Timefulness
Page 7
with the youngest rings in ancient wood preserved in bogs and
archeological sites, the tree ring record can be extended back
more than 10,000 years, and 14C ages can be adjusted accord-
ingly. Growth bands in corals (made of calcite, CaCO3) provide
a somewhat lower- fidelity record than tree rings but make it
possible to calibrate 14C ages still further back in time. Never-
theless, the uncertainties for 14C dates are large— on the order
of hundreds to thousands of years (5% to 10% of an object’s
actual age).
Humans have further complicated radiocarbon dating in two
ways. First, aboveground nuclear tests in the early Cold War
days injected large amounts of 14C into the atmosphere, which
must be corrected for in very recent samples. This is why 14C
ages are typically reported as years before 1950. Second, a cen-
tury of burning fossil fuels with “dead” carbon has shifted the
mix of isotope values in the atmosphere. This is called the Suess
effect, for the Austrian physicist Hans Suess, who first recog-
nized it in 195514 (and who had been working for the German
nuclear program at the time of the Manhattan Project in the
United States). While the Cold War’s 14C will slowly dissipate,
the Suess effect continues to grow.
P R O D I G A L D A U G H T E R S
As mass spectrometers became more accessible to the aca-
demic masses in the late 1950s and ’60s, geochronology came
into its own as a new subdiscipline with dedicated faculty lines
and graduate programs. Among the first isotope systems to be
widely used for geologic dating was the potassium- 40– argon- 40
(40K - 40Ar) parent- daughter pair, because potassium is very
abundant in many igneous and metamorphic rocks, and even
An atlas of time 53
lower- precision instruments could detect both parent and
daughter. The original K- Ar method is still perfectly good for
young rocks with simple thermal histories. It remains an im-
portant tool, for example, in determining the ages of sedimen-
tary deposits containing fossils of human ancestors like “Lucy”
that are conveniently interbedded with volcanic ash layers in
the magmatically active East African Rift Valley.
A problem with the K- Ar system is that the daughter is an
apple that falls far from the parental tree. Potassium is a large,
sociable ion ready to offer an electron to other elements, while
argon is a compact, self- contained noble gas with completely
filled electron shells and no tendency to bond with anything. So
given any chance— a position at the edge of a crystal that allows
an easy exit, a crack that offers a shortcut, a metamorphic heat-
ing event that opens the crystal’s doors to diffusion— daughter
argon atoms will leak out. The calculated age for the host min-
eral will then be younger than the true geologic age, but there is
no way of knowing by how much. The plus or minus value will
reflect the analytical uncertainty arising from the limits of the
laboratory instruments, not the actual imprecision of the date.
The limitations of K- Ar dating began to be especially clear in
the 1960s when the method was applied to old rocks from the
Canadian Shield, which had long, multistage histories of defor-
mation and metamorphism. Age determinations were some-
times inconsistent with field evidence for the relative ages of
rocks. In some cases, so much argon had seeped out of minerals
deep in the subsurface that it lingered in adjacent rocks, leading
to cases in which the K- Ar age determinations were actually
too old. Young- Earth creationists still seize on these ambigu-
ities and suggest that the whole enterprise of geochronology
is hopelessly flawed. But by the 1970s, geochronologists had
developed a powerful variation on K- Ar dating that yields both
54 Ch a pter 2
higher- precision ages and provides information about whether
argon loss (or gain) has occurred.
In the new technique, a potassium- bearing sample is bom-
barded with neutrons, and this converts the 40K in the speci-
men to a short- lived isotope of argon, 39Ar, which then acts as
a proxy for the parent. The sample is next heated slowly in what
amounts to the laboratory equivalent of a metamorphic event.
Both types of argon— 39Ar, representing the parent, and 40Ar,
the daughter produced by radioactive decay— begin to leak out.
As the temperature is increased incrementally, the crystal be-
gins to exhale more argon, which is captured and analyzed in
batches. The 40Ar/39Ar ratio (really the daughter/parent ratio)
is used to determine an apparent age of the sample at each
step. Typically, the ages obtained from the first few samples of
captured argon— representing the outside of the crystal, where
geologic argon escape would have been easiest— are younger
than those for the interior. If the apparent ages obtained with
continued heating stabilize around a consistent value— what
geochronologists call an “40Ar/39Ar plateau age”— then there
is good reason to conclude that the interior of the crystal has
not experienced significant argon loss and that the age is geo-
logically meaningful.
D AT E S W I T H D E S T I N Y
Probably the most famous application of the argon- argon
dating method was the conclusive identification of the crater
formed by the meteorite impact that doomed the dinosaurs at
the end of the Cretaceous Period. The meteorite hypothesis for
the dinosaur extinction was first proposed in 1980 by the father-
son team of Luis Alvarez, a Nobel Prize– winning physicist, and
his son Walter Alvarez, a geologist at Berkeley. Walter had been
An atlas of time 55
working in the central Italian Apennines, where the crinkling of
the crust into recently formed mountains has raised a sequence
of late Mesozoic and early Cenozoic marine limestones above
sea level.15 One of these, the Scaglia Rossa (essentially, “red
rock”)— a beautiful pink limestone that defines the color palette
of many Italian houses, castles, and cathedrals— contains an un-
interrupted chronicle of ocean conditions before, during, and
after the Cretaceous extinction. There are no dinosaur bones in
the Scaglia Rossa, since it accumulated on the seafloor on the
continental shelf of Africa, but the extinction event is clearly
recorded by an abrupt change in the nature and number of
microscopic fossils and by a distinctive half- inch- thick dark-
red layer of clay.
Walter Alvarez wondered how much time this clay layer— a
mute witness to global apocalypse— represented. His father
Luis, another Manhattan Project alumnus, had access to an in-
strument at Lawrence Berkeley Laboratory that could detect
trace elements in materials at the parts per billion (ppb) level.
He suggested measuring the boundary clay’s concentration of
certain rare metals in the platinum group, such as iridium, that
are delivered to earth�
�s surface mainly by a slow but constant rain
of micrometeoritic dust (you can even collect micro meteorites,
many of which are magnetic, from your roof over a period of
months16). The average rate of this metallic “rainfall” over the
past 700,000 years is known from Antarctic ice cores, and as-
suming it was about the same in the Cretaceous, measuring the
metal content of the boundary clay would allow an estimate of
how long it had taken that layer to accumulate. The logic was
essentially the same as that used by Victorian geologists who
attempted to rebut Kelvin: sum up the total amount of accumu-
lated stuff (sediment, or iridium) and divide by the best estimate
of its rate of accumulation to estimate elapsed time.
56 Ch a pter 2
To have some idea of background concentrations of iridium,
the Alvarezes analyzed closely spaced samples not only from
the clay layer but also from the limestone below and above the
boundary. They found that the concentration of iridium went
from about 0.1 ppb in the underlying limestone to more than
6 ppb in the clay. The absolute amount seems small, but the
anomalous spike— a 60- fold increase— was dramatic. It meant
either (1) the clay layer represented a very long period of time
during which meteoritic dust rained down slowly, yet very little
normal sediment accumulated; or (2) a very large amount of
meteoritic material had been delivered all at once to Earth by
an object on the order of 10 km (6 mi) in diameter. Neither of
these seemed likely, but of the two, the second seemed less
unlikely.
However, this deus ex machina explanation ran counter to
the deeply instilled habit of uniformitarian thinking in geol-
ogy and its Lyellian aversion to invoking catastrophic causes.
Also, the seemingly thin thread of evidence— a tiny increase
in a strange element in a thin clay layer— was not convincing
to many paleontologists who had spent their lives studying the
fossil record for clues to the Cretaceous extinction. But as sim-
ilar iridium anomalies were documented at other sites around
the world where uppermost Cretaceous rocks are exposed, the
story gained momentum. The new question became, Where
was the crater?
By the late 1980s, a trail of tektites— spheres and teardrops of
glass formed from the melting of rock in high- energy impacts—
pointed to the Caribbean region as the most likely location of
the end- Cretaceous ground zero. But it wasn’t until 1991, more
than 10 years after the original meteorite impact hypothesis
was proposed, that a crater of the right approximate age and
size was identified— a 190- km (120- mi)- wide depression largely
An atlas of time 57
buried by younger sediment off the north coast of Mexico’s
Yucatan Peninsula. It was named the Chicxulub crater after the
nearest seaside village. The following year, the publication of
argon- argon ages of in situ melt glass from drill cores taken at
the center of the crater was enough to change the minds of geol-
ogists who were still skeptical about whether this was the site of
the cataclysm. The weighted mean of the 40Ar/39Ar plateau ages
for three samples was 65.07 ± 0.10 million years— the Inter-
national Commission on Stratigraphy’s precise chronometric
definition of the end of the Cretaceous Period.17
PA R S I N G T H E P R E C A M B R I A N
In the context of Earth’s history, dinosaurs are like attention-
hogging celebrities who get a disproportionate share of media
coverage when there are so many other important stories to be
followed. While I respect all rocks, I must confess to some prej-
udices. Having grown up on the edge of the Canadian Shield—
the old core of the North American continent— I have a deeply
instilled predilection for rocks with at least a billion years be-
hind them. Like wine and cheese, rocks grow more interesting
as they age, accumulating flavor and character. For one thing,
most Precambrian rocks have survived long enough to have
been caught up in at least one episode of tectonic upheaval and
carried to depths far from their native habitats, then against
all odds, to have found their way back to the surface. Young
rocks communicate in plain prose, which makes them easy to
read, but they typically have only one thing to talk about. The
oldest rocks tend to be more allusive, even cryptic, speaking
in metamorphic metaphor. With patience and close listening,
however, they can be understood, and they generally have
more profound truths to share about endurance and resilience.
58 Ch a pter 2
Even before Claire Patterson’s decisive determination of the
age of the Earth, isotopic dates from Precambrian rocks were
revealing how greatly the Victorian fossil- based timescale had
distorted geologists’ perceptions of geologic time. Rocks of
lowermost Cambrian age were known to be about 550 million
years old, but rocks in the Canadian Shield were yielding ages
greater than 2 billion years. And a 4.5 billion- year- old Earth
meant that the quasi- mystical Precambrian, once viewed as the
irretrievable infancy of the Earth, actually includes its child-
hood, youth, and most of its adulthood— eight- ninths of its total
age. Even today, there is a lingering habit of overemphasizing
the Phanerozoic— the eon of “visible life,” from the Cambrian
to today. Most textbooks of historical geology still devote only a
perfunctory chapter or two to the Precambrian and then move
quickly on to the “real” story. Little by little, high- resolution
geochronology, and in particular a new generation of uranium-
lead analyses, is correcting this persistent temporal bias.
Just as people have no memory of their birth or first year
of life, Earth has no direct record of its formation or earliest
days. Earth’s own chronicle of its past begins with faint, cryptic
entries from between 4.4 to 4.2 billion years ago, in the form of
a few tiny crystals of the durable mineral zircon that were pre-
served as grains in an ancient sandstone in the remote Jack Hills
of western Australia. The significance of these oldest of all Earth
objects has been hotly contested since the announcement of
their discovery in a now- famous paper in Nature in 2001.18
Zircon is a geochronologist’s dream (and was the mineral
Arthur Holmes used in the very first geologic age determina-
tions). It accepts uranium but not lead into its structure at the
time it crystallizes. And because uranium has two radioactive
parent isotopes that decay to different lead daughters, there is
a built- in cross- check for whether any daughter has been lost.
An atlas of time 59
If the 206Pb/238U and 207Pb/235U ages match, the dates are
said to be concordant, and this is good evidence that no lead
loss has occurred. The precisions of concordant U- Pb zircon
dates are astonishing: the oldest Jack Hills zircon gave an age of
44
04 ± 8 million years, or an uncertainty of only 0.1%— far bet-
ter, proportionally, than 14C dates. Still, all is not lost even if lead
was lost; statistical analysis of a collection of discordant zircons
from a given rock can yield not only their crystallization age but
often the age of the metamorphic event that led to the lead loss.
In addition, zircon is a physically tough mineral, capable of
withstanding abrasion and corrosion that others cannot, and
it has a very high melting temperature, so it can come through
metamorphism without losing its “memory” of its earlier days.
As geochronologists are fond of saying, “zircons are forever”
(in contrast with diamonds, which, as high- pressure mantle-
derived minerals slowly but inexorably revert to graphite at
Earth’s surface). Old zircon crystals commonly have con-
centric zones that are almost like tree rings— the core of the
crystal records its original crystallization from a magma, and
the successive rings reflect growth during later metamor-
phic events (figure 6). The most advanced generation of mass
spectrometers— the Super High Resolution Ion Microprobe, or
SHRIMP— can find isotope ratios for individual “growth rings”
as narrow as 10 microns, less than the width of a hair. The ex-
tremely old ages obtained for the Jack Hills zircons came from
the interiors of crystals with complex overgrowths. Just as the
rings of one old tree may contain a climate record for a whole
region, a single ancient zoned zircon crystal can chronicle the
tectonic history of a continent.
The astounding age of the Jack Hills zircon grains is even
more surprising in light of the fact that zircon forms almost
exclusively during the crystallization of granites and similar
60 Ch a pter 2
F I G U R E 6 . Zircon crystals with growth bands
igneous rocks, which are the foundations of the continents.
Granites represent “evolved” magmas, meaning they are diffi-
cult to form in just one stage of melting from the Earth’s mantle
(the ultimate source of all crustal rocks). Today, granitic rocks
come mainly from the forges of subduction- zone volcanoes
like Mount Rainier and are derived by partial melting of pre-
existing crust, usually in the presence of water (more about
this in chapter 3). So if the Jack Hills zircons were forged in
this modern way, their existence suggests the dizzying pros-