simultaneous storage of the parent element.
These objections unfortunately apply in some degree
to any substance of secondary origin. Thus if we have
to deal with a mineral substance probably derived by
metasomatic change of a pre-existing rock, we have of
necessity to ask if it is certain that the waters con-
cerned in the genesis of the mineral brought in the
uranium and radium, and that these elements were not
And again, may not the
latter element have taken up its position in quantities
beyond the true equilibrium amount ? Here again we
seem entitled to ask for the corroboration of results,
before we conclude that these sources of error are
absent. In a word, when we examine into the claims
of the radioactive method, we find that, in a certain
sense, we postulate a degree of uniformity between
present and past conditions which might give pause
to the most advanced uniformitarian geologist. It has,
however, this advantage over any other known method
of approaching the problem of the earth's age, that
each experiment is, even if uncertain, an independent
estimate, and that we may make as many of them as
we please. Thus it may result that out o a large-
number we may haply find a degree of coincidence
sufficing to confer complete credibility on the evidence.
The production of helium from thorium, and the pos-
sibility of helium so produced making an appreciable
228 RADIOACTIVITY AND GEOLOGY.
' part of the whole amount, have been carefully considered
by Strutt. 1 He has brought forward much evidence to
show that this may indeed be the case. It may be
added that the fact that the alpha ray is a helium atom
appears to set at rest the sometime disputed question
whether thorium could indeed give rise to helium.
Direct evidence of the genesis of helium by thorium
has also been obtained by Soddy. 2 Obviously if this
source of helium is of moment, its neglect in any par-
ticular experiment is in the direction of producing an
error of excess in the age.
The idea that lead might be the ultimate disinteg-
ration product of uranium has been made the basis
of an attack on the age problem. In a paper of great
interest Boltwood 3 records that in unaltered primary
minerals from the same locality, and presumably of the
same age, the amount of lead was proportional to the
amount of uranium present ; and, further, that the
amount of lead relative to the uranium was greatest in
minerals from those localities which were of the greatest
geological age. This, it will be admitted, is much in
favour of the view that lead is indeed a disintegration
product of uranium.
On the foregoing basis Boltwood arrives at the age of
a number of uranium- bear ing minerals. His figures,
when corrected according to more recent data, respect-
ing the constants involved (the proportion of uranium
- 1 Strutt, Proc. JR. S., A 76, pp. 101, 312.
2 Soddy, Phil. Mag., October, 1908.
3 Boltwood, Am. Journ. Science, February, 1907, p. 77.
URANIUM AND THE AGE OF THE EARTH. 229
to radium, and the rate of decay of radium), range
from 246 to 1320 millions of years. The geological
positions of these minerals are not cited; the oldest,
from Norway and Ceylon, are presumably from primi-
It is not improbable that the radioactive method of
determining the age will receive a useful development
in the procedure adopted in this research. Although
the introduction of radium, etc., unattended with
uranium or the introduction of radioactive materials
at dates subsequent to the formation of the substance,
would give rise to larger errors than in the method
based on the helium ratio, on the other hand, there
would probably be more security against error from
the loss of the disintegration product chosen for
The apparent tendency of Strutt's results is to assign
ages to the older sedimentary deposits greater than
would be deduced by the legitimate use of any other
method of estimating the age of the earth with which
we are acquainted. Thus Strutt selects the following
results as showing the bearing of his measurements,
regarding them as minor limits :
Phosphatic nodules of the Crag, . . 225,000
Upper Greensand, 3,080,000
Haematite overlying Carboniferous Lime-
stone ...... 141,000,000
If the last of these estimates is correct, there can be
230 EADIOACTIVITY AND GEOLOGY.
little doubt that the Huronian must date back some
400 millions of years, and the " geological age " would
probably be somewhat greater still. It is, therefore,
of importance to consider whether those existing
methods, which have hitherto been interpreted as
indicating from 50 millions to 100 millions of years
as the period since denudation began, can be so far
There exist, in fact, but two methods directly deal-
ing with the denudation period : that which is based
on the records of detrital denudation, and that which
is based on the records of solvent denudation.
The first method seeks to find the period since
sedimentation began by taking as numerator the maxi-
mum total depth of sedimentary deposits, and as
denominator the rate of accumulation. The quotient
is the period sought.
It must be admitted that grave difficulties attend
the determination of both the numerator and denomi-
nator of this fraction :
(a) What is the maximum thickness of the sedi-
ments ? We are first met with the difficulty of not
knowing where to begin. The more ancient sediments
are apparently to be found in those complexes of
igneous and sedimentary materials wherein the dif-
ferentiation of the two classes is all but, or quite,
impossible. About the origin of such rocks we find
the most experienced observers differing. Again,
even when the evidence is adjudged decisive as to
an igneous origin, this may really denote no more
URANIUM AND THE AGE OF THE EARTH. 231
than the complete igneous change of pre-existing
(I) A fine-grained deposit will take a far longer
time to collect than an equal thickness of coarser
materials. A whole formation will include layers of
many different sorts. The time-values of these are
very various; some are of slow, some of rapid depo-
sition, suggesting that the area of maximum rate of
deposition was not fixed in position. In seeking the
denominator (the rate of accumulation) this must be
remembered or serious error will be incurred.
(c) The unconformities often represent great lapses
of time, which not only may be unrepresented, but may
even have resulted in the removal of earlier deposits.
As an instance, consider the almost complete removal
of the mountain- building sediments of the Alps. This
error may or may not be made good in the measurements
at our disposal. Certain deposits are very perishable.
The Chalk is an instance. This substance is not only
perishable, but it is, relatively, a slow-forming deposit.
In a few million years, if exposed so long to denudation,
all records of such formations might be removed. Thus
the time record would be falsified.
To these sources of error might be added such as
may arise by the conditions being homotaxial rather
than contemporaneous. For if Huxley's case have a
semblance of the truth, and it be possible that the
Devonian period in England was overlapped by the
Silurian in America, and in turn overlapped the
Carboniferous in Africa, we evidently are not entitled
232 RADIOACTIVITY AND GEOLOGY.
to sum the several depths of deposit in these countries
as successive time intervals. This latter objection has
hardly more than academic interest at the present
moment. The radioactive method of investigating
the ages of the strata may yet put Huxley's case to
the test. Finally the practical difficulty of measuring
the thickness of strata has to be faced. The difficulty
of the task is shown in the diversity of opinion as to
many of the recorded measurements.
Supposing we have satisfactorily determined the
numerator, we are by no means done with our
troubles. "We have to find the rate at which the
rivers are carrying detrital sediments to the ocean.
This involves measurements of the silt-load of rivers,
at various seasons. The seasonal variation is great,
because the same cause which raises the river to
flood-level increases the supply of suspended matter.
The burden carried may be five-fold as much at one
time as at another time. 1
We have also to find the rate at which this silt
grows in depth when it comes to rest in the ocean.
And here is perhaps the most serious source of error, for
not only is the catchment area difficult to define, but
the troublesome question remains as to whether we
should reckon the silt as laid down over the whole
area, and take this mean rate of deposition, or, having
found this average rate of accumulation, proceed to in-
crease it in order to arrive at the maximum rate, having
in view the fact that we have taken the sedimentary
1 See table in Russell's River Development, p. 72.
URANIUM AND THE AGE OF THE EARTH. 233
records at their maximum thickness. The final result
will vary considerably according to which procedure
we adopt. Thus while Geikie 1 is satisfied that the
whole geological column may have taken about 100
million years to accumulate, and De Lapparent, 2 on a
rather vague definition of the area of deposition, arrives
at seventy-five to eighty millions, Sollas, 3 adding more
than 100 per cent, to the mean rate of collection to
represent the maximum, obtains twenty-six millions.
But it would appear as if some moderate increase of the
mean rate was the more correct procedure; for, as
already pointed out, when reckoning up the beds we
include both fast- and slow-forming deposits ; admitting
thereby that the area in question was at one time the
place of rapid growth and at another time the place of
slow growth. In other words, the deposit is not truly
wedge-shaped, but rather takes the form of a shallow
basin. Hence, unless we reckoned coarse detrital rocks
exclusively, we are not entitled to assume the fastest
rate of deposition for the whole formation.
The foregoing method of finding the age, which has
given rise to so much discussion, appears to afford its
answer by a somewhat different mode of approach
from that which is generally adopted. We take first
the average ratio of the mass of suspended sediment to
the mass of the river water.
1 Geikie, Brit. Assoc. Rep., Address to Sect. C, 1899.
2 Traite de Geologic, vol. iii., p. 1860.
3 Sollas, Brit. Assoc. Rep., Address to Sect. C, 1900.
RADIOACTIVITY AND GEOLOGY.
The following table contains estimates of this
factor 1 :
Po, . .
It will be seen that the ratio of the weight of
sediment to the weight of water in the cases of all
the rivers cited, save two, does not differ very widely.
The arithmetic mean of the ratios, as given in the
table, fails to allow for the relative importance of
the rivers. When we ascribe to the ratio given above
for each separate river the weight proper to the dis-
charge, the mean ratio of sediment to water is found
to be 1 : 2521.
If we now take Sir J. Murray's estimate of the
total annual discharge of the world's rivers, 27 x 10 3
cubic kilometres, we find that the total carriage of
detrital sediments to the ocean is about 4*4 cubic
1 Russell, loc. cvt., p. 14.
URANIUM AND THE AGE OF THE EARTH. 235
kilometres per annum. This is for the whole earth :
we have now to find an approximate measure of the
area of deposition. On the basis of the measure-
ments made in the case of the Mississippi, 1 we take the
area of deposition to be one-eighteenth that of denu-
dation. To find the latter we have first to subtract
from the land area of the globe the rainless areas.
These are estimated as making one-fifth the total
area (which is 14*4 x 10 7 square kilometres), leaving
11*5 x 10 7 square kilometres; and one-eighteenth of
this receives the detritus, i.e., 6*4 x 10 6 square kilo-
metres. Spreading the sediment over this, we get a
rate of growth of 6*9 centimetres in 100 years. This
rate we increase by 50 per cent, to allow for the fact
that we compute the depth of sedimentary deposits
where these are most developed.
According to Sollas 2 we may reckon the geological
column as about eighty kilometres. The geological
age derived would be about eighty million years. The
neglect of the calcareous precipitates in this computa-
tion tends to increase somewhat the estimated age.
The question we have to ask here is : Can we be
four- or five-fold in error in such an estimate ? In the
foregoing pages we have been at pains to set forth the
directions in which this method may err. But con-
sidering each weak point separately, we must admit
the great improbability of failure so complete. The
integral of the sediments can scarcely be 100 miles
instead of fifty miles thick as estimated; much less
1 Magee, Am. Journ. of Science, 1892, p. 188.
2 Sollas, The Age of the Earth, p. 14.
236 RADIOACTIVITY AND GEOLOGY.
can it be really 200 miles. Can the ratio of the
areas of denudation and deposition be on the average
four or five times greater than what has been measured
in the case of the Mississippi? Or can the mean found
from the nine rivers for the ratio of silt to water be
equally far out? Or, finally, can the several rates
have been so much less in the past? The answer in
each case seems to be a negative, and only on the
chance that several of these factors have been computed
with error of the same sign, does it appear possible to
reconcile the indications of the sedimentary accumu-
lations with the storage of the products of radioactive
The probability that such an accumulation of errors
has occurred is greatly lessened by the fact that a really
quite independent mode of computing the geological
age gives a similar result.
The method of estimating the geological age by the
accumulated results of solvent denudation requires
knowledge of two quantities : the quantity of sodium
discharged per annum into the ocean by all the rivers
of the world, and the amount of this element in the
sea. The latter is evidently the numerator and the
former the denominator of the fraction, the quotient
of which is the sought-for geological age.
As regards the determination of the numerator
that is the saltness of the sea the correct value,
within a few per cent., is known. Two quantities
are involved in it: the average composition of sea-
water and the mass of the ocean. Our knowledge of
the first is favoured by the fair degree of uniformity
URANIUM AND THE AGE OF THE EARTH. 237
in chemical composition prevailing in all the great
seas; our knowledge of the latter by the fact that
soundings are certainly sufficiently numerous to afford
the average depth of the ocean : undiscovered deeps
or shallows are not likely to seriously affect the
volume inferred from the average depths.
In the geological application of this knowledge we
are, however, exposed to the possibility that the early
ocean received salts at an accelerated rate owing to
the prevalence of an atmosphere possessing acid proper-
ties. It will be shown that error from this source is
calculable and its limiting amount definable. Neglect
of this possible correction renders our estimate of
geological time proportionately too long, for it is a sub-
tractive correction on the numerator. The reception,
by the ocean, of salts dissolved directly from the coasts,
or from the submerged sediments, gives rise to yet
another correction which can be shown to be probably
small. Its neglect again tends to an unduly great
estimate of the period of denudation. The rock-salt
deposits are obviously negligible as a source of error.
This will appear from the fact that the rock salt in the
ocean, if spread upon the dry land of the globe, would
cover it to a depth of 122 metres (400 feet). The
quantity of sodium now in the ocean is about
1,555 x 10 13 tonnes. The allowance to be made for
the sources of error referred to we consider later.
The denominator (the sodium content of the rivers)
is not so well known. It is indeed quite determinable,
but we have to go on insufficient knowledge at present.
We have Sir John Murray's estimate of the total
238 RADIOACTIVITY AND GEOLOGY.
amount of river water, annually discharged into the
ocean, as 27,191 cubic kilometres ; and his average
analysis of 19 rivers, many of them principal rivers
of the world. Combining these data, we find that the
total amount of elemental sodium discharged yearly is
15,976 x 10* tonnes. As the 19 rivers do not repre-
sent a sufficiently large fraction of the total drainage
of the continents, this figure cannot be considered final.
The most important correction on the denominator
is concerned with the cyclical nature of part of the
sodium. 1 For some part of it is believed with consider-
able probability to be wind-borne from the sea, reaching
the rivers in the rain-water. The facts connected with
and giving rise to this belief may be briefly stated.
(a) The rain falling on islands and coastal countries
carries down with it a considerable proportion of
chlorides and other salts, which might be referred to
oceanic origin. The proportion of such salts in rain
falls off as we go inland. Thus 200 miles from the
coast there may be no more than -fa of what would
be observed in rain-water collected 30 or 40 miles
from the sea. In India, 300 miles from the sea, a
measurement showed chlorine to the extent of only
0'04 part in 100,000 ; whereas in British rain-water it
may be from O'l to 21 parts per 100,000, and even
more. The rate of decrease is, apparently, rapid as we
proceed inland. Kemote from the sea some of the
chlorine in rain-water must be due to dust raised from
the ground by winds : in arid, saline regions this might
affect considerably the results of rain- water analyses.
1 July, Gtoloyical Magazine, Aug., 1901, p 341.
URANIUM AND THE AGE OF THE EARTH. 239
(b) Coastal rivers arid the rivers of small islands
show abnormal percentages of chlorine compared with
the sodium, indicating that but a small fraction of the
latter is referable to the rocks, and the greater part in
these cases to the supply from the rains.
(c) The chlorine in average river- water is more than
the observed amount of this element in the rocks
would warrant. Thus Clarke, as an approximate
estimate, gives the percentage of chlorine in the igneous
lithosphere as 0*07. There is apparently no measurable
amount in detrital rocks ; and in limestones, which con-
stitute only some 5 per cent, of the sediments, there
is very little, so that we must conclude that in
the process of denudation the whole of the chlorine
passes into solution. Now the sodium in igneous rocks
averages (according to Clarke) 2*43 per cent., and of this
it can be shown that about one-half, or a little more,
goes into solution. Thus we should expect that the
break-up of an average igneous rock would liberate
chlorine and sodium in the ratio of 0'07 to 1*4 (say), or
as 1 to 20, about. However, the rivers show a pro-
portion of chlorine to sodium which, on the average, is
1 to 2. Hence there is about ten times as much
chlorine as would be inferred from the mean chlorine
content of igneous rocks, and this fact is taken as in
favour of the view that much of the chlorine of rivers
is wind-borne from the ocean.
This last argument must not be pushed too far.
Clarke's estimate of the elements in average igneous
rocks excludes extravasated or non-magmatic sub-
stances. It is based on typical .rock analyses. Now
240 RADIOACTIVITY AND GEOLOGY.
chlorine, as it happens, enters with difficulty into
the constitution of silicates. We find it accordingly
free (or forming hydrochloric acid) in association with
melted magmas in volcanoes, and rising in springs in
volcanic regions. And it is found in combination with
metallic minerals and in springs associated with ore
deposits. The sources of supply are in these cases
probably deep-seated, and would not enter into rock
analyses. Possibly the slow leaching of the rocks
within the hydrosphere is ultimately responsible ; but,
however this may be, it is evident that, so far as the
river supply is concerned, we require to include all
these sources. Nor must we exclude the ancient
relics of congested denudation which appear in old
salt lakes, bad lands, and probably in saline waters
found in deep-seated bedded rocks. Again, if we go
upon this argument, we may, in the case of certain
rivers, arrive at the absurdity that no sodium is being
denuded : for cases occur where (if the analyses are
correct) we have insufficient sodium to combine with
the chlorine ; so that if the chlorine and sodium reach
the rivers from the ocean, there can be no sodium
arising from solvent effects. This, of course, is an
untenable view, and demonstrates the unsoundness of
the argument which gives rise to it.
The possibility must be kept in mind that chlorine
is liberated from the ocean unattended by the less
volatile elements. The escape of free chlorine, or,
more probably, of chlorine combined with hydrogen,
seems by no means improbable. In justification of such
a view we have to consider the very large amount of
URANIUM AND THE AGE OF THE EARTH. 241
chlorine present ; and that if the chloride of hydrogen
was formed by fortuitous ionic combinations near the
surface especially in agitated water a certain per-
centage of the molecules should escape. The radioactive
substances in the ocean will, probably, be instrumental
in liberating a certain amount of free chlorine ; but it
is not hard to show that this contribution must be
insignificant. On this view the chlorine of rain-water
might only to a subordinate degree be present as sodium
chloride. This, if true, would reconcile the various
observations in the simplest manner.
Admitting, however, that some part of the chloride
of sodium of rivers is derived from the ocean, conveyed
by the winds and falling as rain, the question arises
what to allow for this source of error; for evidently
we cannot include wind-borne sodium in the calculation
of the geological age. The only really satisfactory mode
of investigating the amount of the wind-borne sodium
would be by rain-water analysis. Unfortunately we
possess very little knowledge on this subject. Let us
proceed, however, on the assumption that the chlorine
of inland rains is a clue to the cyclical sodium. The
great bulk of the river- water of the world is collected
at distances greater than 300 miles from the sea. If
we take 0*04 per 100,000 as the proportion of chlorine
in rain-water, and, making the usual allowance for
concentration by evaporation before it reaches the
rivers, conclude that in rivers, chlorine to the extent
of O'l in 100,000 is wind-borne ; and, further, that this
chlorine enters the rivers with the same proportion of
combined sodium as it would show in sea-water, we
242 RADIOACTIVITY AND GEOLOGY.
can apply the appropriate correction to Murray's
estimate of river sodium.
Following Murray, we have 84'7 x 10 6 tonnes of
chlorine annually discharged by rivers into the ocean,
and this chlorine is contained in 27'2 x 10 12 tonnes of
water. This is a proportion of about 0'3 in 100,000.
The correction then amounts to taking one-third of the
chlorine as coming from the ocean. We have to find
what correction this is on the sodium. Now, in the
ocean the chlorine is considerably in excess of the
sodium, so that if all the sodium in the ocean is
assumed to exist in the form of the chloride, only
82 per cent, of the chlorine is used. Hence, ocean-
water contains at least 18 per cent, of its chlorine
otherwise attached than to sodium : and, accordingly,
what circulates from the sea to the land can only bring
with it to the rivers 82 per cent, of the equivalent
One-third of the chlorine annually discharged by
the rivers, on the foregoing basis, represents 15 x 10 6
tonnes of sodium. The true denudative supply of