importance to us--_i.e._ that the y-or x-rays give rise to
ß-rays.

This relation gives us additional evidence of the identity of the
physical effects of y-, x-, and light-rays --using the term light
rays in the usual sense of spectral rays. For it has long been
known that light waves liberate electrons from atoms. It has been
found that these electrons possess a certain initial velocity
which is the greater the shorter the wave length of the light
concerned in their liberation. The whole science of
"photo-electricity" centres round this phenomenon. The action of
light on the photographic plate, as well as many other physical
and chemical phenomena, find an explanation in this liberation of
the electron by the light wave.

Here, then, we have spectral light waves liberating
electrons--_i.e._ very minute negatively-charged particles, and we
find that, as we use shorter light waves, the initial velocity of
these particles increases. Again, we have x-rays which are far
smaller in wave length than spectral light, liberating much
faster negatively electrified particles. Finally, we have
y-rays--the shortest nether waves of all-liberating negative
particles of the highest velocity known. Plainly the whole series
of phenomena is continuous.

We can now look closer at the actions involved in the therapeutic
influence of the several rays and in

248

this way, also, see further the correlation between what may be
called photo-therapeutics and radioactive therapeutics.

The ß-ray, whether we obtain it directly from the transforming
radioactive atom or whether we obtain it as a result of the
effects of the y- or x-rays upon the atom, is an ionising agent
of wonderful power. What is meant by this? In its physical aspect
this means that the atoms through which it passes acquire free
electric charges; some becoming positive, some negative. This can
only be due to the loss of an electron by the affected atom. The
loss of the small negative charge carried in the electron leaves
the atom positively electrified or creates a positive ion. The
fixing of the wandering electron to a neutral atom creates a
negative ion. Before further consideration of the importance of
the phenomenon of ionisation we must fix in our minds that the
agent, which brings this about, is the ß-ray. There is little
evidence that the y-ray can directly create ions to any large
extent. But the action of liberating high-speed ß-rays results in
the creation of many thousands of ions by each ß-ray liberated.
As an agent in the hands of the medical man we must regard the
y-ray as a light wave of extremely penetrating character, which
creates high-speed ß-rays in the tissues which it penetrates,
these ß-rays being most potent ionising agents. The ß-rays
directly obtained from radioactive atoms assist in the work of
ionisation. ß-rays do not

249

penetrate far from their source. The fastest of them would not
probably penetrate one centimetre in soft tissues.

We must now return to the phenomenon of ionisation. Ionisation is
revealed to observation most conspicuously when it takes place in
a gas. The + and - electric charges on the gas particles endow it
with the properties of a conductor of electricity, the + ions
moving freely in one direction and the - ions in the opposite
direction under an electric potential. But there are effects
brought about by ionisation of more importance to the medical man
than this. The chemist has long come to recognise that in the ion
he is concerned with the inner mechanism of a large number of
chemical phenomena. For with the electrification of the atom
attractive and repulsive forces arise. We can directly show the
chemical effects of the ionising ß-rays. Water exposed to their
bombardment splits up into hydrogen and oxygen. And, again, the
separated atoms may be in part recombined under the influence of
the radiation. Ammonia splits up into hydrogen and nitrogen.
Carbon dioxide forms carbon, carbon monoxide, and oxygen;
hydrochloric acid forms chlorine and hydrogen. In these cases,
also, recombination can be partially effected by the rays.

We can be quite sure that within the complex structure of the
living cell the ionising effects which everywhere accompany the
ß-rays must exert a profound influence. The sequence of chemical
events which as yet seem

250

beyond the ken of science and which are involved in metabolism
cannot fail to be affected. Any, it is not surprising that as the
result of eaperinient it is found that the radiations are agents
which may be used either for the stimulation of the natural
events of growth or used for the actual destruction of the cell.
It is easy to see that the feeble radiation should produce the
one effect, the strong the other. In a similar way by a moderate
light stimulus we create the latent image in the photographic
plate; by an intense light we again destroy this image. The inner
mechanism in this last case can be logically stated.[1]

_There is plainly a true physical basis here for the efficacy of
radioactive treatment and, what is more, we find when we examine
it, that it is in kind not different from that underlying
treatment by spectral radiations. But in degree it is very
different and here is the reason for the special importance of
radioactivity as a therapeutic agent._ The Finsen light is capable
of influencing the soft tissues to a short depth only. The reason
is that the wave length of the light used is too great to pass
without rapid absorption through the tissues; and, further, the
electrons it gives rise to--_i.e._ the ß-rays it liberates--are too
slow-moving to be very efficient ionisers. X-rays penetrate in
some cases quite freely and give rise to much faster and more
powerful ß-rays

[1] See _The Latent Image_, p. 202.

251

than can the Finsen light. But far more penetrating than x-rays
are the y-rays emitted in certain of the radioactive changes.
These give rise to ß-rays having a velocity approximate to that
of light.

The y-rays are, therefore, very penetrating and powerfully
ionising light waves; light waves which are quite invisible to
the eye and can beam right through the tissues of the body. To
the mind's eye only are they visible. And a very wonderful
picture they make. We see the transmuting atom flashing out this
light for an inconceivably short instant as it throws off the
ß-ray. And "so far this little candle throws his beams" in the
complex system of the cells, so far atoms shaken by the rays send
out ß-rays; these in turn are hurled against other atomic
systems; fresh separations of electrons arise and new attractions
and repulsions spring up and the most important chemical changes
are brought about. Our mental picture can claim to be no more
than diagrammatic of the reality. Still we are here dealing with
recognised physical and chemical phenomena, and their description
as "occult" in the derogatory sense is certainly not
justifiable.

Having now briefly reviewed the nature of the rays arising in
radioactive substances and the rationale of their influence, we
must turn to more especially practical considerations.

The Table given opposite shows that radium itself is responsible
for a- and ß-rays only. It happens that

252

Period in whioh ½ element is transformed.

URANIUM 1 & 2 { a 6 } x 109 years.

URANIUM X { a ß } 24.6 days.

IONIUM { a 8 } x 104 years.

RADIUM { a ß } 2 x 102 years.

EMANATION { a } 8.85 days.

RADIUM A { a 8 } minutes.

RADIUM B { ß y } 26.7 minutes.

RADIUM C { a ß y } 13.5 minutes.

RADIUM D { ß } 15 years.

RADIUM E { ß y } 4.8 days.

RADIUM (Polonium) F { a } 140 days.

Table showing the successive generations of the elements of the
Uranium-radium family, the character of their radiations and
their longevity.

253

the ß-rays emitted by radium are very "soft"--_i.e._ slow and
easily absorbed. The a-ray is in no case available for more than
mere surface application. Hence we see that, contrary to what is
generally believed, radium itself is of little direct therapeutic
value. Nor is the next body in succession--the emanation, for it
gives only a-rays. In fact, to be brief, it is not till we come
to Radium B that ß-rays of a relatively high penetrative quality
are reached; and it is not till we come to Radium C that highly
penetrative y-rays are obtained.

It is around this element, Radium C, that the chief medical
importance of radioactive treatment by this family of radioactive
bodies centres. Not only are ß-rays of Radium C very penetrating,
but the y-rays are perhaps the most energetic rays of the, kind
known. Further in the list there is no very special medical
interest.

Now, how can we get a supply of this valuable element Radium C?
We can obtain it from radium itself. For even if radium has been
deprived of its emanation (which is easily done by heating it or
bringing it into solution) in a few weeks we get back the Radium
C. One thing here we must be clear about. With a given quantity
of Radium only a certain definitely limited amount of Radium C,
or of emanation, or any other of the derived bodies, will be
associated. Why is this? The answer is because the several
successive elements are themselves decaying --_i.e._ changing one
into the other. The atomic per-

254

centage of each, which decays in a second, is a fixed quantity
which we cannot alter. Now if we picture radium which has been
completely deprived of its emanation, again accumulating by
automatic transmutation a fresh store of this element, we have to
remember:-- (i) That the rate of creation of emanation by the
radium is practically constant; and (2) that the absolute amount
of the emanation decaying per second increases as the stock of
emanation increases. Finally, when the amount of accumulated
emanation has increased to such an extent that the number of
emanation atoms transmuting per second becomes exactly equal to
the number being generated per second, the amount of emanation
present cannot increase. This is called the equilibrium amount.
If fifteen members are elected steadily each year into a
newly-founded society the number of members will increase for the
first few years; finally, when the losses by death of the members
equal about fifteen per annum the society can get no bigger. It
has attained the equilibrium number of members.

This applies to every one of the successive elements. It takes
twenty-one days for the equilibrium quantity of emanation to be
formed in radium which has been completely de-emanated; and it
takes 3.8 days for half the equilibrium amount to be formed.
Again, if we start with a stock of emanation it takes just three
hours for the equilibrium amount of Radium C to be formed.

255

We can evidently grow Radium C either from radium itself or from
the emanation of radium. If we use a tube of radium we have an
almost perfectly constant quantity of Radium C present, for as
fast as the Radium C and intervening elements decay the Radium,
which only diminishes very slowly in amount, makes up the loss.
But, if we start off with a tube of emanation, we do not possess
a constant supply of Radium C, because the emanation is decaying
fairly rapidly and there is no radium to make good its loss. In
3.8 days about one half the emanation is transmuted and the
Radium C decreases proportionately and, of course, with the
Radium C the valuable radiations also decrease. In another 3.8
days--that is in about a week from the start--the radioactive value
of the tube has fallen to one-fourth of its original value.

But in spite of the inconstant character of the emanation tube
there are many reasons for preferring its use to the use of the
radium tube. Chief of these is the fact that we can keep the
precious radium safely locked up in the laboratory and not
exposed to the thousand-and-one risks of the hospital. Then,
secondly, the emanation, being a gas, is very convenient for
subdivision into a large number of very small tubes according to
the dosage required.

In fact the volume of the emanation is exceedingly minute. The
amount of emanation in equilibrium with one gramme of radium is
called the curie, and with one

256

milligramme the millicurie. Now, the volume of the curie is only
a little more than one half a cubic millimetre. Hence in dealing
with emanation from twenty or forty milligrammes of radium we are
dealing with very small volumes.

How may the emanation be obtained? The process is an easy one in
skilled and practised hands. The salt of radium--generally the
bromide or chloride--is brought into acid solution. This causes
the emanation to be freely given off as fast as it is formed. At
intervals we pump it off with a mercury pump.

Let us see how many millicuries we will in future be able to turn
out in the week in our new Dublin Radium Institute.[1] We shall
have about 130 milligrammes of radium. In 3.8 days we get 65
millicuries from this--_i.e._ half the equilibrium amount of 130
millicuries. Hence in the week, we shall have about 130
millicuries.

This is not much. Many experts consider this little enough for
one tube. But here in Dublin we have been using the emanation in
a more economical and effective manner than is the usage
elsewhere; according to a method which has been worked out and
developed in our own Radium Institute. The economy is obtained by
the very simple expedient of minutely subdividing the' dose. The
system in vogue, generally, is to treat the tumour by inserting
into it one or two very active

[1] Then recently established by the Royal Dublin Society.

257

tubes, containing, perhaps, up to 200 millicuries, or even more,
per tube. Now these very heavily charged tubes give a radiation
so intense at points close to the tube, due to the greater
density of the rays near the tube, and, also, to the action of
the softer and more easily absorbable rays, that it has been
found necessary to stop these softer rays--both the y and ß--by
wrapping lead or platinum round the tube. In this lead or
platinum some thirty per cent. or more of the rays is absorbed
and, of course, wasted. But in the absence of the screen there is
extensive necrosis of the tissues near the tubes.

If, however, in place of one or two such tubes we use ten or
twenty, each containing one-tenth or one-twentieth of the dose,
we can avail ourselves of the softer rays around each tube with
benefit. Thus a wasteful loss is avoided. Moreover a more uniform
"illumination" of the tissues results, just as we can illuminate
a hall more uniformly by the use of many lesser centres of light
than by the use of one intense centre of radiation. Also we get
what is called "cross-radiation,"which is found to be beneficial.
The surgeon knows far better what he is doing by this method.
Thus it may be arranged for the effects to go on with approximate
uniformity throughout the tumour instead of varying rapidly
around a central point or--and this may be very important-- the
effects may be readily concentrated locally.

Finally, not the least of the benefit arises in the easy
technique of this new method. The quantities of

258

emanation employed can fit in the finest capillary glass tubing
and the hairlike tubes can in turn be placed in fine exploring
needles. There is comparatively little inconvenience to the
patient in inserting these needles, and there is the most perfect
control of the dosage in the number and strength of these tubes
and the duration of exposure.[1]

The first Radium Institute in Ireland has already done good work
for the relief of human suffering. It will have, I hope, a great
future before it, for I venture, with diffidence, to hold the
opinion, that with increased study the applications and claims of
radioactive treatment will increase.

[1] For particulars of the new technique and of some of the work
already accomplished, see papers, by Dr. Walter C. Stevenson,
_British Medical Journal_, July 4th, 1914, and March 20th, 1915.

259

SKATING [1]

IT is now many years ago since, as a student, I was present at a
college lecture delivered by a certain learned professor on the
subject of friction. At this lecture a discussion arose out of a
question addressed to our teacher: "How is it we can skate on ice
and on no other substance?"

The answer came back without hesitation: "Because the ice is so
smooth."

It was at once objected: "But you can skate on ice which is not
smooth."

This put the professor in a difficulty. Obviously it is not on
account of the smoothness of the ice. A piece of polished plate
glass is far smoother than a surface of ice after the latter is
cut up by a day's skating. Nevertheless, on the scratched and
torn ice-surface skating is still quite possible; on the smooth
plate glass we know we could not skate.

Some little time after this discussion, the connection between
skating and a somewhat abstruse fact in physical science occurred
to me. As the fact itself is one which has played a part in the
geological history of the earth,

[1] A lecture delivered before the Royal Dublin Society in 1905.

260

and a part of no little importance, the subject of skating,
whereby it is perhaps best brought home to every one, is
deserving of our careful attention. Let not, then, the title of
this lecture mislead the reader as to the importance of its
subject matter.

Before going on to the explanation of the wonderful freedom of
the skater's movements, I wish to verify what I have inferred as
to the great difference in the slipperiness of glass and the
slipperiness of ice. Here is a slab of polished glass. I can
raise it to any angle I please so that at length this brass
weight of 250 grams just slips down when started with a slight
shove. The angle is, as you see, about 12½ degrees. I now
transfer the weight on to this large slab of ice which I first
rapidly dry with soft linen. Observe that the weight slips down
the surface of ice at a much lower angle. It is a very low angle
indeed: I read it as between 4 and 5 degrees. We see by this
experiment that there is a great difference between the
slipperiness of the two surfaces as measured by what is called
"the angle of friction." In this experiment, too, the glass
possesses by far the smoother surface although I have rubbed the
deeper rugosities out of the ice by smoothing it with a glass
surface. Notwithstanding this, its surface is spotted with small
cavities due to bubbles and imperfections. It is certain that if
the glass was equally rough, its angle of friction towards the
brass weight would be higher.

261

We have, however, another comparative experiment to carry out. I
made as you saw a determination of the angle at which this weight
of 250 grams just slipped on the ice. The lower surface of the
weight, the part which presses on the ice, consists of a light,
brass curtain ring. This can be detached. Its mass is only 6½
grams, the curtain ring being, in fact, hollow and made of very
thin metal. We have, therefore, in it a very small weight which
presents exactly the same surface beneath as did the weight of
250 grams. You see, now, that this light weight will not slip on
ice at 5 or 6 degrees of slope, but first does so at about io
degrees.

This is a very important experiment as regards our present
inquiry. Ice appears to possess more than one angle of friction
according as a heavy or a light weight is used to press upon it.
We will make the same experiment with the plate of glass. You see
that there is little or no difference in the angle of friction of
brass on glass when we press the surfaces together with a heavy
or with a light weight. The light weight requires the same slope
of 12½ degrees to make it slip.

This last result is in accordance with the laws of friction. We
say that when solid presses on solid, for each pair of substances
pressed together there is a constant ratio between the force
required to keep one in motion over the other, and the force
pressing the solids together. This ratio is called"the
coefficient of friction."The coefficient is, in fact, constant or
approximately

262

so. I can determine the coefficient of friction from the angle of
friction by taking the tangent of the angle. The tangent of the
angle of friction is the coefficient of friction. If, then, the
coefficient is constant, so, of course, must the angle of
friction be constant. We have seen that it is so in the case of
metal on glass, but not so in the case of metal on ice. This
curious result shows that there is something abnormal about the
slipperiness of ice.

The experiments we have hitherto made are open to the reproach
that the surface of the ice is probably damp owing to the warmth
of the air in contact with it. I have here a means of dealing
with a surface of cold, dry ice. This shallow copper tank about
18 inches (45 cms.) long, and 4 inches (10 cms.) wide, is filled
with a freezing 'mixture circulated through it from a larger
vessel containing ice melting in hydrochloric acid at a
temperature of about -18° C. This keeps the tank below the
melting point of ice. The upper surface of the tank is provided
with raised edges so that it can be flooded with water. The water
is now frozen and its temperature is below 0° C. It is about
10° C. I can place over the ice a roof-shaped cover made of two
inclined slabs of thick plate glass. This acts to keep out warm
air, and to do away with any possibility of the surface of the
ice being wet with water thawed from the ice. The whole tank
along with its roof of glass can be adjusted to any angle, and a,
scale at the

263

raised end of the tank gives the angle of slope in degrees. A
weight placed on the ice can be easily seen through the glass
cover.

The weight we shall use consists of a very light ring of
aluminium wire which is rendered plainly visible by a ping-pong
ball attached above it. The weight rests now on a copper plate
provided for the purpose at the upper end of the tank. The plate
being in direct contact beneath with the freezing mixture we are
sure that the aluminium ring is no hotter than the ice. A light
jerk suffices to shake the weight on to the surface of the ice.

We find that this ring loaded with only the ping-pong ball, and
weighing a total of 2.55 grams does not slip at the low angles. I
have the surface of the ice at an angle of rather over 13½, and
only by continuous tapping of the apparatus can it be induced to
slip down. This is a coefficient of 0.24, and compares with the
coefficient of hard and smooth solids on one another. I now
replace the empty ping-pong ball by a similar ball filled with
lead shot. The total weight is now 155 grams. You see the angle
of slipping has fallen to 7°.

Every one who has made friction experiments knows how
unsatisfactory and inconsistent they often are. We can only
discuss notable quantities and broad results, unless the most
conscientious care be taken to eliminate errors. The net result
here is that ice at about -10° C. when pressed on by a very light
weight possesses a

264

coefficient of friction comparable with the usual coefficients of
solids on solids, but when the pressure is increased, the
coefficient falls to about half this value.

The following table embodies some results obtained on the
friction of ice and glass, using the methods I have shown you. I
add some of the more carefully determined coefficients of other
observers.

          Wt. in      On Plate         On Ice         On Ice
          Grams.      Glass.           at 0° C.       at 10° C.

                   Angle. Coeff.    Angle. Coeff.  Angle. Coeff
Aluminium   2.55    12½°  0.22        12°   0.21     13½°   0.24
Same      155       12½°   0.22        6°   0.11       7°   0.12
Brass       6.5     12½°  0.22        10°   0.17     10½°   0.18
Same      107       12½°   0.22        5°   0.09       6°   0.10

Steel on steel (Morin) - - - - 0.14
Brass on cast iron (Morin) - - 0.19
Steel on cast iron (Morin) - - 0.20
Skate on ice (J. Müller) - - - 0.016--0.032
Best-greased surfaces (Perry) - 0.03--0.036

You perceive from the table that while the friction of brass or
aluminium on glass is quite independent of the weight used, that
of brass or aluminium on ice depends in some way upon the weight,
and falls in a very marked degree when the weight is heavy. Now,
I think that if we had been on the look out for any abnormality
in the friction of hard substances on ice, we would have rather
anticipated a variation in the

265

other direction. We would have, perhaps, expected that a heavy
weight would have given rise to the greater friction. I now turn
to the explanation of this extraordinary result.

You are aware that it requires an expenditure of heat merely to
convert ice to water, the water produced being at the temperature
of the ice, _i.e._ at 0° C., from which it is derived. The heat
required to change the ice from the solid to the liquid state is
the latent heat of water. We take the unit quantity of heat to be
that which is required to heat 1 kilogram of water 1° C. Then if
we melt 1 kilogram of ice, we must supply it with 80 such units
of heat. While melting is going on, there is no change of
temperature if the experiment is carefully conducted. The melting
ice and the water coming from it remain at 0° C. throughout the
operation, and neither the thermometer nor your own sensations
would tell you of the amount of heat which was flowing in. The
heat is latent or hidden in the liquid produced, and has gone to
do molecular work in the substance. Observe that if we supply
only 40 thermal units, we get only one-half the ice melted. If
only 10 units are supplied, then we get only one eighth of a
kilogram of water, and no more nor less.

I have ventured to recall to you these commonplaces of science
before considering a mode of melting ice which is less generally
known, and which involves no supply of heat on your part. This
method involves for its

266

understanding a careful consideration of the thermal properties
of water in the solid state.

It must have been observed a very long time ago that water
expands when it freezes. Otherwise ice would not float on water;
and, what is perhaps more important in your eyes, your water
pipes would not burst in winter when the water freezes therein.
But although the important fact of the expansion of water on
freezing was so long presented to the observation of mankind, it
was not till almost exactly the middle of the last century that
James Thomson, a gifted Irishman, predicted many important
consequences arising from the fact of the expansion of water on
becoming solid. The principles lie enunciated are perfectly
general, and apply in every case of change of volume attending
change of state. We are here only concerned with the case of
water and ice.

James Thomson, following a train of thought which we cannot here
pursue, predicted that owing to the fact of the expansion of
water on becoming solid, pressure will lower the melting point of
ice or the freezing point of water. Normally, as you are aware,
the temperature is 0° C. or 32° F. Thomson said that this would
be found to be the freezing point only at atmospheric pressure.
He calculated how much it would change with change of pressure.
He predicted that the freezing point would fall 0.0075 of a
degree Centigrade for each additional atmosphere of pressure
applied to the water. Suppose,

267

for instance, our earth possessed an atmosphere so heavy to as
exert a thousand times the pressure of the existing atmosphere,
then water would not freeze at 0° C., but at -7.5° C. or about
18° F. Again, in vacuo, that is when the pressure has been
reduced to the relatively small vapour pressure of the water, the
freezing point is above 0° C., _i.e._ at 0.0075° C. In parts of
the ocean depths the pressure is much over a thousand
atmospheres. Fresh water would remain liquid there at
temperatures much below 0° C.

It will be evident enough, even to those not possessed of the
scientific insight of James Thomson, that some such fact is to be
anticipated. It is, however, easy to be wise after the event. It
appeals to us in a general way that as water expands on freezing,
pressure will tend to resist the turning of it to ice. The water
will try to remain liquid in obedience to the pressure. It will,
therefore, require a lower temperature to induce it to become
ice.

James Thomson left his thesis as a prediction. But he predicted
exactly what his distinguished brother, Sir William Thomson--later
Lord Kelvin--found to happen when the matter was put to the test
of experiment. We must consider the experiment made by Lord
Kelvin.

According to Thomson's views, if a quantity of ice and water are
compressed, there must be _a fall of temperature_. The nature of
his argument is as follows:

268

Let the ice and water be exactly at 0° C. to start with. Then
suppose we apply, say, one thousand atmospheres pressure. The
melting point of the ice is lowered to -7.5° C. That is, it will
require a temperature so low as -7.5° C. to keep it solid. It
will therefore at once set about melting, for as we have seen,
its actual temperature is not -7.5° C., but a higher temperature,
_i.e._ 0° C. In other words, it is 7.5° above its melting point.
But as soon as it begins melting it also begins to absorb heat to
supply the 80 thermal units which, as we know, are required to
turn each kilogram of the ice to water. Where can it get this
heat? We assume that we give it none. It has only two sources,
the ice can take heat from itself, and it can take heat from the
water. It does both in this case, and both ice and water drop in
temperature. They fall in temperature till -7.5° is reached. Then
the ice has got to its melting point under the pressure of one
thousand atmospheres, or, as we may put it, the water has reached
its freezing point. There can be no more melting. The whole mass
is down to -7.5° C., and will stay there if we keep heat from
flowing either into or out of the vessel. There is now more water
and less ice in the vessel than when we started, and the
temperature has fallen to -7.5° C. The fall of temperature to the
amount predicted by the theory was verified by Lord Kelvin.

Suppose we now suddenly remove the pressure; what will happen? We
have water and ice at -7.5° C.

269

and at the normal pressure. Water at -7.5° and at the normal
pressure of course turns to ice. The water will, therefore,
instantly freeze in the vessel, and the whole process will be
reversed. In freezing, the water will give up its latent heat,
and this will warm up the whole mass till once again 0° C. is
attained. Then there will be no more freezing, for again the ice
is at its melting point. This is the remarkable series of events
which James Thomson predicted. And these are the events which
Lord Kelvin by a delicate series of experiments, verified in
every respect.

Suppose we had nothing but solid ice in the vessel at starting,
would the experiment result in the same way? Yes, it assuredly
would. The ice under the increased pressure would melt a little
everywhere throughout its mass, taking the requisite latent heat
from itself at the expense of its sensible heat, and the
temperature of the ice would fall to the new melting point.

Could we melt the whole of the ice in this manner? Again the
answer is "yes." But the pressure must be very great. If we
assume that all the heat is obtained at the expense of the
sensible heat of the ice, the cooling must be such as to supply
the latent heat of the whole mass of water produced. However, the
latent heat diminishes as the melting point is lowered, and at a
rate which would reduce it to nothing at about 18,000
atmospheres. Mousson, operating on ice enclosed in a conducting
cylinder and cooled to -18° at starting

270

appears to have obtained very complete liquefaction. Mousson must
have attained a pressure of at least an amount adequate to lower
the melting point below -18°. The degree of liquefaction actually
attained may have been due in part to the passage of heat through
the walls of the vessel. He proved the more or less complete
liquefaction of the ice within the vessel by the fall of a copper
index from the top to the bottom of the vessel while the pressure
was on.

I have here a simple way of demonstrating to you the fall of
temperature attending the compression of ice. In this mould,
which is strongly made of steel, lined with boxwood to diminish
the passage of conducted heat, is a quantity of ice which I
compress when I force in this plunger. In the ice is a
thermoelectric junction, the wires leading to which are in
communication with a reflecting galvanometer. The thermocouple is
of copper and nickel, and is of such sensitiveness as to show by
motion of the spot of light on the screen even a small fraction
of a degree. On applying the pressure, you see the spot of light
is displaced, and in such a direction as to indicate cooling. The
balancing thermocouple is all the time imbedded in a block of ice
so that its temperature remains unaltered. On taking off the
pressure, the spot of light returns to its first position. I can
move the spot of light backwards and forwards on the screen by
taking off and putting on the pressure. The effects are quite
instantaneous.

271

The fact last referred to is very important. The ice, in fact, is
as it were automatically turned to water. It is not a matter of
the conduction of heat from point to point in the ice. Its own
sensible heat is immediately absorbed throughout the mass. This
would be the theoretical result, but it is probable that owing to
imperfections throughout the ice and failure in uniformity in the
distribution of the stress, the melting would not take place
quite uniformly or homogeneously.

Before applying our new ideas to skating, I want you to notice a
fact which I have inferentially stated, but not specifically
mentioned. Pressure will only lead to the melting of ice if the
new melting point, _i.e._ that due to the pressure, is below the
prevailing temperature. Let us take figures. The ice to start
with is, say, at -3° C. Suppose we apply such a pressure to this
ice as will confer a melting point of -2° C. on it. Obviously,
there will be no melting. For why should ice which is at -3° C.
melt when its melting point is -2° C.? The ice is, in fact,
colder than its melting point. Hence, you note this fact: The
pressure must be sufficiently intense to bring the melting point
below the prevailing temperature, or there will be no melting;
and the further we reduce the melting point by pressure below the
prevailing temperature, the more ice will be melted.

We come at length to the object of our remarks I don't know who
invented skating or skates. It is said that in the thirteenth
century the inhabitants of

272

England used to amuse themselves by fastening the bones of an
animal beneath their feet, and pushing themselves about on the
ice by means of a stick pointed with iron. With such skates, any
performance either on inside or outside edge was impossible. We
are a conservative people. This exhilarating amusement appears to
have served the people of England for three centuries. Not till
1660 were wooden skates shod with iron introduced from the
Netherlands. It is certain that skating was a fashionable
amusement in Pepys' time. He writes in 1662 to the effect: "It
being a great frost, did see people sliding with their skates,
which is a very pretty art." It is remarkable that it was the
German poet Klopstock who made skating fashionable in Germany.
Until his time, the art was considered a pastime, only fit for
very young or silly people.

I wish now to dwell upon that beautiful contrivance the modern
skate. It is a remarkable example of how an appliance can develop
towards perfection in the absence of a really intelligent
understanding of the principles underlying its development. For
what are the principles underlying the proper construction of the
skate? After what I have said, I think you will readily
understand. The object is to produce such a pressure under the
blade that the ice will melt. We wish to establish such a
pressure under the skate that even on a day when the ice is below
zero, its melting

273

point is so reduced just under the edge of the skate that the ice
turns to water.

It is this melting of the ice under the skate which secures the
condition essential to skating. In the first place, the skate no
longer rests on a solid. It rests on a liquid. You are aware how
in cases where we want to reduce friction--say at the bearing of a
wheel or under a pivot--we introduce a liquid. Look at the
bearings of a steam engine. A continuous stream of oil is fed in
to interpose itself between the solid surfaces. I need not
illustrate so well-known a principle by experiment. Solid
friction disappears when the liquid intervenes. In its place we
substitute the lesser difficulty of shearing one layer of the
liquid over the other; and if we keep up the supply of oil the
work required to do this is not very different, no matter how
great we make the pressure upon the bearings. Compared with the
resistance of solid friction, the resistance of fluid friction is
trifling. Here under the skate the lubrication is perhaps the
most perfect which it is possible to conceive. J. Müller has
determined the coefficient by towing a skater holding on by a
spring balance. The coefficient is between 0.016 and 0.032. In
other words, the skater would run down an incline so little as 1
or 2 degrees; an inclination not perceivable by the eye. Now
observe that the larger of these coefficients is almost exactly
the same as that which Perry found in the case of well-greased
surfaces. But evidently no

274

artificial system of lubrication could hope to equal that which
exists between the skate and the ice. For the lubrication here
is, as it were, automatic. In the machine if the lubricant gets
squeezed out there instantly ensues solid friction. Under the
skate this cannot happen for the squeezing out of the lubricant
is instantly followed by the formation of another film of water.
The conditions of pressure which may lead to solid friction in
the machine here automatically call the lubricant into
existence.

Just under the edge of the skate the pressure is enormous.
Consider that the whole weight of the skater is born upon a mere
knife edge. The skater alternately throws his whole weight upon
the edge of each skate. But not only is the weight thus
concentrated upon one edge, further concentration is secured in
the best skates by making the skate hollow-ground, _i.e._
increasing the keenness of the edge by making it less than a
right angle. Still greater pressure is obtained by diminishing
the length of that part of the blade which is in contact with the
ice. This is done by putting curvature on the blade or making it
what is called "hog-backed." You see that everything is done to
diminish the area in contact with the ice, and thus to increase
the pressure. The result is a very great compression of the ice
beneath the edge of the skate. Even in the very coldest weather
melting must take place to some extent.

As we observed before, the melting is instantaneous,

275

Heat has not to travel from one point of the ice to another;
immediately the pressure comes on the ice it turns to water. It
takes the requisite heat from itself in order that the change of
state may be accomplished. So soon as the skate passes on, the
water resumes the solid state. It is probable that there is an
instantaneous escape, and re-freezing of some of the water from
beneath the skate, the skate instantly taking a fresh bearing and
melting more ice. The temperature of the water escaping from
beneath the skate, or left behind by it, immediately becomes what
it was before the skate pressed upon it.

Thus, a most wonderful and complex series of molecular events
takes place beneath the skate. Swift as it passes, the whole
sequence of events which James Thomson predicted has to take
place beneath the blade Compression; lowering of the melting
point below the temperature of the surrounding ice; melting;
absorption of heat; and cooling to the new melting point, _i.e._
to that proper to the pressure beneath the blade. The skate now
passes on. Then follow: Relief of pressure; re-solidification of
the water; restoration of the borrowed heat from the congealing
water and reversion of the ice to the original temperature.

If we reflect for a moment on all this, we see that we do not
skate on ice but on water. We could not skate on ice any more
than we could skate on glass. We saw that with light weights and
when the pressure

276

{Diagram}

Diagram showing successive states obtaining in ice, before,
during, and after the passage of the skate. The temperatures and
pressures selected for illustration are such as might occur under
ordinary conditions. The edge of the skate is shown in magnified
cross-section.

277

Was not sufficient to melt the ice, the friction was much the
same as that of metal on glass. Ice is not slippery. It is an
error to say that it is. The learned professor was very much
astray when he said that you could skate on ice because it is so
smooth. The smoothness of the ice has nothing to do with the
matter. In short, owing to the action of gravity upon your body,
you escape the normal resistance of solid on solid, and glide
about with feet winged like the messenger of the Gods; but on
water.

A second condition essential to the art of skating is also
involved in the melting of the ice. The sinking of the skate
gives the skater "bite." This it is which enables him to urge
himself forward. So long as skates consisted of the rounded bones
of animals, the skater had to use a pointed staff to propel
himself. In creating bite, the skater again unconsciously appeals
to the peculiar physical properties of ice. The pressure required
for the propulsion of the skater is spread all along the length
of the groove he has cut in the ice, and obliquely downwards. The
skate will not slip away laterally, for the horizontal component
of the pressure is not enough to melt the ice. He thus gets the
resistance he requires.

You see what a very perfect contrivance the skate is; and what a
similitude of intelligence there is in its evolution. Blind
intelligence, because it is certain the true physics of skating
was never held in view by

278

the makers of skates. The evolution of the skate has been truly
organic. The skater selected the fittest skate, and hence the fit
skate survived.

In a word, the possibility of skating depends on the dynamical
melting of ice under pressure. And observe the whole matter turns
upon the apparently unrelated fact that the freezing of water
results in a solid more bulky than the water which gives rise to
it. If ice was less bulky than the water from which it was
derived, pressure would not melt it; it would be all the more
solid for the pressure, as it were. The melting point would rise
instead of falling. Most substances behave in this manner, and
hence we cannot skate upon them. Only quite a few substances
expand on freezing, and it happens that their particular melting
temperatures or other properties render them unsuitable to
skating. The most abundant fluid substance on the earth, and the
most abundant substance of any one kind on its surface, thus
possesses the ideally correct and suitable properties for the art
of skating.

I have pointed out that the pressure must be such as to bring the
temperature of melting below that prevailing in the ice at the
time. We have seen also, that one atmosphere lowers the melting
point of ice by the 1/140 of a degree Centigrade; more exactly by
0.0075°. Let us now assume that the skate is so far sunken in the
ice as to bear for a length of two inches, and for a width of
one-hundredth of an inch. The skater weighs,

279

let us say--150 pounds. If this weight was borne on one square
inch, the pressure would be ten atmospheres. But the skater rests
his weight, in fact, upon an area of one-fiftieth of an inch. The
pressure is, therefore, fifty times as great. The ice is
subjected to a pressure of 500 atmospheres. This lowers the
melting point to -3.75° C. Hence, on a day when the ice is at
this temperature, the skate will sink in the ice till the weight
of the skater is concentrated as we have assumed. His skate can
sink no further, for any lesser concentration of the pressure
will not bring the melting point below the prevailing
temperature. We can calculate the theoretical bite for any state
of the ice. If the ice is colder the bite will not be so deep. If
the temperature was twice as far below zero, then the area over
which the skater's weight will be distributed, when the skate has
penetrated its maximum depth, will be only half the former area,
and the pressure will be one thousand atmospheres.

An important consideration arises from the fact that under the
very extreme edge of the skate the pressure is indefinitely
great. For this involves that there will always be some bite,
however cold the ice may be. That is, the narrow strip of ice
which first receives the skater's weight must partially liquefy
however cold the ice.

It must have happened to many here to be on ice which was too
cold to skate on with comfort. The

280

skater in this case speaks of the ice as too hard. In the
Engadine, the ice on the large lakes gets so cold that skaters
complain of this. On the rinks, which are chiefly used there, the
ice is frequently renewed by flooding with water at the close of
the day. It thus never gets so very cold as on the lakes. I have
been on ice in North France, which, in the early morning, was too
hard to afford sufficient bite for comfort. The cause of this is
easily understood from what we have been considering.

We may now return to the experimental results which we obtained
early in the lecture. The heavy weights slip off the ice at a low
angle because just at the points of contact with the ice the
latter melts, and they, in fact, slip not on ice but on water.
The light weights on cold, dry ice do not lower the melting point
below the temperature of the ice, _i.e._ below -10° C., and so
they slip on dry ice. They therefore give us the true coefficient
of friction of metal on ice.

This subject has, more recently been investigated by H. Morphy,
of Trinity College, Dublin. The refinement of a closed vessel at
uniform temperature, in which the ice is formed and the
experiment carried out, is introduced. Thermocouples give the
temperatures, not only of the ice but of the aluminium sleigh
which slips upon it under various loads. In this way we may be
certain that the metal runners are truly at the temperature of
the ice. I now quote from Morphy's paper

281

"The angle of friction was found to remain constant until a
certain stage of the loading, when it suddenly fell to about half
of its original value. It then remained constant for further
increases in the load.

"These results, which confirmed those obtained previously with
less satisfactory apparatus, are shown in the table below. In the
first column is shown the load, _i.e._ the weight of sleigh +
weight of shot added. In the second and third columns are shown,
respectively, the coefficient and angle of friction, whilst the
fourth gives the temperature of the ice as determined from the
galvanometer deflexions.

Load.        Tan y.          y.        Temp.

5.68 grams.  0.36±.01     20°±30'    -5.65° C.
10.39                                -5.65°
11.96                                -5.75°
12.74                                -5.60°
13.53                                -5.65°
14.31                                -5.65°
15.10 grams. 0.17±.01    9°.30'±30'  -5.60°
16.67                                -5.55°
19.81                                -5.60°
24.52                                -5.60°
5.68 grams. 0.36±.01      20°±30'    -5.60°

"These experiments were repeated on another occasion with the same
result and similar results had been obtained with different
apparatus.

"As a result of the investigation the following points are
clearly shown:--

282

"(1) The coefficient of friction for ice at constant temperature
may have either of two constant values according to the pressure
per unit surface of contact.

"(2) For small pressures, and up to a certain well defined limit
of pressure, the coefficient is fairly large, having the value
0.36±.01 in the case investigated.

"(3) For pressures greater than the above limit the coefficient
is relatively small, having the value 0.17±.01 in the case
investigated."

It will be seen that Morphy's results are similar to those
arrived at in the first experimental consideration of our
subject; but from the manner in which the experiments have been
carried out, they are more accurate and reliable.

A great deal more might be said about skating, and the allied
sports of tobogganing, sleighing, curling, ice yachting, and
last, but by no means least, sliding--that unpretentious pastime
of the million. Happy the boy who has nails in his boots when
Jack-Frost appears in his white garment, and congeals the
neighbouring pond. But I must turn away at the threshold of the
humorous aspect of my subject (for the victim of the street
"slide" owes his injured dignity to the abstruse laws we have
been discussing) and pass to other and graver subjects intimately
connected with skating.

James Thomson pointed out that if we apply compressional stress
to an ice crystal contained in a vessel

283

which also contains other ice crystals, and water at 0° C., then
the stressed crystal will melt and become water, but its
counterpart or equivalent quantity of ice will reappear elsewhere
in the vessel. This is, obviously, but a deduction from the
principles we have been examining. The phenomenon is commonly
called "regelation." I have already made the usual regelation
experiment before you when I compressed broken ice in this mould.
The result was a clear, hard and almost flawless lens of ice. Now
in this operation we must figure to ourselves the pieces of ice
when pressed against one another melting away where compressed,
and the water produced escaping into the spaces between the
fragments, and there solidifying in virtue of its temperature
being below the freezing point of unstressed water. The final
result is the uniform lens of ice. The same process goes on in a
less perfect manner when you make--or shall I better say--when you
made snowballs.

We now come to theories of glacier motion; of which there are
two. The one refers it mainly to regelation; the other to a real
viscosity of the ice.

The late J. C. M'Connel established the fact that ice possesses
viscosity; that is, it will slowly yield and change its shape
under long continued stresses. His observations, indeed, raise a
difficulty in applying this viscosity to explain glacier motion,
for he showed that an ice crystal is only viscous in a certain
structural

284

direction. A complex mixture of crystals such, as we know glacier
ice to be, ought, we would imagine, to display a nett or
resultant rigidity. A mass of glacier ice when distorted by
application of a force must, however, undergo precisely the
transformations which took place in forming the lens from the
fragments of ice. In fact, regelation will confer upon it all the
appearance of viscosity.

Let us picture to ourselves a glacier pressing its enormous mass
down a Swiss valley. At any point suppose it to be hindered in
its downward path by a rocky obstacle. At that point the ice
turns to water just as it does beneath the skate. The cold water
escapes and solidifies elsewhere. But note this, only where there
is freedom from pressure. In escaping, it carries away its latent
heat of liquefaction, and this we must assume, is lost to the
region of ice lately under pressure. This region will, however,
again warm up by conduction of heat from the surrounding ice, or
by the circulation of water from the suxface. Meanwhile, the
pressure at that point has been relieved. The mechanical
resistance is transferred elsewhere. At this new point there is
again melting and relief of pressure. In this manner the glacier
may be supposed to move down. There is continual flux of
conducted heat and converted latent heat, hither and thither, to
and from the points of resistance. The final motion of the whole
mass is necessarily slow; a few feet in the day or, in winter,

285

even only a few inches. And as we might expect, perfect silence
attends the downward slipping of the gigantic mass. The motion
is, I believe, sufficiently explained as a skating motion. The
skate is, however, fixed, the ice moves. The great Aletsch
Glacier collects its snows among the highest summits of the
Oberland. Thence, the consolidated ice makes its way into the
Rhone Valley, travelling a distance of some 20 miles. The ice now
melting into the youthful Rhone fell upon the Monch, the Jungfrau
or the Eiger in the days when Elizabeth ruled in England and
Shakespeare lived.

The ice-fall is a common sight on the glacier. In great lumps and
broken pinnacles it topples over some rocky obstacle and falls
shattered on to the glacier below. But a little further down the
wound is healed again, and regelation has restored the smooth
surface of the glacier. All such phenomena are explained on James
Thomson's exposition of the behaviour of a substance which
expands on passing from the liquid to the solid state.

We thus have arrived at very far-reaching considerations arising
out of skating and its science. The tendency for snow to
accumulate on the highest regions of the Earth depends on
principles which we cannot stop to consider. We know it collects
above a certain level even at the Equator. We may consider, then,
that but for the operation of the laws which James Thomson
brought to light, and which his illustrious brother,

286

Lord Kelvin, made manifest, the uplands of the Earth could not
have freed themselves of the burthen of ice. The geological
history of the Earth must have been profoundly modified. The
higher levels must have been depressed; the general level of the
ocean relatively to the land thereby raised, and, it is even
possible, that such a mean level might have been attained as
would result in general submergence.

During the last great glacial period, we may say the fate of the
world hung on the operation of those laws which have concerned us
throughout this lecture. It is believed the ice was piled up to a
height of some 6,000 feet over the region of Scandinavia. Under
the influence of the pressure and fusion at points of resistance,
the accumulation was stayed, and it flowed southwards the
accumulation was stayed, and it flowed southwards over Northern
Europe. The Highlands of Scotland were covered with, perhaps,
three or four thousand feet of ice. Ireland was covered from
north to south, and mighty ice-bergs floated from our western and
southern shores.

The transported or erratic stones, often of great size, which are
found in many parts of Ireland, are records of these long past
events: events which happened before Man, as a rational being,
appeared upon the Earth.

287

A SPECULATION AS TO A PREMATERIAL UNIVERSE [1]

"And therefore...these things likewise had a birth; for things
which are of mortal body could not for an infinite time back...
have been able to set at naught the puissant strength of
immeasurable age."--LUCRETIUS, _De Rerum Natura._

"O fearful meditation! Where, alack! Shall Time's best jewel
from Time's chest lie hid?" --SHAKESPEARE.

IN the material universe we find presented to our senses a
physical development continually progressing, extending to all,
even the most minute, material configurations. Some fundamental
distinctions existing between this development as apparent in the
organic and the inorganic systems of the present day are referred
to elsewhere in this volume.[2] In the present essay, these
systems as having a common origin and common ending, are merged
in the same consideration as to the nature of the origin of
material systems in general. This present essay is occupied by
the consideration of the necessity of limiting material
interactions in past time. The speculation originated in the
difficulties which present themselves when we ascribe to these
interactions infinite duration in the past. These difficulties
first claim our consideration.

[1] Proc. Royal Dublin Soc., vol. vii., Part V, 1892.

[2] _The Abundance of Life._

288

Accepting the hypothesis of Kant and Laplace in its widest
extension, we are referred to a primitive condition of wide
material diffusion, and necessarily too of material instability.
The hypothesis is, in fact, based upon this material instability.
We may pursue the sequence of events assumed in this hypothesis
into the future, and into the past.

In the future we find finality to progress clearly indicated. The
hypothesis points to a time when there will be no more
progressive change but a mere sequence of unfruitful events, such
as the eternal uniform motion of a mass of matter no longer
gaining or losing heat in an ether possessed of a uniform
distribution of energy in all its parts. Or, again, if the ether
absorb the energy of material motion, this vast and dark
aggregation eternally poised and at rest within it. The action is
transferred to the subtle parts of the ether which suffer none of
the energy to degrade. This is, physically, a thinkable future.
Our minds suggest no change, and demand none. More than this,
change is unthinkable according to our present ideas of energy.
Of progress there is an end.

This finality _â parte post_ is instructive. Abstract
considerations, based on geometrical or analytical illustrations,
question the finiteness of some physical developments. Thus our
sun may require eternal time to attain the temperature of the
ether around it, the approach to this condition being assumed to
be asymptotic in

289

character. But consider the legitimate _reductio ad absurdum_ of
an ember raked from a fire 1000 years ago. Is it not yet cooled
down to the constant temperature of its surroundings? And we may
evidently increase the time a million-fold if we please. It
appears as if we must regard eternity as outliving every
progressive change, For there is no convergence or enfeeblement
of time. The ever-flowing present moves no differently for the
occurrence of the mightiest or the most insignificant events. And
even if we say that time is only the attendant upon events, yet
this attendant waits patiently for the end, however long
deferred.

Does the essentially material hypothesis of Kant and Laplace
account for an infinite past as thinkably as it accounts for the
infinite future? As this hypothesis is based upon material
instability the question resolves itself into this:-- Is the
assumption of an infinitely prolonged past instability a probable
or possible account of the past? There are, it appears to me,
great difficulties involved in accepting the hypothesis of
infinitely prolonged material instability. I will refer here to
three principal objections. The first may be called a
metaphysical objection; the second is partly metaphysical and
partly physical, the third may be considered a physical
objection, as it is involved directly in the phenomena presented
by our universe.

The metaphysical objection must have presented itself to every
one who has considered the question. It may

290

be put thus:--If present events are merely one stage in an
infinite progress, why is not the present stage long ago passed
over? We are evidently at liberty to push back any stage of
progress to as remote a period as we like by putting back first
the one before this and next the stage preceding this, and so on,
for, by hypothesis, there is no beginning to the progress.

Thus, the sum of passing events constituting the present universe
should long ago have been accomplished and passed away. If we
consider alternative hypotheses not involving this difficulty, we
are at once struck by the fact that the future of material
development is free of the objection. For the eternity of
unprogressive events involved in the future on Kant's hypothesis,
is not only thinkable, but any change is, as observed,
irreconcilable with our ideas of energy. As in the future so in
the past we look to a cessation to progress. But as we believe
the activity of the present universe must in some form have
existed all along, the only refuge in the past is to imagine an
active but unprogressive eternity, the unprogressive activity at
some period becoming a progressive activity--that progressive
activity of which we are spectators. To the unprogressive
activity there was no beginning; in fact, beginning is as
unthinkable and uncalled for to the unprogressive activity of the
past as ending is to the unprogressive activity of the future,
when all developmental actions shall have ceased. There is no
beginning or ending to the activity of the universe.

291

There is beginning and ending to present progressive activity.
Looking through the realm of nature we seek beginning and ending,
but "passing through nature to eternity" we find neither. Both
are justified; the questioning of the ancient poet regarding the
past, and of the modern regarding the future, quoted at the head
of this essay.

The next objection, which is in part metaphysical, is founded on
the difficulty of ascribing any ultimate reality or potency to
forces diminishing through eternal time. Thus, against the
assumption that our universe is the result of material
aggregation progressing over eternal time, which involves the
primitive infinite separation of the particles, we may ask, what
force can have acted between particles sundered by infinite
distance? The gravitational force falling off as the square of
the distance, must vanish at infinity if we mean what we say when
we ascribe infinite separation to them. Their condition is then
one of neutral stability, a finite movement of the particles
neither increasing nor diminishing interaction. They had then
remained eternally in their separated condition, there being no
cause to render such condition finite. The difficulty involved
here appears to me of the same nature as the difficulty of
ascribing any residual heat to the sun after eternal time has
elapsed. In both cases we are bound to prolong the time, from our
very idea of time, till progress is no more, when in the one case
we can imagine no mutual approximation of the

292

particles, in the other no further cooling of the body. However,
I will riot dwell further upon this objection, as it does not, I
believe, present itself with equal force to every mind. A reason
less open to dispute, as being less subjective, against the
aggregation of infinitely remote particles as the origin of our
universe, is contained in the physical objection.

In this objection we consider that the appearance presented by
our universe negatives the hypothesis of infinitely prolonged
aggregation. We base this negation upon the appearance of
simultaneity ~ presented by the heavens, contending that this
simultaneity is contrary to what we would expect to find in the
case of particles gathered from infinitely remote distances.
Whether these particles were endowed with relative motions or not
is unimportant to the consideration. In what respects do the
phenomena of our universe present the appearance of simultaneous
phenomena? We must remember that the suns in space are as fires
which brighten only for a moment and are then extinguished. It is
in this sense we must regard the longest burning of the stars.
Whether just lit or just expiring counts little in eternity. The
light and heat of the star is being absorbed by the ether of
space as effectually and rapidly as the ocean swallows the ripple
from the wings of an expiring insect. Sir William Herschel says
of the galaxy of the milky way:-- "We do not know the rate of
progress of this mysterious chronometer, but it is nevertheless
certain that it cannot

293

last for ever, and its past duration cannot be infinite." We do
not know, indeed, the rate of progress of the chronometer, but if
the dial be one divided into eternal durations the consummation
of any finite physical change represents such a movement of the
hand as is accomplished in a single vibration of the balance
wheel.

Hence we must regard the hosts of glittering stars as a
conflagration that has been simultaneously lighted up in the
heavens. The enormous (to our ideas) thermal energy of the stars
resembles the scintillation of iron dust in a jar of oxygen when
a pinch of the dust is thrown in. Although some particles be
burnt up before others become alight, and some linger yet a
little longer than the others, in our day's work the
scintillation of the iron dust is the work of a single instant,
and so in the long night of eternity the scintillation of the
mightiest suns of space is over in a moment. A little longer,
indeed, in duration than the life which stirs a moment in
response to the diffusion of the energy, but only very little. So
must an Eternal Being regard the scintillation of the stars and
the periodic vibration of life in our geological time and the
most enduring efforts of thought. The latter indeed are no more
lasting than

"... the labour of ants In the light of a million million of
suns."

But the myriad suns themselves, with their generations, are the
momentary gleam of lights for ever after extinguished.

294

Again, science suggests that the present process of material
aggregation is not finished, and possibly will only be when it
prevails universally. Hence the very distribution of the stars,
as we observe them, as isolated aggregations, indicates a
development which in the infinite duration must be regarded as
equally advanced in all parts of stellar space and essentially a
simultaneous phenomenon. For were we spectators of a system in
which any very great difference of age prevailed, this very great
difference would be attended by some such appearance as the
following:--

The aupearance of but one star, other generations being long
extinct or no others yet come into being; or, perhaps, a faint
nebulous wreath of aggregating matter somewhere solitary in the
heavens; or no sign of matter beyond our system, either because
ungathered or long passed away into darkness.[1]

Some such appearances were to be expected had the aggregation of
matter depended solely on chance encounters of particles
scattered through infinite space.

For as, by hypothesis, the aggregation occupies an infinite time
in consummation it is nearly a certainty that each particle
encountered after immeasurable time, and then for the first time
endowed with actual gravitational potential energy, would have
long expended this energy

[1] It is interesting to reflect upon the effect which an entire
absence of luminaries outside our solar system would have had
upon the views of our philosophers and upon our outlook on life.

295

before another particle was gathered. But the fact that so many
fires which we know to be of brief duration are scattered through
a region of space, and the fact of a configuration which we
believe to be a transitory ore, suggest their simultaneous
aggregation here and there. And in the nebulous wreaths situated
amidst the stars there is evidence that these actually originated
where they now are, for in such no relative motion, I believe,
has as yet been detected by the spectroscope. All this, too, is
in keeping with the nebular hypothesis of Kant and Laplace so
long as this does not assume a primitive infinite dispersion of
matter, but the gathering of matter from finite distances first
into nebulous patches which aggregating with each other have
given rise to our system of stars. But if we extend this
hypothesis throughout an infinite past by the supposition of
aggregation of infinitely remote particles we replace the
simultaneous approach required in order to accotnt for the
simultaneous phenomena visible in the heavens, by a succession of
aggregative events, by hypothesis at intervals of nearly infinite
duration, when the events of the universe had consisted of fitful
gleams lighted after eternities of time and extinguished for yet
other eternities.

Finally, if we seek to replace the eternal instability involved
in Kant's hypothesis when extended over an infinite past, by any
hypothesis of material stability, we at once find ourselves in
the difficulty that from the known properties of matter such
stability must have been

296

permanent if ever existent, which is contrary to fact. Thus the
kinetic inertia expressed in Newton's first law of motion might
well be supposed to secure equilibrium with material attraction,
but if primevally diffused matter had ever thus been held in
equilibrium it must have remained so, or it was maintained so
imperfectly, which brings us back to endless evolution.

On these grounds I contend that the present gravitational
properties of matter cannot be supposed to have acted for all
past duration. Universal equilibrium of gravitating particles
would have been indestructible by internal causes. Perpetual
instability or evolution is alike unthinkable and contrary to the
phenomena of the universe of which we are cognisant. We therefore
turn from gravitating matter as affording no rational account of
the past. We do so of necessity, however much we feel our
ignorance of the nature of the unknown actions to which we have
recourse.

A prematerial condition of the universe was, we assume, a
condition in which uniformity as regards the average distribution
of energy in space prevailed, but neterogeneity and instability
were possible. The realization of that possibility was the
beginning we seek, and we today are witnesses of the train of
events involved in the breakdown of an eternal past equilibrium.
We are witnesses on this hypothesis, of a catastrophe possibly
confined to certain regions of space, but which is, to the
motions and configurations concerned, absolutely unique,
reversible to

297

its former condition of potential by no process of which we can
have any conception.

Our speculation is that we, as spectators of evolution, are
witnessing the interaction of forces which have not always been
acting. A prematerial state of the universe was one of unfruitful
motions, that is, motions unattended by progressing changes, in
our region of the ether. How extended we cannot say; the nature
of the motions we know not; but the kinetic entities differed
from matter in the one important particular of not possessing
gravitational attraction. Such kinetic configurations we cannot
consider to be matter. It was _possible_ to construct matter by
their summation or linkage as the configuration of the crystal is
possible in the clear supersaturated liquid.

Duration in an ether filled with such motions would pass in a
succession of mere unfruitful events; as duration, we may
imagine, even now passes in parts of the ether similar to our
own. An endless (it may be) succession of unprogressive,
fruitless events. But at one moment in the infinite duration the
requisite configuration of the elementary motions is attained;
solely by the one chance disposition the stability of all must
go, spreading from the fateful point.

Possibly the material segregation was confined to one part of
space, the elementary motions condensing upon transformation, and
so impoverishing the ether around till the action ceased. Again
in the same sense as the

298

stars are simultaneous, so also they may be regarded as uniform
in size, for the difference in magnitude might have been anything
we please to imagine, if at the same time we ascribe sufficient
distance sundering great and small. So, too;, will a dilute
solution of acetate of soda build a crystal at one point, and the
impoverishment of the medium checking the growth in this region,
another centre will begin at the furthest extremities of the
first crystal till the liquid is filled with loose feathery
aggregations comparable in size with one another. In a similar
way the crystallizing out of matter may have given rise, not to a
uniform nebula in space, but to detached nebula, approximately of
equal mass, from which ultimately were formed the stars.

That an all-knowing Being might have foretold the ultimate event
at any preceding period by observing the motions of the parts
then occurring, and reasoning as to the train of consequences
arising from these nations, is supposable. But considerations
arising from this involve no difficulty in ascribing to this
prematerial train of events infinite duration. For progress there
is none, and we can quite as easily conceive of some part of
space where the same Infinite Intelligence, contemplating a
similar train of unfruitful motions, finds that at no time in the
future will the equilibrium be disturbed. But where evolution is
progressing this is no longer conceivable, as being contradictory
to the very idea of progressive development. In this case
Infinite Intelligence

299

_necessarily_ finds, as the result of his contemplation, the
aggregation of matter, and the consequences arising therefrom.

The negation of so primary a material property as gravitation to
these primitive motions of (or in) the ether, probably involves
the negation of many properties we find associated with matter.
Possibly the quality of inertia, equally primary, is involved
with that of gravitation, and we may suppose that these two
properties so intimately associated in determining the motions of
bodies in space were conferred upon the primitive motions as
crystallographic attraction and rigidity are first conferred upon
the solid growing from the supersaturated liquid. But in some
degree less speculative is the supposition that the new order of
motions involved the transformation of much energy into the form
of heat vibrations; so that the newly generated matter, like the
newly formed crystal, began its existence in a medium richly fed
with thermal radiant energy. We may consider that the thermal
conditions were such as would account for a primitive
dissociation of the elements. And, again, we recall how the
physicist finds his estimate of the energy involved in mere
gravitational aggregation inadequate to afford explanation of
past solar heat. It is supposable, on such a hypothesis as we
have been dwelling on, that the entire subsequent gravitational
condensation and conversion of material potential energy, dating
from the first formation of matter to the stage of star
formation

300

may be insignificant in amount compared with the conversion of
etherial energy attending the crystallizing out of matter from
the primitive motions. And thus possibly the conditions then
obtaining involved a progressively increasing complexity of
material structure the genesis of the elements, from an
infra-hydrogen possessing the simplest material configuration,
resulting ultimately in such self-luminous nebula as we yet see
in the heavens.

The late James Croll, in his _Stellar Evolution_, finds objections
to an eternal evolution, one of which is similar to the
"metaphysical" objection urged in this paper. His way out of the
difficulty is in the speculation that our stellar system
originated by the collision of two masses endowed with relative
motion, eternal in past duration, their meeting ushering in the
dawn of evolution. However, the state of aggregation here
assumed, from the known laws of matter and from analogy, calls
for explanation as probably the result of prior diffusion, when,
of course, the difficulty is only put back, not set at rest. Nor
do I think the primitive collision in harmony with the number of
relatively stationary nebula visible in space.

The metaphysical objection is, I find, also urged by George
Salmon, late Provost of Trinity College, in favour of the
creation of the universe.--(_Sermons on Agnosticism_.)

A. Winchell, in _World Life_, says: "We have not

301

the slightest scientific grounds for assuming that matter existed
in a certain condition from all eternity. The essential activity
of the powers ascribed to it forbids the thought; for all that we
know, and, indeed, as the _conclusion_ from all that we know,
primal matter began its progressive changes on the morning of its
existence."

Finally, in reference to the hypothesis of a unique determination
of matter after eternal duration in the past, it may not be out
of place to remind the reader of the complexity which modern
research ascribes to the structure of the atom.

302

INDEX

A.

Abney, Sir Wm., on sensitisers, 210.

Abundance of life, numerical, 98-100.

Adaptation and aggressiveness of the organism, 80.

Additive law, the, with reference to alpha rays, 220.

Age of Earth, comparison of denudative and radioactive methods of
finding, 23-29.

Aletsch glacier, 286.

Allen, Grant, on colour of Alpine plants, 104.

Allen, H. Stanley, on photo-electricity, 203.

Alpha rays, nature of, 214; velocity of, 214; effects of, on
gases, 214; range of, in air, 215; visualised, 218; ionisation
curve of, 216; number of, from one gram of radium, 237; number of
ions made by, 237.

Alpine flowers, intensity of colour of, 102.

Alps, history of, 141; Tertiary denudation of, 148; depth of
sedimentary covering of, 148; evidence of high pressures and
temperatures in, 149; recent theories of formation of, 150 _et
seq._; upheaval of, 147; age of, 147; volcanic phenomena
attending elevation of, 147.

Andes, trough parallel to, 123; not volcanic in origin, 118.

Angle of friction on ice, 261-265, 281-283; on glass, 261-265.

Animate systems, dynamic conditions of, 67; and transfer of
energy, 71; and old age, 72; mechanical imitation of, 76, 77.

Animate and inanimate systems compared, 73-75.

Appalachian range, formation of, 120.

Arrhenius, on elevation of continents, 17.

Aryan Era of India, 136.

Asteroids, probable origin of, 175; discovery of, 175; dimensions
of, 176; orbits of, 176; Mars' moons derived from, 177.

B.

Babbage and Herschel, theory of mountain building, 123.

Babes (and Cornil), size of spores, 98.

Becker, G. F., age of Earth by sodium collection, 14; age of
minerals by lead ratio, 20.

Berthelot, law of maximum work, 62.

Bertrand, Marcel, section of Mont Blanc Massif, 154.

Beta rays, nature of, 246; accompanied by gamma rays, 247;
production of, by gamma rays, 247; as ionising agents, 249.

Biotite, containing haloes, 223; pleochroism of, 235; intensified
pleochroism in halo, 235.

Body and mind, as manifestations of progressiveness of the
organism, 86.

Boltwood, age of minerals by lead ratio, 20.

Bose, theory of latent image, 203.

Bragg and Kleeman, on path of the alpha ray, 215; stopping power,
219; laws affecting ionisation by alpha rays, 220; curve of
ionisation and structure of the halo, 232.

Brecciendecke, sheet of the, 154.

Brdche, sheet of the, 154.

Burrard and Hayden on the Himalaya, 138; sections of the
Himalaya, 139.

C.

Canals and "canali," 166; curvature of, and path of a satellite,
188 _et seq._; double and triple accounted for, 186, 187;
doubling of, 195; disappearance and reappearance of, 196-198;
photography of, 198; not due to cracks, 167; not due to rivers,
167; of Mars, double nature of, 166, 170; crossing dark regions
of planet's surface, 168; of Mars, Lowell's views on, 168 _et
seq._; shown on Lowell's map, investigation of, 192 _et seq._;
radiating, explanation of, 193, 194; number of, 194; developed by
secondary disturbances, 194; nodal development of, due to raised
surface features, 195.

Chamberlin and Salisbury, the Laramide range, 121.

Clarke, F. W., estimate of mass of sediments, 9; age of Earth by
sodium collection, 14; average composition of sedimentary and
igneous rocks, 42; on average composition of the crust, 126;
solvent denudation of the continents, 17, 40.

Claus, protoplasm the test of the cell, 67; abortion of useless
organs, 69.

Coefficient of friction, definition of, 262; deduction of, from
angle of friction, 263; abnormal values on ice, 261-265, 282; for
various substances, 265.

Continental areas, movements of, 144.

Cornil and Babes, size of spores, 98.

Croll, James, dawn of evolution, 301.

Crust of the Earth, average composition of, 126; depth of
softening in, 128.

Curie, definition of the, 256.

D.

Dana, on mountain building, 120.

Dawson, reduction of surface represented by Laramide range, 123.

Deccan traps, 137

_déferlement_, theory of, 155; explanation of, 155 _et seq._;
temperature involved in, 156.

Deimos, dimensions of, 177; orbit of, 577.

De Lapparent, exotic nature of the Préalpes, 150.

De Montessus and the association of earthquakes with
geosynclines, 142.

Denudation as affected by continental elevation, 17; factors
promoting, 30 _et seg._; relative activity in mountains and on
plains, 35-40; solvent, by the sea, 40; the sodium index of,
46-50; thickness of rock-layer removed from the land, 51.

De Quincy, System of the Heavens, 200.

Dewar, Sir James, latent image formed at low temperatures, 202.

Dixon, H. H., and AGnadance of Life, 60.

Double canals, formation by attraction of a satellite, 585-187.

Douglass, A. E., observations on Mars, 167.

Dravidian Era of India, 135.

E.

Earth, early history of, 3, 4; dimensions of, relative to surface
features, 117.

Earth's age determined by thickness of sediments, 5; determined
by mass of the sediments, 7; determined by sodium in the ocean,
12; determined by radioactive transformations, 19; significance
of, 2.

Earthquakes associated with geosynclincs, 142.

Efficiency, tendency to maximum, in organisms, 113, 114.

Elements, probable wide diffusion of rare, 230; rarity of
radioactive, 241.

Elster and Geitel, photo-electric activity and absorption, 207;
photo-electric properties of gelatin, 212; Emanation of radium,
therapeutic use of, 256-259; advantages of, in medicine, 256;
volume of, 257; how obtained, 257; use of, in needles, 258.

Equilibrium amount, meaning of, 254, 255.

Evolution and acceleration of activity, 79; of the universe not
eternal a pane ante, 298.

F.

Faraday and ionisation, 57.

Finality of progress a part, post, 289.

Flahault, experiments on colour of flowers, 108.

Fletcher, A. L., proportionality of thorium and uranium, 26,

G.

Galileo, discovery of Jupiter's moons, 162.

Gamma rays, nature of, 247: production of, by beta rays, 247; as
ionising agents, 249.

Geddes and Thomson, hunger and living matter, 71.

Geiger, range of alpha rays in air, 215; ionisation affected by
alpha rays in air, 216; on "scattering," 217; scattering and the
structure of the halo, 232.

Geikie, Sir A., uniformity in geological history, 15.

Geosynclines, 119; association with earthquakes and volcanoes,
142; of the tethys, 142; radioactive heat in, due to sediments,
130; temperature effects due to lateral compression of, 131.

Glacial epoch, phenomena of, 287.

Glacier motion, cause of. 285.

Glossopteris and Gangamopteris flora, 136.

Gondwanaland, 136.

Gradient of temperature in Earth's surface crust, 126.

H.

Haimanta period of India, 135.

Halley, Edmund, finding age by saltness of ocean, 13.

Hallwachs, photo-electric activity and absorption, 207.

Haloes, pleochroic, finding age of rocks by, 21; due to uranium
and thorium families, 227; radii of, 227; over-exposed and
underexposed, 228; intimate structure of, 229 _et seq._;
artificial, 229; tubular, in mica, 230; extreme age of, 231;
effect of nucleus on structure of, 232; inference from spherical
form of, in crystals, 233; structure of, unaffected by cleavage,
235; origin of the name "pleochroic,"235; colouration due to
iron, 235; colouration not due to helium, 236; age Of, 236; slow
formation of, 237, 238; number of rays required to build, 237;
and age of the Earth, 238-241.

Hayden, H.H., geology of the Himalaya, 134, 138, 139.

Heat-tendency of the universe, 62.

Heat emission from the Earth's surface, 126; from average igneous
rock due to radioactivity, 126.

Helium and the alpha ray, 214, 222; colouration of halo not due
to, 236.

Hering, E., and physiological or unconscious memory, 111.

Herschel and Babbage theory of mountain building, 123.

Herschel, Sir W., on galaxy of milky way, 293.

Hertz, negative electrification discharged by light, 204.

Himalaya, geological history of, 134-139.

Hobbs, on association of earthquakes and geosynclines, 143.

Holmes, A., original lead in minerals, 20; age of Devonian, 21.

Horst concerned in Alpine _déferlement_, objections to, 156.

Hyperion, dimensions of, 177.

I.

Ice, melting of, by pressure, 267 _et seq._; expansion of water
in becoming, 267; lowering of melting-point by pressure, 267;
fall of temperature under pressure, 268 _et seq._; viscosity of,
284.

Igneous rocks, average composition of, 43.

Inanimate actions, dynamic conditions of, 61.

Inanimate systems, secondary effects in, 63-65; transfer of
energy into, 66.

Indian geology, equivalent nomenclature of, 139.

Initial recombination of ions due to alpha rays, 221, 222, 231;
and structure of the halo, 231.

Insect life in the higher Alps, 104, 105; destruction of, on the
Alpine snows, 106.

Ionisation by alpha ray, density of, 221; importance in chemical
actions, 250; in living cell, 250.

Ions, number of, produced by an alpha ray, 237.

Isostasy, 53; and preservation of continents, 53.

Ivy, inconspicuous blossoms of, 107; delay in ripening seed,
107.

K.

Kant and Laplace, material hypothesis of, does not account for
the past, 290.

Kelvin, Lord, experiment on effects of pressure on ice, 268-270.

Kleeman and Bragg. See Bragg.

Klopstock introduces skating into Germany, 273.

L.

Lakes, cause of blue colour of, 55.

Land, movements of the, 53, 54.

Laukester, Ray, the soma and reproductive cells, 85.

Lapworth, structure of the Scottish Highlauds, 153.

Latent heat of water, 266.

Latent image, formed at low temperatures, 202; Bose's theory of,
203; photo-electric theory of, 204, 209 _et seq._

Least action, law of, 66.

Lembert and Richards, atomic weight of lead, 27.

Length of life dependent on conditions of structural development,
93; dependent on rate of reproduction, 94.

Life-curves of organisms having different activities, 92.

Life, length of, 91.

Life waves of a cerial, 95; of Ausaeba, 87; of a species, 90.

Light, effects of, in discharging negative electrification, 204;
chemical effects of, 205; experiment showing effect of, in
discharging electrified body, 205.

Lindemann, Dr., duration of solar heat, 29.

Lowell, Percival, observations on Mars, 167 _et seq._; map of
Mars, reliability of, 198.

Lucretius, birth-time of the world, 1.

Lugeon, formation of the Préalpes, 171; sections in the Alps,
154.

Lyell, uniformity in geological history, 15.

M.

Magee, relative areas of deposition and denudation, 16.

Mars, climate of, 170; position in solar system, 174, 175;
dimensions of satellites of, 177; snow on, 169; water on, 169;
clouds on, 169; atmosphere of, 170; melting of snow on, 170;
dimensions of canals, 171; signal on, 172; times of opposition,
164; orbit of, 165; distance from the Earth, 165; eccentricity of
his orbit, 165; observations of, by Schiaparelli, 165, 166;
Lowell's observations on, 167 _et seq._

Maxwell, Clerk, changes made under constraints, 65; on
conservation of energy, 61.

M'Connel, J. C., viscosity and rigidity of ice, 284.

Memory, physiological, 111, 112.

Metamorphism, thermal, in Alpine rocks, 132, 149

Millicurie, definition of, 256.

Molasse, accumulations of, 148.

Morin, coefficients of friction, 265.

Morphy, H., experiments on coefficient of friction of ice, 281.

Mountain-building and the geosynclines, 119-121; conditioned by
radioactive energy, 125; energy for, due to gravitation, 122;
reduction of surface attending, 123; depression attending, 123;
instability due to thermal effects of compression, 132; igneous
phenomena attending, 132; rhythmic character of, accounted for,
133; movements confined to upper crust, 122; movements due to
compressive stresses in crust, 122; movements, rhythmic character
of, 121.

Mountain ranges built of sedimentary materials, 118.

Müller, J., coefficient of friction of skate on ice, 265, 274.

Muth deposits of India, 135.

N.

Newton, Professor, of Yale, on origin of Mars' satellites, 177.

Nucleus, dimensions of, 237; amount of radium in, 238.

Nummulitic beds of Himalaya, 138.

O.

Ocean, amount of rock salt in, 50; cause of black colour of, 55;
estimated mass of sediments in, 48; increase of bulk due to
solvent denudation, 52; its saltness due to denudation, 41.

Old age and death, 82-85; not at variance with progressive
activity, 83.

Organic systems, origin of, 78.

Organic vibrations, 86 _et seq._

Organism and accelerative absorption of energy, 79; and economy,
109-111; and periodic rigour of the environment, 94,95.

Organism and sleep, 95; ultimate explanation of rythmic events
in, 96, 97; law of action of, 68 _et seq._; periodicity of; and
law of progressive activity, 82 _et seq._

P.

Penjal traps, 135.

Pepys and skating, 273.

Perry, coefficient of friction of greased surfaces, 265.

Phobos, dimensions of, 177; orbit of, 177.

Photoelectric activity and absorption, 207; persists at low
temperatures, 208, 209; not affected by solution, 213.

Photo-electric experiment, 205; sensitiveness of the hands, 207;
theory of latent image, 204, 209 _et seq._

Photographic reversal, experiments on, by Wood, 211; theory of,
210.

Piazzi, discovery of first Asteroid, 175.

Pickering, W. H., observations on Mars, 167.

Planet, slowing of axial rotation of, 189.

Plant, expectant attitude of, 109.

Pleochroic haloes, measurements of, 224; theory of, 224 _et
seq._; true form of, 226; radius of, and the additive law, 225;
absence of actinium haloes, 225; see _also_ Haloes; mode of
occurrence of, 223 _et seq._

Poole, J. H. J., proportionality of thorium and uranium, 26.

Poulton, uniformity of past climate, 17.

Pratt, Archdeacon, and isostasy, 53.

Préalpes, exotic nature of, 150, 151.

Prematerial universe, nature of a, 297, 300.

Prestwich and thickness of rigid crust, 128; history of the
Pyrenees, 140.

Primitive organisms, interference of, 89; life-curves of, 88.

Proctor and orbits of Asteroids, 176.

Protoplasm, encystment of, 68.

Purana Era of India, 134.

Pyrenees, history of, 140.

R.

Radioactive elements concerned in mountain building, 125.

Radioactive layer, failure to account for deep-seated
temperatures, 127; assumed thickness of, 128; temperature at base
of, due to radioactivity, 129; in the upper crust of the Earth,
125; thickness of, 126-128.

Radioactive treatment, physical basis of, 251.

Radioactivity and heat emission from average igneous rock, 126;
rarity of, established by haloes, 241, 243.

Radium, chemical nature and transmutation of, 244-245; emanation
of, 245; rays from, 253, 254; table of family of, 253; period of,
253; small therapeutic value of, 254.

Radium C, therapeutic value of, 254; rays from. 254; generation
of, 254.

Rationality, conditions for development of, 163.

Rays, similarity in nature of gamma, X, and light rays, 248;
effects on living cell, 251; penetration of, 251.

Reade, T. Mellard, finding age of ocean by calcium sulphate, 13.

Recumbent folds, formation of, 155 _et seq._

Regelation, 284; affecting glacier motion, 285.

Reversal, photographic, explanation of, 211.

Richards and Lembert, atomic weight of lead, 27.

Richter, Jean Paul, Dream of the Universe, 200.

Rock salt in the ocean, amount of, 13.

Rocks, average composition of, 43; radioactive heat from, 126;
rate of solution of, 36.

Russell, I. C., river supply of sediments, 10.

Rutherford, Sir E., determination of age of minerals, 19, 20; age
of rocks by haloes, 22; derivation of actinium, 226; artificial
halo, 229; number of alpha rays from one gram of radium, 237.

S.

Salt range deposits of India, 134. 135.

Saltness of the ocean due to denudation, 41-46.

Salisbury (and Chamberlin), the Larimide range, 121.

Salmon, Rev. George, on creation, 301.

Satellite, velocity of, in its orbit, 191; method of finding path
of, over a rotating primary, 189 _et seq._; direct and
retrograde, 178; ultimate end of, 178; path of, when falling into
primary, 179; effect of Mars' atmosphere on infalling satellite,
179; stability of close to primary, 180; effects of, on crust of
primary, 180 _et seq._

Schiaparelli, observations on Mars, 165 166.

Schmidt, C., original depth of Alpine layer, 131-148; structure
of the Alps, 152.

Schmidt, G. C., on photo-electricity, 207, 208; effect of
solution on photo-electric activity, 213.

Schuchert, C., average area of N. America during geological time,
16.

Sedimentary rocks, average composition of, 43; mass of,
determined by sodium index, 47.

Sedimentation a convection of energy, 133.

Sediments, average river supply of, 11; on ocean floor, mass of,
48; average thickness of, 49; precipitation of, by dissolved
salts, 56-58; radioactivity of 130; radioactive heat of,
influential in mountain building, 130, 131; rate of collecting,
7; determination of mass of, 8; river supply of, 10; total
thickness of, 6.

Semper, energy absorption of vegetable and animal systems, 78.

Sensitisers, effects of low temperature on, 210.

Simplon, radioactive temperature in rocks of, before denudation,
132.

Skates, early forms of, 273; principles of construction of, 273
_et seq._; action of, on ice, 276; bite of, 278-280.

Skating not dependent on smoothness of ice, 260; history of,
273.

Skating only possible on very few substances, 279.

Soddy, F., on isotopes, 24.

Sodium, deficiency of, in sediments, 44; discharge of rivers,
14.

Soils, formation of, 37-39; surface area exposed in, 39.

Sollas, W. J., age of Earth by sodium in ocean, 14; thickness of
sediments, 6.

Spencer, on division of protoplasm, 67.

Spores, number of molecules in, 97.

Stevenson, Dr. Walter C., and technique of radioactive treatment,
259.

Stoletow, photo-electric activity anal absorption, 207.

Stopping power of substances with reference to alpha rays, 219.

Struggle for existence, dynamic basis of, 80.

Strutt, Prof. the Hon. R. J., age of geological periods, 20;
radioactivity of zircon, 223.

Sub-Apennine series of Italy, 148.

Suess, nature of earthquakes. 143.

Survival of the fittest and the organic law, 80.

T.

Talchir boulder-bed, 136.

Temperature gradient in Earth's crust, 126.

Termier, section of the Pelvoux Massif, 254.

Tethys, early extent of, 135-137; geosynclines of, 142.

Thermal metamorphism in Alpine rocks, 132, 149.

Thomson, James, prediction of melting of ice by pressure, 267.

Thorium and uranium, proportionality of, in older rocks, 26.

Triple canals, formation of, by attraction of a satellite, 187.

Tyndall, colour of ocean water, 55.

U.

Uniformitarian view of geological history, 15-18.

Universe, simultaneity of the, 293-295.

Uranium-radium family of elements, table of, 253.

V.

Val d'Hérens, earth pillars of, 33.

Van Tillo, nature of continental rock covering, 9.

Vegetable and animal systems, relative absorption of energy of,
78.

Vegetative organs, struggle between, 105, 106.

Volcanoes and mountain ranges, 118; associated with geosynclines,
142; Oligocene and Miocene of Europe, 147.

W.

Weinschenk and thermal metamorphism, 132,

149.

Weismaun, encystment of protoplasm, 68; length of life and
somatic cells, 96; origin of death, 83; tendency to early
reproductiveness, 98.

Wilson, C. T. R., visualised alpha rays, 218.

Winchell, progressive changes of matter not eternal, 302.

Wood, R. W., on photographic reversal, 211.

Z.

Zircon, radioactivity of, 223; as nucleus of halo, 223.