pect of a still- earlier crust that had formed, cooled, and then
remelted within the first 150 million years of the planet’s ori-
gin. Equally surprising, the ratios of different oxygen isotopes
in the old zircons suggest that the magma from which they
crystallized had interacted with relatively cool surface water.
An atlas of time 61
Abandoning customary scientific restraint in their conclud-
ing paragraph, the authors of the 2001 Nature paper boldly
suggested— on the basis of a few crystals smaller than fleas—
that not only continents and oceans existed on Earth 4.4 bil-
lion years ago but that if surface water was around, perhaps
there was even life.
T H I N K I N G L I K E A P L A N E T
The Jack Hills zircon paper, one of the most cited in all the
geologic literature, was a virtuosic culmination of a century
of isotope geochemistry and required the most advanced ana-
lytical methods available. Yet, in its audacious inductive spec-
ulations and strong predilection for uniformitarianism, it is
remarkably similar to the very first work of modern geology:
Hutton’s Theory of the Earth. Whether, in fact, the early Earth
should be viewed through strictly uniformitarian spectacles is
currently a matter of lively debate among geologists. There are
compelling reasons to suggest that Earth’s habits were different
in its first 2 billion years.
But the story of how the still- unfinished Atlas of Deep Time
has evolved, from Siccar Point to Chicxulub to Jack Hills, makes
it clear that mapping time has been a very human endeavor that
requires just this kind of give- and- take. It has involved a great
variety of minds— visionary thinkers like Hutton and Lyell not
too obsessed with details; attentive fossil- hunters like William
Smith who are; polymaths like Darwin and Holmes who see
connections across disciplines; fastidious instrumentalists like
Nier and Patterson; bureaucracies like the International Com-
mission on Stratigraphy; and legions of hardy, anonymous field
mappers (including a few jocks) who understand both chronos
and kairos, and how to turn rocks into verbs.
C H A P T E R 3
T H E PAC E O F
T H E E A R T H
How many years can a mountain exist
before it is washed to the sea?
— B O B DY L A N , 1 9 6 3
E P H E M E R A L G E O G R A P H I E S
One of my earliest school memories is of watching a film about
the emergence of Surtsey, a volcanic island off the coast of Ice-
land that began to rise from the Atlantic late in 1963. The black-
and- white footage showed explosive spires of steam and ash
creating a blank new world of coal- dark cinder not yet on any
map. A ship captain had been the first to notice the eruption
and had initially thought it was another vessel on fire. To my
impressionable young mind, the idea of new land forming was
thrilling; it suggested a secret life- force inside the impassive,
stony- faced Earth. Between 1963 and 1967, Surtsey built itself
up from a submarine ridge 130 m (425 ft) below sea level to a
small cone more than 170 m (570 ft) above it. At its maximum
extent, Surtsey had an area of about a square mile. But after the
eruptions ceased, its destruction by erosion, settling, and sub-
sidence was nearly as rapid. Today, it has been reduced to about
half its 1967 size and is expected to disappear entirely by 2100
(or sooner, depending on the rate of sea level rise). To my still
impressionable middle- aged mind, it is somehow unsettling to
The pace of the earth 63
have seen the arc of Surtsey’s life— the birth, youth, brief prime,
and inexorable demise of a landmass.
To Hutton, Lyell, and Darwin, most geologic processes
seemed imperceptibly slow, and for decades, geologists
drummed this idea into the public consciousness. But today,
thanks to high- precision geochronology, direct satellite obser-
vation of Earth processes from space, and a century of moni-
toring the planet’s vital signs— temperature, precipitation, river
flow, glacier behavior, groundwater reserves, sea level, seismic
activity— many of the geologic processes that once seemed be-
yond the reach of direct human observation can now be clocked
in real time. And we are finding that the pace of the planet is
neither as slow nor as constant as was previously thought.
B A S A LT O F T H E E A R T H
Hutton’s original epiphany that the age of the earth was effec-
tively infinite compared with human lifespans arose from his
recognition that the unconformity at Siccar Point represented
the time needed for a mountain belt to form and be beveled
again to a flat plain. So how long, exactly, would this take? The
forces behind mountain building were not known until about
175 years after Hutton’s death— in fact, around the time of
Surtsey’s birth in the 1960s, when plate tectonic theory finally
explained how the solid Earth works. Today we realize that the
tempo of mountain growth is ultimately set by the formation
and destruction of ocean basins.
Unlike the continents, which are a messy amalgam of many
different rock types of a wide range of ages and individual histo-
ries, the ocean crust is simple and homogeneous. It’s all basalt—
the black volcanic rock of Surtsey— and it’s all produced in the
same way: by partial melting of the Earth’s mantle beneath
64 Ch a pter 3
submarine volcanic rifts, marked by high- standing midocean
ridges. Counter to fanciful depictions in fiction and film, the
mantle (which constitutes more than 80% of Earth’s volume)
is not a vat of seething magma but solid rock— though it flows
over geologic timescales. Every few hundred million years, the
mantle overturns itself in the manner of a gargantuan lava lamp,
through the process of thermal convection: hotter, buoyant
rock from depth rises while cooler, denser rock sinks. Mantle
convection is Earth’s main heat- loss mechanism (contrary to
Lord Kelvin’s erroneous assumption that the mantle was static
and Earth had cooled over its lifespan through conduction).
Arthur Holmes was among the first to suggest, in the 1930s,
that the mantle convects; today, high- pressure experiments
simulating the behavior of minerals at mantle depths show that
convection of rock in Earth’s interior is inevitable.
Midocean ridges are thought to coincide with areas of con-
vective upwelling, where the Earth’s crust is forced to stretch
and thin above the rising plume of hot rock. Paradoxically,
however, no melt forms until the ascending rock has lost
much of it heat. So what makes still- solid mantle rock melt as
it nears the surface? The mechanism is counterintuitive— not
an input of heat but a decrease in pressure. Unlike water, a
completely abnormal compound from which most of us derive
our under standing of phase changes, rock behaves the way a
proper substance should: it expands on melting and contracts
on freezing. This means that if a rock is close to its melting
temperature at some depth in the Earth and is depressurized
(e.g., by rising closer to the surface), the lower- density phase—
melt— becomes favored, and magma is formed. This phenome-
non is called decompression melting and can happen even if the
rock is actually cooling, as long the pressure is decreasing faster
than the temperature is. (Decompression melting is especially
The pace of the earth 65
hard for skiers and skaters to understand, since the opposite
be havior by water— melting under higher pressure— is the very
basis for winter sports that involve slippery surfaces).
On Earth today, after 4.5 billion years of cooling through
mantle convection, upwelling mantle rock does not carry
enough thermal clout to undergo wholesale melting. Instead,
magmas at ocean ridges represent the components in mantle
rock that melt at the lowest temperature. This partial, or frac-
tional, melting is what generates basalt, which has a different
composition— higher in silica, aluminum, and calcium, and
lower in magnesium— than its parent, the mantle.
As each new pulse of basaltic melt rises and fills the central
axis of an oceanic rift, the previous batches, now frozen into
rock again, are displaced symmetrically outward in the pro-
cess called seafloor spreading (see figure 7). The most recently erupted basalt is warmer and less dense than the slightly older
rock it has pushed aside, and each generation cools progres-
sively as it moves away from its birthplace at the rift. This is the
reason that the midocean ridges stand high, like a soufflé fresh
from the oven. In fact, one of the clues that led to the epiphany
of plate tectonic theory in the early 1960s, when deep- seafloor
maps first became available, was that the cross- sectional form
of the ocean ridges is essentially a pair of mirror- image cooling
curves— the shape of two skis placed on the floor tip to tip.
A L L O V E R T H E M A P
Let us pause to contemplate how incredible it is that most of the
Earth’s surface— the deep- ocean floor— had not been mapped
until the middle of the twentieth century. Even today, the
topography of much of the seafloor is known to a resolution of
only about 3 miles; bathymetric charts of the ocean are about
66 Ch a pter 3
F I G U R E 7. Midocean ridge, seafloor spreading, and magnetic reversals
100 times “blurrier” than current maps of the surface of Venus
and Mars.1 Still more incredible is the fact that one person al-
most single- handedly created the first maps of two- thirds of
the planet yet is unknown to the average citizen of Earth (while
Amerigo Vespucci, whose cartographic credentials are suspect,
has two continents named for him). The unsung mapmaker was
The pace of the earth 67
Marie Tharp, who earned a master’s degree in geology from
the University of Michigan, worked briefly for an oil company,
and then in 1948 became a drafter for a new oceanographic
project led by Maurice Ewing at Columbia University.2 For
years, Ewing’s all- male team of graduate students collected
sonar soundings of the ocean floor while Tharp laboriously
transformed the linear strings of depth readings into three-
dimensional topography.
Tharp’s exquisite shaded relief maps, painstakingly drawn in
pen and ink, revealed that the seabed— previously thought flat
and featureless— had a rugged, globe- encircling range of ridges
and terrifyingly deep trenches. By 1953, she had noticed that
the high ridges all had central down- dropped valleys and spec-
ulated that this might be evidence for crustal stretching. She
shared her idea with another member of Ewing’s group, Bruce
Heezen, who infamously dismissed it as “girl talk.” But Heezen
and Tharp became close collaborators at Columbia, producing
a series of seafloor maps that revolutionized geologists’ view of
the Earth. In 1963, when two British geologists first articulated
the concept of seafloor spreading in a paper in Nature 3 (and
Surtsey was demonstrating the process), Heezen— and much
later, the rest of the geologic community— acknowledged that
Tharp had been right.
The authors of the 1963 paper, Fredrick Vine and Drum-
mond Matthews, proposed seafloor spreading on the basis of a
perceptive geometric argument rather than firsthand geologic
observation (the ridges would not be directly seen or sampled
for another decade). Vine and Matthews had access not only
to Tharp’s maps but also to data from the U.S. and Royal Navies
on the magnetic signatures of rocks at the bottom of the ocean.
They noted that both the ridge topography and the magnetic
intensity readings had mirror symmetry moving outward from
68 Ch a pter 3
the ridge crest— that is, bands of similarly magnetized rocks ran
in parallel stripes on either side of the ridge (see figure 7). The ridge heights fell off with distance in just the way one would
expect for deflating soufflés or cooling and contracting rocks.
The symmetrical pattern of magnetic stripes suggested that
successive generations of ocean crust had formed at the ridge,
cooled enough for their iron- bearing minerals to align with
the ambient magnetic field, and then been cleaved in half and
displaced outward in a great conveyor system. Meanwhile, the
polarity of Earth’s magnetic field had repeatedly reversed itself,
the north and south geomagnetic poles switching places on an
erratic schedule (a second revolutionary inference in a paper
that is barely three pages long).
By the early 1970s, age determinations for seafloor samples
from deep- ocean drilling, as well as correlation of the marine
magnetic record with magnetic reversals in well- dated volca-
nic sequences on land, had created a new way to demarcate
geologic time, and the geomagnetic timescale was grafted
onto the biostratigraphic (fossil- based) and geochronologic
(radioisotope- calibrated) timescales. Today, with the date of
each magnetic field reversal well constrained, it is possible to
determine the age of a rock anywhere on the seafloor without
even getting a physical sample— simply by counting how many
magnetic stripes it is away from the ridge.
On a map showing the ages of seafloor in the world’s oceans,
the most striking pattern is that the swaths of rock of any given
age are much wider in the Pacific Ocean than in the Atlantic.
Since the start of the Cenozoic Era 65 million years ago (i.e.,
since the demise of the dinosaurs), seafloor spreading rates in
the Atlantic have averaged about 1 cm (ca. 1/2 inch) per year,
which is on the order of the rate at which one’s fingernails
grow. It’s fast enough that at Thingvellir in Iceland, one of the
The pace of the earth 69
few places where an ocean ridge stands above sea level— and
the site the Vi
kings chose in AD 930 for their annual parlia-
ment meeting, the Althing— the visitor center was built to be
as wide as the amount by which the crust has stretched since
Viking times.
On the other hand, the rate of spreading in the Atlantic is
slow enough that a species of green sea turtle ( Chelonia mydas)
from Brazil that has made an annual swim to a high spot on the
Mid- Atlantic Ridge to breed and nest since the time of the dino-
saurs hasn’t seemed to notice that the ridge is now nearly 1100
km (700 mi) farther distant. Luckily the turtle’s natal beaches
weren’t in the Pacific, where spreading rates are almost an order
of magnitude faster, at close to 10 cm (4 in.) per year (a little
slower than the “velocity” of hair growth). If these rates simply
reflected the pace of mantle convection, why would that pace
be more vigorous beneath one ocean than another?
P L AT E S P U L L T H E I R W E I G H T
Marie Tharp’s marvelous maps hold clues to the disparity in
rates of plate motion in the two oceans. In particular, they show
profound differences between the edges of the Pacific and At-
lantic basins: the margins of the Atlantic Ocean are mainly shal-
low continental shelves, like the area off the coast of the eastern
United States, where water depths are less than about 200 m
(660 ft), and submerged crust gives way gradually to emergent
land. The margins of the Pacific Ocean, in contrast, are delin-
eated by vertiginous chasms, like the one off the west coast
of South America, whose deepest points lie more than 8000
m (24,000 ft) below sea level. These trenches mark the sites
of subduction, where old, cold, ocean crust—with the same
instinct as the Brazilian turtles—returns to its place of origin.
70 Ch a pter 3
F I G U R E 8 . Subduction zones and volcanic arcs
When sea floor basalt is around 150 million years old, and
hundreds of miles from its natal ridge, it has become about as
dense as the underlying mantle and sinks back into Earth’s inte-
rior at a slant, pulling the rest of the plate behind it, like a blan-
ket sliding off a bed (figure 8). This “slab pull” force is almost certainly what sets the tempo for the rapid seafloor spreading in the Pacific— its rifts are simply keeping up with rates
Timefulness Page 8