Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Earth science

In the beginning. . .

Grains of the mineral zircon have survived from way back in Earth's history. Analysis of these grains provides information on the state of our planet as long as 4.4 billion years ago.

When did the Earth's continents and oceans first form? On pages 175 and 178 of this issue, Wilde et al.1 and Mojzsis et al.2 address the question with reports of uranium–lead (U–Pb) ages and oxygen isotopic compositions of extremely old zircon grains. Their results provide evidence that continents and liquid water were surface features of the earliest Earth. Part of one grain appears to have formed 4.4 billion years ago: this is the oldest terrestrial solid yet identified. Although it is unclear how general a picture a few tiny zircon grains can provide, the results represent a significant advance in reconstructing Earth's Dark Ages (Fig. 1).

Figure 1: Earth's earliest history, in millions of years, starting from the origin of the Solar System; numbers on the right show absolute ages in billions of years.

Dynamic modelling, U–Pb dating of meteorites, and the use of short-lived nuclides all provide evidence for how the Earth formed over the first 100 million years of the Solar System's existence. The Dark Ages of before 4 billion years ago, strictly termed the Hadean, is a period from which no rocks seem to have survived. All we have from this time are a few zircon grains. Wilde et al.1 and Mojzsis et al.2 are the latest to show that zircons nonetheless represent a wonderful archive of information about the early Earth.

The U–Pb dating technique depends on the two main isotopes of uranium, 235U and 238U, which undergo spontaneous chain decay until they form stable 207Pb and 206Pb, respectively. The probabilities of decay of 235U and 238U are extremely small, so their half-lives are long — 700 million and over 4 billion years, respectively. The unique feature that you can simply use the ratio of 207Pb to 206Pb to determine an age, even if the U/Pb ratio has been recently perturbed, makes U–Pb dating the most powerful way of determining the absolute age of material from the early Solar System3. The technique has been used to date the origin of the Solar System at 4.566±0.002 billion years ago4.

However, the method is inadequate for defining the rates of formation of planets and their component parts, such as the core and atmosphere. In the 1950s came the first convincing evidence of a now-extinct isotope, 129I, that was present at the birth of the Solar System5. This discovery sparked the beginning of the use of extinct nuclides, such as 53Mn, 107Pd, 129I and 182Hf, which have half-lives that are two orders of magnitude shorter than that of 235U. Although the parent isotope is extinct, its former abundance can be estimated from a correlation between the isotopic abundance of the daughter and the parent/daughter element ratio in an object of independently known age. One can then use the isotopic abundance of the daughter in other objects to determine their precise age relative to that of the original object.

From these studies, we know that the accretion (aggregation) of small bodies in the solar nebula occurred within about ten million years of the birth of the Solar System6. Similarly, we can show that Earth's growth was protracted, and dominated by impacts and planetary collisions7. By 4.51 billion to 4.45 billion years ago the Earth had reached its present mass, with a metal core and primitive atmosphere7,8,9. These timescales are consistent with computer models of planetary accretion10. In its early stages Earth probably had a magma ocean, sustained by heat from impacts and the blanketing effects of a dense atmosphere, but much of that atmosphere would have been lost with the dispersion of the solar nebula and during planetary collisions8,9,11. The Earth would then have cooled quickly, the outer portions would have solidified, and the first primitive crust would have developed.

There was, however, no direct evidence of what such a crust might have looked like. Unlike on the Moon and Mars, no rock older than 4 billion years old seems to have been preserved on Earth. An intense bombardment of the Moon occurred up until roughly 3.9 billion years ago12. So Earth's earlier crust may have been destroyed by impacts at around the same time, or it could be that a hotter Earth had an inherently unstable surface. Some argued that the earliest crust was like the Lunar Highlands — made from a welded mush of crystals that had previously floated on the magma ocean. Others suggested that it was made of denser rocks like those of the Earth's present ocean floor. But firm evidence was scarce.

In 1983, with the exciting report13 of zircon grains over 4 billion years old, we at last had some evidence. But it upheld neither of the above theories. Instead the data constituted evidence that the early crust was like the continental crust today. The original host rocks for the zircons had been destroyed by erosion. But zircon is especially hard and enduring, and the grains survived to tell their tale after becoming incorporated in sedimentary rocks now exposed in Australia. The U–Pb ages of most of the grains turned out to be between 4.1 and 4.2 billion years, with a few grains being slightly older13,14.

But why does the presence of zircons suggest the existence of continental crust? Nearly all zircons grow from granite magmas, which are nothing like those forming the ocean floor or Lunar Highlands. Such magmas usually form at high temperatures and depths of more than 20 km, mainly by melting of pre-existing continental crust above the subduction zones where ocean crust is being dragged down into Earth's mantle. But the granites are buoyant, and rise up to eventually create mountain regions such as the Andes. So the zircons described in the new papers1,2 are consistent with the existence of continental crust as far back as 4.4 billion years ago.

We could do with more data to confirm this age, and inferring the existence of entire continents from zircon grains requires quite a leap of the imagination. Nonetheless, large ocean islands such as Iceland contain small amounts of granitic magma, and one can conceive of that magma acting as a nucleus for other material to accrete and create a proto-continent, and then a continent.

Both Wilde et al.1 and Mojzsis et al.2 also include oxygen-isotope evidence for the presence of liquid water on the early Earth's surface. The oxygen isotopic composition of zircons reflects that of the magma from which they crystallized, which in turn reflects the composition of the rocks that were melted to form the magma. Heavy oxygen (with a high proportion of the isotope 18O) is produced by low-temperature interactions between a rock and liquid water. The zircons are rich in heavy oxygen, which implies that the rocks that were melted to form the magma included components that had been at the surface in the presence of liquid water. Earlier on, when Earth was hotter, they might also have formed by melting of once-wet ocean-floor rocks during a process akin to modern subduction. Either way, surface rocks affected by low-temperature fluids were probably being transported to considerable depths and melted. However, there will be suspicions that the 18O-rich composition was acquired by diffusion after the zircons had been formed, so further information on that possibility is required.

Zircon grains have further potential in investigations of the early Earth. For example, Wilde et al.1 also analysed trace elements and tiny inclusions to reconstruct the composition of the parent magma, and Hf isotopes can be used to investigate the even earlier history of the Earth15. And zircons might even contain some isotopic or chemical tracer for early biological activity. For studies such as these we need more grains, and lots of them.


  1. 1

    Wilde, S. A., Valley, J. W., Peck, W. H. & Graham, C. M. Nature 409, 175–178 ( 2001).

    CAS  Article  ADS  Google Scholar 

  2. 2

    Mojzsis, S. J., Harrison, T. M. & Pidgeon, R. T. Nature 409, 178– 181 (2001).

    CAS  Article  ADS  Google Scholar 

  3. 3

    Patterson, C. C. Geochim. Cosmochim. Acta 10, 230–237 (1956).

    CAS  Article  ADS  Google Scholar 

  4. 4

    Allègre, C. J., Manhès, G. & Göpel, C. Geochim. Cosmochim. Acta 59, 1445–1456 (1995).

    Article  ADS  Google Scholar 

  5. 5

    Reynolds, J. H. Phys. Rev. Lett. 4, 8 (1960).

    CAS  Article  ADS  Google Scholar 

  6. 6

    Lugmair, G. W. & Shukolyukov, A. Geochim. Cosmochim. Acta 62, 2863–2886 (1998).

    CAS  Article  ADS  Google Scholar 

  7. 7

    Halliday, A. N. Earth Planet. Sci. Lett. 176, 17–30 (2000).

    CAS  Article  ADS  Google Scholar 

  8. 8

    Sasaki, S. & Nakazawa, K. J. Geophys. Res. 91 , B9231–B9238 (1986).

    Article  ADS  Google Scholar 

  9. 9

    Porcelli, D., Cassen, P., Woolum, D. & Wasserburg, G. J. LPI Contrib. 597 35–36 (Lunar & Planetary Inst., Houston, 1998).

    Google Scholar 

  10. 10

    Wetherill, G. W. Annu. Rev. Astron. Astrophys. 18, 77– 113 (1980).

    CAS  Article  ADS  Google Scholar 

  11. 11

    Benz, W. & Cameron, A. G. W. in Origin of the Earth (eds Newsom, H. E. & Jones, J. H.) 61–67 (Oxford Univ. Press, 1990).

    Google Scholar 

  12. 12

    Wasserburg, G. J., Papanastassiou, D. A., Tera, F. & Huneke, J. C. Phil. Trans. R. Soc. Lond. A 285, 7– 22 (1977).

    CAS  Article  ADS  Google Scholar 

  13. 13

    Froude, D. O. et al. Nature 304, 616–618 (1983).

    CAS  Article  ADS  Google Scholar 

  14. 14

    Compston, W. & Pidgeon, R. T. Nature 321, 766–769 (1986).

    CAS  Article  ADS  Google Scholar 

  15. 15

    Amelin, Y. et al. Nature 399, 252–255 (1999).

    CAS  Article  ADS  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Alex N. Halliday.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Halliday, A. In the beginning. . .. Nature 409, 144–145 (2001).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing