24
28
30
pende
ad
nit
s ind i
s of r
le 2: I
ount of
al an
atic
xamp
m
E
A
parent
32
16
8
4
2
er
em
ath
:
atio
ge of a min
e m
er
e a
ghtau
g th
arent = 100
D
parent r
0
1
3
7
15
minin
ation of th
ter
ount of p
er
str
ount of
ght
o de
llu
ial am
mA dau 0 50
75
87.5
93.75
ey t
le i
nit
e k
s th
.
imp
le 1: I
sent
S
ount of
atio i
1.
xamp
m
E
A
parent
100
50
25
12.5
6.25
ial pre
arent r: p ater
s
edm
er
ive
or
ABLE 2.
ght
T
alf-l
al f
au
arent m
ryst
e d
e in him ce c
te: Th
ount of p
T
sin
0
1
2
3
4
No
am
An atlas of time 35
the parent, which is essentially the probability that any given
atom will decay in a certain amount of time— analogous to one’s
chances of winning the lottery in any year. The units of the
decay constant are thus 1/time. Rutherford had estimated the
decay constant for uranium from the number of radioactive
emissions detected from a mass of uranium in a given time in-
terval. The decay constant is inversely proportional to the more
familiar idea of a half- life— the time it takes for half the parent
material to decay to the daughter form. In other words, a small
decay constant (low probability of a lottery win) means a long
half- life (long wait to get rich), while a large decay constant
means a short half- life (easy money!).
By 1911, in spite of the still- rudimentary understanding of
the phenomenon of radioactivity and rather primitive labora-
tory facilities, Arthur Holmes had determined the absolute ages
of a half- dozen igneous rocks whose relative ages on the fossil-
based geologic timescale were bracketed by their relationships
with sedimentary rocks. Three samples were from the fossil-
iferous Paleozoic and three from the murky, undifferentiated
Precambrian. Even though some of the lead Holmes measured
was not from the decay of the parent uranium but from another
radioactive element, thorium, his dates are amazingly close to
modern values (within tens of millions of years).
The very first rock analyzed, a granite from Norway thought
to have formed in the Devonian Period (based on its cross-
cutting relationships with fossil- rich sedimentary strata),
yielded an approximate age of 370 million years— 18 times
longer than Kelvin’s estimates of the age of the Earth. And a
Precambrian metamorphic gneiss from Ceylon (Sri Lanka) was
found to be 1.64 billion years old— two full orders of magnitude
greater. Darwin’s intuition was vindicated. Holmes would go on
to become one of the preeminent geologists of the twentieth
36 Ch a pter 2
century. Kelvin’s long- reigning proclamations became imme-
diately irrelevant, because radioactivity not only provided a
means of directly dating rocks but was also a source of internal
heat that Kelvin had not incorporated into his calculations of
the rate of planetary cooling. (Years later, Holmes would chal-
lenge another of Kelvin’s fundamental assumptions, arguing
that Earth cools mainly by convective, rather than conduc-
tive, heat loss). Most important, the geologic timescale could
now be calibrated. Even the deepest reaches of geologic time
could be fathomed; the Precambrian would no longer be an
uncharted primordial wilderness.
T H E G E N E R A L S E D I M E N T
In reality, it would take many more decades for the new science
of geochronology (Earth time) to mature. The use of radio-
active isotopes as high- precision geologic clocks required ad-
vances in nuclear physics, cosmochemistry (which concerns
the stellar origins of the elements), petrology (study of igneous
and metamorphic rocks), and mineralogy, as well as the devel-
opment of new analytical instruments, particularly mass spec-
trometers capable of distinguishing among isotopes of a single
element. There was also the nontrivial problem that geologic
timescale so laboriously built by the Victorians, with fossils as
timekeepers, was entirely based on sedimentary rocks. Any
isotopic dates derived from these would reflect not the age of
the sedimentary deposit but the time of crystallization of the
igneous or metamorphic precursors from which its grains were
derived. Assigning absolute ages to the fossil- based timescale
has thus required finding serendipitous outcrops where sedi-
mentary rocks of well- constrained biostratigraphic age happen
to be interlayered with, or cut by, igneous rocks in such a way
An atlas of time 37
F I G U R E 4 . Cross- cutting relationships between igneous and sedimentary rocks allow calibration of the fossil- based timescale.
that allows an isotopic age to be tied directly with the fossil
record (figure 4). Volcanic ash layers are ideal for this purpose, since they represent fresh igneous crystals that fell from the air
in a geologic instant and were interleaved with the sedimentary
and paleontologic archive of their day.
Ash layers within sedimentary strata reveal a subtle but fun-
damental idea about the way in which the rock record is writ-
ten. Looking at layered rocks like the extraordinary sequence
in the Grand Canyon, one tends to imagine that each stratum
accumulated in the manner of a snowfall, blanketing a given
38 Ch a pter 2
area all at once, in a well- defined period of time. But this is not
necessarily the right way to think about rock layers. Consider
the beautiful white, almost pure- quartz St. Peter Sandstone
of Ordovician age, exposed along river valleys in Minnesota,
Iowa, Wisconsin, and northern Illinois. The St. Peter forms the
picturesque hol
low at Minnehaha Falls in Minneapolis and was
for decades the source of silica for window glass made at the
Ford plant in Saint Paul. During Prohibition, natural pockets
in the St. Peter along the Mississippi River were enlarged into
a network of caverns that housed speakeasies and secret ware-
houses beneath the Twin Cities.
The St. Peter Sandstone is crumbly, hardly even a proper
rock, and when it falls apart into uniform, rounded grains in
one’s hand, it is easy to see that this is an ancient beach sand.
But the St. Peter is found at the surface in four states, and is
known from drilling to continue beneath Michigan, Indiana,
and Ohio. No beach would cover such a vast area at any partic-
ular moment. Instead, the St. Peter records the gradual migra-
tion of beaches across the land surface as ancient shallow seas
waxed and waned over millions of years. One day in the Ordo-
vician, clouds of ash from a supervolcano eruption in the infant
Appalachian Mountains hundreds of miles away fell out of the
air over the midcontinent seas, leaving a thin, greenish clay
layer across the region that is like a clearly dated diary entry.
In some places, the ash occurs near the top of the St. Peter, but
elsewhere, the sandstone lies far below this level, having been
buried long before by other sediments at the time the volcano
erupted. Thus, although the unmistakable St Peter sandstone
is a continuous layer for hundreds of miles, it is not the same
age everywhere. The more general idea is that except for layers
that mark sudden regional or global events, like a great eruption
or a meteorite impact, laterally extensive sedimentary units
An atlas of time 39
are not strictly isochronous— markers of the same moment
in time. Instead, they record the slow march of depositional
settings across the Earth’s surface over time, as sea levels and
environmental conditions changed. In geologic parlance, they
are diachronous— that is, they transect time.
T H E T I M E B U R E A U C R AT S
These days the geologic timescale is not merely a chart or even
a multivolume treatise but a gigantic digital database that is
administered by the formidable International Commission
on Stratigraphy (ICS), the oldest and most important body
within the International Union of Geological Sciences. The ICS
maintains strict rules about how geologic units are named and
defined, and it catalogues outcrops, rock formations, fossils,
isotopic dates, geochemical data, and analytical protocols, in
the never- ending task of mapping geologic time at higher and
higher resolution.
Since the 1970s, the ICS has sought to identify specific
sites around the world to serve as the international standards
for the boundaries of each division of the geologic timescale.
Such an outcrop is formally called a Global Boundary Strati-
graphic Section and Point, or GSSP, but known colloquially
among geologists as a “golden spike.” These sites must have
well- exposed rocks with biostratigraphically diagnostic fossils
that straddle the boundary between the two time intervals,
and they must be in places that can be protected from de-
velopment or destruction. The location of the exact stratum
representing the boundary at a given GSSP is often described
in charmingly idiosyncratic detail. For example, the golden
spike outcrop for the Cenomanian division of the Upper Cre-
taceous lies high in the French Alps and begins “36 meters
40 Ch a pter 2
below the top of the Marnes Bleues Formation on the south
side of Mont Risou.”7
The primary divisions of the geologic timescale— the eons,
eras and periods— were largely defined by the work of Brit-
ish geologists in the nineteenth century, and the names of the
Paleo zoic Periods more strongly reflect that geographic influ-
ence: Cambrian, from the Latin name for Wales; Devonian for
the county of cream teas; Carboniferous for the coal measures
of northern England. But the finer subdivisions— the epochs
and ages— reveal the subsequent, wholly international nature
of the time- mapping project: the Jiangshanian and Guzhangian
of the Cambrian; the Eifelian and Pragian of the Devonian; the
Moscovian and Bashkirian in the Carboniferous. The ICS is like
a temporal counterpart to the United Nations— a parliament of
the past, whose jurisdiction is geologic time.
And the ICS, somewhat fussily, insists on maintaining the
subtle but important distinction between time and the rock re-
cord of time. Geologic time is divided into eons, eras, periods,
epochs and ages, and the corresponding rocks into eonothems,
erathems, systems, series, and stages. Similarly, one should say
“Early” or “Late” Ordovician (for example) when referring to
time, but “Lower” and “Upper” when speaking of rocks. Time
( chronos) could happen without rocks (representing kairos),
but not the other way around. However, time vanishes, while
rocks persist.
P L U M B I N G T H E D E P T H S O F T I M E
Arthur Holmes’s early efforts to obtain absolute ages from
rocks, carried out before the structure of the atom and the ex-
istence of isotopes were even known, are analogous to Darwin’s
insights about heredity, which predated the discovery of genes
An atlas of time 41
and DNA. In both cases, it would take years before the rest of
science developed the capacity to explore fully the implications
of their visionary ideas. It was not until the 1930s that the com-
plexity of lead isotope geochemistry was fully understood, one
might say “plumbed.” In 1929, Ernest Rutherford showed that
there were two different parent isotopes of uranium, 238U and
235U, which produced two different isotopes of lead (206Pb and
207Pb, respectively) at the end of long radioactive decay series
with very different overall half- lives (4.47 billion and 710 mil-
lion years, respectively). Soon after, Alfred Nier, a physicist at
the University of Minnesota, identified another lead isotope,
204Pb, which was nonradiogenic— that is, lead that started as
lead, and was not the product of radioactive decay. Nier had
developed the essential instrument in isotope analysis, the mass
spectrometer, which allows isotopes of a single element to be
sifted out according to their atomic weight. And with the dis-
covery of 204Pb, Nier recognized the potential application of
these three lead species for dating rocks, and even the Earth.
Over geologic time, he realized, the abundances of 206Pb
and 207Pb would have grown in a mathematically predictable
way while the absolute amount of 204Pb remained constant. In
particular, the comparatively short half- life of 235U would have
caused Earth’s inventory of 207Pb to increase rapidly early in the
planet’s history, but then flatten off, like the cumulative earn-
ings from a savings account with a high interest rate but from
wh
ich rapid withdrawals are made. Meanwhile, the global stock
of 206Pb would have continued to accumulate from the slower
decay of 238U— like the money earned at a lower interest rate in
an account that is drawn down more slowly. (The unchanging
amount of 204Pb would be like money hidden under a mattress).
In 1940, Nier and his students were about to put these ideas
to the test using geologic samples. The work was interrupted
42 Ch a pter 2
when Nier— the son of German immigrants— was asked by En-
rico Fermi to work on the Manhattan Project, which required
the separation of the fissionable isotope 235U from nonfission-
able 238U.8 Nier’s spectrometer was the only instrument that
could distinguish the two isotopes, and his lab was required
to focus on the uncertain future rather than questions of the
geologic past.
Immediately after the war, however, Nier set about mea-
suring the Pb isotope ratios in deposits of galena (lead sulfide,
PbS), the primary ore of lead, of different ages from around
the world. Galena obviously has plenty of lead in it, but the
lead doesn’t take in uranium when it crystallizes. This means
that lead isotope ratios in galena do not change over time and
should instead reflect the particular mix of lead species that
were available in the environment at the time the mineral
formed. As Nier had predicted, the older samples had lower
ratios of 207Pb/204Pb and 206Pb/204Pb (lead from “interest” vs.
“mattress” lead). If the Earth had started with no 207Pb or 206Pb,
these ratios would be enough to determine the age the planet.
But Nier knew that at the time of its formation, Earth had al-
most certainly inherited some “interest” lead from what had
accumulated in the “bank accounts” of ancestor solar systems.
Thus, determining the age of the Earth required knowing the
primordial ratios of the various lead isotopes.
Nier also recognized a subtler problem: even a very ancient
galena sample would not represent the primordial lead ratios
for the Earth as a whole. Earth is not one uniform geochemical
reservoir, like a planetary milkshake. Instead, it has unmixed
itself over time. In its earliest days, the planet differentiated
into a metallic core of iron and nickel and a rocky mantle that
got most of everything else, including virtually all Earth’s ura-
nium. Ever since then, repeated partial melting of the mantle
An atlas of time 43
has generated the crust, which is much richer in uranium than
either the bulk Earth or the mantle, in the way that butterfat
is concentrated in the cream at the top of a bottle of raw milk.
Timefulness Page 5