Wrestling monsters in deep space

Observing distant galaxies is always problematic. But when it comes to the biggest star-forming galaxies, far across the Universe, only indirect approaches can give astronomers any handle at all.

One of the big questions in observational cosmology is when and where the stars in galaxies formed. Giant optical telescopes can now pick out galaxies at huge distances across the Universe, but curiously they miss the galaxies in which the greatest number of stars are forming. Much of the star formation in the Universe occurs in bursts, and the processes that fire up these monster galaxies also produce huge amounts of dust that blot out starlight at most wavelengths, including optical ones. The energy that is absorbed by the dust is re-radiated at much longer wavelengths (in the submillimetre range), and observations at these wavelengths have only recently revealed the presence of these hidden systems.

The fact that so little optical light escapes from the giant star-formers also makes it very difficult to determine how far away they are — distance is usually derived from the spectra of optical emission from galaxies. Two new approaches to finding the distances, or redshifts, of monster galaxies are now reported: by Chapman et al.1 on page 695 of this issue, and by Wiklind2 in the Astrophysical Journal. Wiklind uses the spectral shape of the galaxies at long wavelengths to estimate their redshifts. Chapman et al. use radio data to locate the galaxies accurately and then obtain the best possible optical spectra.

In 1996, the COBE satellite measured the energy density of the Universe at submillimetre wavelengths and made the remarkable discovery that, over the lifetime of the Universe, galaxies have radiated as much energy at submillimetre as at optical wavelengths3. Clearly there had to be an as-yet undetected population of galaxies that has produced this energy — which is remarkable because such a population would have to have formed as many stars as all the galaxies seen in optical observations. But this population does exist: the Submillimetre Common-User Bolometer Array, or SCUBA, on the 15-m James Clerk Maxwell Telescope on Mauna Kea, Hawaii4, has found a huge number of luminous galaxies5,6 radiating at submillimetre wavelengths — enough to produce nearly all of the submillimetre light seen by COBE7.

Simply imaging the galaxies, however, is not sufficient. To measure how they evolved with time, and to map the history of star formation, we also need to know the distances to these galaxies. This has proved to be a hard problem. Identifying the optical counterparts to the submillimetre sources is difficult because observations at long wavelengths are limited in their positional accuracy, even when very large telescopes are used. Ultimately, this problem will be resolved with a new generation of submillimetre-telescope arrays, the first of which is now coming online on Mauna Kea8. But for the moment it is still difficult to match up the images of the same object seen at optical and at submillimetre wavelengths.

The first attempts to measure redshifts for the submillimetre sources involved identifying every one of the handful or so optical galaxies near the position of each detection. The method was tedious, but distances could be measured for about a quarter of the submillimetre sources and were typically about two-thirds of the way across the Universe9. A more promising alternative, however, is to locate the submillimetre counterparts through radio observations. This works because there is a relatively tight empirical correlation in local star-forming galaxies between radio emission and thermal emission from dust10. The locations of about 60% of the bright submillimetre sources can be pinpointed in this way11.

Chapman et al.1 have drawn on this result and used radio observations to locate the optical counterparts to the submillimetre sources. They then obtained optical spectra over very long exposure times to identify emission lines in the small amount of optical light that emerges from the galaxies. Chapman et al. have measured redshifts for ten submillimetre sources, which is a substantial increase in the number of such sources with known redshifts. Even so, this approach unfortunately only samples sources from the same bright submillimetre population that had previously been identified. Although these redshifts are obtained with much less effort, distances to most of the sources still cannot be measured.

In the not so distant future, it should be possible to measure redshifts directly using the Atacama Large Millimeter Array — an ensemble of 64 dishes, each 12 m in diameter, to be built in Chile. Until then, the distances to most submillimetre sources can only be measured using crude redshift estimators. The most widely used estimator is based on the ratio of the amount of emission, or flux, at submillimetre wavelengths to the flux at radio wavelengths12. Unfortunately, this method fails when the submillimetre sources become too faint — too distant — to detect in the radio. Wiklind2 shows that redshifts for these distant sources can be estimated using an alternative method that is based on the shape of the submillimetre region of their spectra. But even this method requires more sensitive submillimetre imaging observations than have yet been made and will depend for its success on the next generation of submillimetre instruments.

Despite their limitations, these methods devised by Chapman et al.1 and by Wiklind2 take us towards a better understanding of star-forming systems, while the next generation of telescopes is awaited. But for the moment, many of the star-forming monsters of deep space will keep their secrets.


  1. 1

    Chapman, S. C., Blain, A. W., Ivison, R. J. & Smail, I. R. Nature 422, 695–698 (2003).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Wiklind, T. Astrophys. J. (in the press); preprint astro-ph/0301403 at <> (2003).

  3. 3

    Puget, J.-L et al. Astron. Astrophys. 308, L5–L8 (1996).

  4. 4

    Holland, W. S. et al. Mon. Not. R. Astron. Soc. 303, 659–672 (1999).

    ADS  Article  Google Scholar 

  5. 5

    Barger, A. J. et al. Nature 394, 248–251 (1998).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Hughes, D. H. et al. Nature 394, 241–247 (1998).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Cowie, L. L., Barger, A. J. & Kneib, J.-P Astron. J. 123, 2197–2205 (2002).

    ADS  Article  Google Scholar 

  8. 8

    Moran, J. M. Proc. SPIE 3357, 208–219 (1998).

    ADS  Article  Google Scholar 

  9. 9

    Barger, A. J. et al. Astron. J. 117, 2656–2665 (1999).

    ADS  CAS  Article  Google Scholar 

  10. 10

    Condon, J. J. Annu. Rev. Astron. Astrophys. 30, 575–611 (1992).

    ADS  Article  Google Scholar 

  11. 11

    Barger, A. J., Cowie, L. L. & Richards, E. A. Astron. J. 119, 2092–2109 (2000).

    ADS  Article  Google Scholar 

  12. 12

    Carilli, C. L. & Yun, M. S. Astrophys. J. 513, L13–L16 (1999).

    ADS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Lennox Cowie.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Cowie, L. Wrestling monsters in deep space. Nature 422, 670–671 (2003).

Download citation


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.