Researchers collect sediments from a rocky stream with a helicopter and steep rock hills in the background
Researchers collect stream sediments in Greenland. Zircons from these sediments have provided tantalizing clues to how today’s continents came to be. Credit: Agnete Steenfelt

The jigsaw fit of Earth’s continents, which long intrigued map readers and inspired many theories, was explained about 60 years ago when the foundational processes of plate tectonics came to light. Topographic and magnetic maps of the ocean floor revealed that the crust—the thin, rigid top layer of the solid Earth—is split into plates. These plates were found to shift gradually around the surface atop a ductile upper mantle layer called the asthenosphere. Where dense oceanic crust abuts thicker, buoyant continents, the denser crust plunges back into the mantle beneath. Above these subduction zones, upwelling mantle melt generates volcanoes, spewing lava and creating new continental crust.

From these revelations, geologists had a plausible theory for how the continents formed and perhaps how Earth’s earliest continents grew—above subduction zones. Unfortunately, the process is not that simple, and plate tectonics have not always functioned this way. Subsequent research since the advent of plate tectonic theory has shown that subduction and associated mantle melting provide only a partial explanation for the formation and growth of today’s continents. To better understand the production and recycling of crust, some scientists, including our team, have shifted from studying the massive moving plates to detailing the makeup of tiny mineral crystals that have stood the test of time.

Starting in the 1970s, geologists from the Greenland Geological Survey collected stream sediments from all over Greenland, sieving them to sand size and chemically analyzing them to map the continent-scale geochemistry and contribute to finding mineral occurrences. Unbeknownst to them at the time, tiny grains of the mineral zircon contained in the samples held clues about the evolution of Earth’s early crust. After decades in storage in a warehouse in Denmark, the zircon grains in those carefully archived bottles of sand—and the technology to analyze them—were ready to reveal their secrets.

Magnified view showing the internal structure of zircon crystals
This cathodoluminescence image shows the internal structure of magnified zircons analyzed by laser ablation. Credit: Chris Kirkland

Zircon occurs in many rock types in continental crust, and importantly, it is geologically durable. These tiny mineral time capsules preserve records of the distant past—as far back as 4.4 billion years—which are otherwise almost entirely erased. More than just recording the time at which a crystal grew, zircon chemistry records information about the magma from which it grew, including whether the magma originated from a melted piece of older crust, from the mantle, or from some combination of these sources. Through the isotopic signatures in a zircon grain, we can track its progression, from the movement of the magma up from the mantle, to its crystallization, to the grain’s uplift to the surface and its later erosion and redeposition.

The Past Is Not Always Prologue

New continental crust is formed above subduction zones, but it is also destroyed at subduction zones [e.g., Scholl and von Heune, 2007]. Formation and destruction occur at approximately equal rates in a planetary-scale yin and yang [Stern and Scholl, 2010; Hawkesworth et al., 2019]. So crust formation above subduction zones cannot satisfactorily account for growth of the continents.

If subduction didn’t generate the volume of continental crust we see today, what did?

What’s more, plate tectonic movements like we see on Earth today did not operate the same way during Earth’s early history. Although there are indications that subduction may have occurred in Earth’s early history (at least locally), many geochemical, isotopic, petrological, and thermal modeling studies of crust formation processes suggest that plate tectonics started gradually and didn’t begin operating as it does today until about 3 billion years ago, after more than a quarter of Earth’s history had already passed [e.g., McClennan and Taylor, 1983; Dhuime et al., 2015; Hartnady and Kirkland, 2019]. Because the mantle was much hotter at that time, more of it melted than it does now, producing large amounts of oceanic crust that was both too thick and too viscous to subduct.

Nonetheless, although subduction was apparently not possible on a global scale before about 3 billion years ago, geochemical and isotopic evidence shows that a large volume of continental crust had already formed by that time [e.g., Hawkesworth et al., 2019; Condie, 2014; Taylor and McClennan 1995].

If subduction didn’t generate the volume of continental crust we see today, what did?

How Did Earth’s Early Crust Form?

The nature of early Earth dynamics and how and when the earliest continental crust formed have remained topics of intense debate, largely because so little remains of Earth’s ancient crust for direct study. Various mechanisms have been proposed.

Perhaps plumes of hot material rising from the mantle melted the oceanic crustal rock above [Smithies et al., 2005]. If dense portions of this melted rock “dripped” back into the mantle, they could have stirred convection cells in the upper mantle. These drips might have also added water to the mantle, lowering its melting point and producing new melts that ascended into the crust [Johnson et al., 2014].

Or maybe meteorite impacts punched through the existing crust into the mantle, generating new melts that, again, ascended toward the surface and added new crust [Hansen, 2015]. Another possibility is that enough heat built up at the base of the thick oceanic crust on early Earth that parts of the crust remelted, with the less dense, buoyant melt portions then rising and forming pockets of continental crust [Smithies et al., 2003].

By whichever processes Earth’s first continental crust formed, how did the large volume of continental crust we have now build up? Our research helps resolve this question [Kirkland et al., 2021].

Answers Hidden in Greenland Zircons

We followed the histories of zircon crystals through the eons by probing the isotopes preserved in grains from the archived stream sediment samples from an area of west Greenland. These isotopes were once dissolved within molten mantle before being injected into the crust by rising magmas that crystallized zircons and lifted them up to the surface. Eventually, wind and rain erosion released the tiny crystals from their rock hosts, and rivulets of water tumbled them down to quiet corners in sandy stream bends. There they rested until geologists gathered the sand, billions of years after the zircons formed inside Earth.

In the laboratory, we extracted thousands of zircon grains from the sand samples. These grains—mounted inside epoxy resin and polished—were then imaged with a scanning electron microscope, revealing pictures of how each zircon grew, layer upon layer, so long ago.

Scene in a laboratory with computers and monitors in front of a laser ablation mass spectrometer
Researchers used the laser ablation mass spectrometer at Curtin University to study isotopic ratios in zircon crystals. Credit: Chris Kirkland

In a mass spectrometer, the zircons were blasted with a laser beam, and a powerful magnetic field separated the resulting vapor into isotopes of different masses. We determined when each crystal formed using the measured amounts of radioactive parent uranium and daughter lead isotopes. We also compared the hafnium isotopic signature in each zircon with the signatures we would expect in the crust and in the mantle on the basis of the geochemical and isotopic fractionation of Earth through time. Using these methods, we determined the origins of the magma from which the crystals grew and thus built a history of the planet from grains of sand.

The oldest continental crust might have survived to serve as scaffolding for successive additions of younger continental crust.

Our analysis revealed that the zircon crystals varied widely in age, from 1.8 billion to 3.9 billion years old—a much broader range than what’s typically observed in Earth’s ancient crust. Because of both this broad age range and the high geographic density of the samples in our data set, patterns emerged in the data.

In particular, some zircons of all ages had hafnium isotope signatures that showed that these grains originated from rocks that formed as a result of the melting of a common 4-billion-year-old parent continental crust. This common source implied that early continental crust did not form anew and discretely on repeated occasions. Instead, the oldest continental crust might have survived to serve as scaffolding for successive additions of younger continental crust.

In addition to revealing this subtle, but ubiquitous, signature of Earth’s ancient crust in the Greenland samples, our data also showed something very significant about the evolution of Earth’s continental crust around 3 billion years ago. The hafnium signature of most of the zircons from that time that we analyzed showed a distinct isotopic signal linked to the input of mantle material into the magma from which these crystals grew. This strong mantle signal in the hafnium signature showed us that massive amounts of new continental crust formed in multiple episodes around this time by a process in which mantle magmas were injected into and melted older continental crust.

Geologists work atop a coastal rock outcrop
Geologists work atop a rock outcrop in the Maniitsoq region of western Greenland. Credit: Julie Hollis

The idea that ancient crust formed the scaffolding for later growth of continents was intriguing, but was it true? And was this massive crust-forming event related to some geological process restricted to what is now Greenland, or did this event have wider significance in Earth’s evolution?

A Global Crust Formation Event

To test our hypotheses, we looked at data sets of isotopes in zircons from other parts of the world where ancient continental crust is preserved. As with our Greenland data, these large data sets all showed evidence of repeated injection of mantle melts into much more ancient crust. Ancient crust seemed to be a prerequisite for growing new crust.

This “hot flash” in the deep Earth may have driven a planetary continent growth spurt.

Moreover, the data again showed that these large volumes of mantle melts were injected into older crust everywhere at about the same time, between 3.2 billion and 3.0 billion years ago, timing that coincides with the estimated peak in Earth’s mantle temperatures. This “hot flash” in the deep Earth may have enabled huge volumes of melt to rise from the mantle and be injected into existing older crust, driving a planetary continent growth spurt.

The picture that emerges from our work is one in which buoyant pieces of the oldest continental crust melted during the accrual and trapping of new mantle melts in a massive crust-forming event about 3 billion years ago. This global event effectively, and rapidly, built the continents. With the onset of the widespread subduction that we see today, these continents have since been destroyed, remade, and shifted around the surface like so many jigsaw pieces in perpetuity through the eons.


Condie, K. (2014), Growth of continental crust: A balance between preservation and recycling, Mineral. Mag., 78(3), 623–637,

Dhuime, B., A. Wuestefeld, and C. J. Hawkesworth (2015), Emergence of modern continental crust about 3 billion years ago, Nat. Geo­sci., 8, 552–555,

Hansen, V. L. (2015), Impact origin of Archean cratons, Lithosphere, 7, 563–578,

Hartnady, M. I. H., and C. L. Kirkland (2019), A gradual transition to plate tectonics on Earth between 3.2 and 2.7 billion years ago, Terra Nova, 31, 129–134,

Hawkesworth, C. J., B. Dhuime, and P. A. Cawood (2019), Rates of generation and growth of the continental crust, Geosci. Front., 10(1), 165–173,

Johnson, T. E., et al. (2014), Delamination and recycling of Archaean crust caused by gravitational instabilities, Nat. Geosci., 7, 47–52,

Kirkland, C. L., et al. (2021), Widespread reworking of Hadean-to-Eoarchean continents during Earth’s thermal peak, Nat. Commun., 12, 331,

McLennan, S. M., and S. R. Taylor (1983), Continental freeboard, sedimentation rates and growth of continental crust, Nature, 306, 169–172,

Scholl, D. W., and R. von Huene (2007), Crustal recycling at modern subduction zones applied to the past—Issues of growth and preservation of continental basement crust, mantle geochemistry, and supercontinent reconstruction, in 4-D Framework of Continental Crust, edited by R. D. Hatcher et al., Mem. Geol. Soc. Am., 200, 9–32,

Smithies, R. H., D. C. Champion, and K. F. Cassidy (2003), Formation of Earth’s early Archaean continental crust, Precambrian Res., 127(1–3), 89–101,

Smithies, R. H., M. J. Van Kranendonk, and D. C. Champion (2005), It started with a plume — Early Archaean basaltic proto-continental crust, Earth Planet. Sci. Lett., 238, 284–297,

Stern, R. J., and D. W. Scholl (2010), Yin and yang of continental crust creation and destruction by plate tectonic processes, Int. Geol. Rev., 52(1), 1–31,

Taylor, S. R., and S. M. McLennan (1995), The geochemical evolution of the continental crust, Rev. Geophys., 33(2), 241–265,

Author Information

Julie Hollis (, Curtin University, Perth, Western Australia; also at Government of Greenland, Nuuk; Chris Kirkland, Michael Hartnady, and Milo Barham, Curtin University, Perth, Western Australia; and Agnete Steenfelt, Geological Survey of Denmark and Greenland, Copenhagen


Hollis, J., C. Kirkland, M. Hartnady, M. Barham, and A. Steenfelt (2021), Earth’s continents share an ancient crustal ancestor, Eos, 102, Published on 23 August 2021.

Text © 2021. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.