PROBING THE SKY

SPECIAL ISSUE 1999

Course: Galaxy formation and evolution: recent progress

Prof. Richard Ellis
California Institute of Technology (Caltech)
USA

A pioneer in the history of extensive galaxy redshift surveys, Richard S. Ellis, from the California Institute of Technology (Caltech) and Plumian professor of Astronomy (University of Cambridge), provides us in the following intervew with a perspective on the results of the pioneering galaxy redshift surveys as well as those of their current counterparts.

"Faint redshift surveys go back to the early 80’s and were originally motivated by Beatrice Tinsley’s suggestion that the faint blue galaxies found by Richard Kron and others were, in fact, high redshift young ellipticals. Some continued to believe that the joint N(m,z) relation contained useful information on the cosmological model. However, it soon became clear that from the redshift-dependent field galaxy luminosity function that galaxy evolution up to redshift one is rather complex. There is a rapid decline in the number of small irregular blue galaxies whereas the massive regular systems evolve more slowly. I would say this is the major outcome of the redshift surveys.

But standing back a little, I would also say the surveys were essential in opening up our appreciation of the high redshift Universe. At the IAU Symposium 92 "Objects of High Redshift" at Los Angeles in 1979 (yes, I’m that old!) you’ll find the odd radio galaxy and cluster galaxy beyond z=0.5 but no significant field galaxy survey fainter than a limit of B=15.5! Whilst there is still a lot to learn physically about evolution within the redshift range 0 to 1, the first surveys were essential in revealing the depth probed at various limits and this has been useful for many subsequent purposes.

"It’s a natural reaction isn’t it? Let’s push fainter and deeper now we have larger aperture telescopes. However, surprisingly little progress has been made in this direction so far. In the optical, Len Cowie and collaborators have reached B=24.5-25 with Keck, only about a magnitude fainter than was achieved on the 4.2m WHT and I think it’s fair to say the redshift distribution has not given radically new insights because we are just probing a little fainter down the same luminosity function. Greater progress has come from K-limited surveys by Cohen and Cowie which would be sensitive to the presence of massive galaxies beyond z=1. It seems there are fewer such galaxies beyond z=1.5 than expected in models where galaxies formed the bulk of their stars at early times.

The problem we face is that magnitude-limited surveys are terribly inefficient ways to probe a given redshift interval (say 1<z<2) because of foreground contamination. I am not inspired by plugging onwards surveying everything to a fixed limit in the style of CFRS or LDSS. There is also the problem of recognising features which could be anywhere from the UV right through to the near-IR. This implies using a combination of optical and IR spectrographs and somehow deciding a distinct strategy for each instrument.

Fortunately, photometric redshifts have come along following the HDF images. Whilst some dismiss these as imprecise, particularly for blue featureless SEDs, the hope is that they can be used to divide up the samples for more detailed spectroscopy. There is much to do in this area as, until recently, those observers with access to Keck were (perhaps justifiably) diverted by the study of Lyman break galaxies at higher redshift. We now desperately need to learn more about galaxies in the 1<z<2.5 region but designing the appropriate surveys will require some care and a new generation of infrared multi-object spectrographs.

"HST has opened our eyes to what we can learn from resolved images. Although the physical basis of galaxy morphology is poorly-understood, few would argue that the presence of a regular spiral or a well-formed elliptical with passively-evolving colours at z=1 is not important information. We must now ensure our ground-based telescopes match these resolved images with appropriate 2-D spectroscopy. Some distant galaxy rotation curves have been obtained by Nicole Vogt and others but this is just the beginning. Such programmes require excellent image quality through adaptive optics and versatile 2-D optical-IR spectrographs equipped with large format optical and IR detectors.

Thinking of non-optical/IR facilities, we need to unravel the role of dust in galaxy evolution through second-generation sub-mm receivers and space missions like SIRTF as well as to detect the "first light" of stars beyond a redshift 5. In the latter case, I don’t think we should just sit back and wait for NGST but hope the promised growth in IR detector area will encourage innovative searches for emission line signals from the ‘dark ages’. There is still a lot in these areas that can be done from the ground.

So, adaptive optics, integral field spectrographs, sub-mm receivers and large format IR arrays are the key ingredients to develop."

"Well, I look forward to reading Simon White’s response on this one! Let me therefore play "devil’s advocate" a bit. CDM is a remarkably versatile "framework" which can, through the ingenuity of its advocates (who are articulate individuals one has to admit) be "tuned" to match most data currently around because the number of free parameters is still rather large. It is basically a theory for the growth of structure rather than a theory of galaxy formation. So far as the latter is concerned, CDM needs to invoke physical details concerned with star formation, feedback and dynamical assembly in order to predict observables. These are not always CDM-specific which should be remembered when one reads claims that some CDM picture matches the observations.

Even if CDM matched the data, for example through highly-developed semi-analytic models, that would not be good enough in my view. We really need to test the theory’s most basics ingredients. Foremost we need to find clear evidence of non-baryonic dark matter, hopefully independently of some theoretical framework, as well, frankly, of unambiguous evidence for biasing on large scales. I think weak lensing outside clusters is a promising place to invest. Secondly, we need to nail down the cosmological parameters so CDM has less room to maneovre; I suppose we’re making progress there. Finally, observers must rise to the challenge to and try to independently measure stellar and total masses of systems at various look-back times to test the rate of hierarchical growth. That’s going to be very tough but well worth the effort.

So is CDM making progress? Sure! The most unsatisfactory feature of where we are, however, is that CDM is the only show in town. Its proponents have developed a very versatile picture, but are they right? It would be so nice to have more than one well-developed picture to keep these guys in check. Actually I’m surprised theorists are so slow off the mark here; it’s unhealthy for the subject and a little bit boring!"

RICHARD S. ELLIS was born in 1950, in Colwyn Bay, Wales (United Kingdom).

At the age of 18 he entered the Department of Astrony, University College London, where he obtained his B. Sc. First Class Honours in Astronomy (1971).

In 1974 he was awarded his PhD in Astrophysics by the Unievrsity of Oxford, with a thesis entitled "Stellar Abundances and Nucleosynthesis".

Up to 1993 Ellis taught Astronomy at the University of Durham.

Between1994 and 1999 he was Director of the Institute of Astronomy, Cambridge, where he has been Plumian Professor of Astronomy since 1993. Ellis is currently Professor of Astronomy at Caltech (USA).

Among many other Committees, he belonged to the NASA Space Telescope Scientific Working Group: Deep Surveys (1984-85); Space Telescope European Coordination Facility Users' Committee (1988-90); GEMINI UK Scientific Advisory Committee (1990-93); Hubble Space Telescope Time Allocation Committee(1992) and PPARC Science Committee (since 1998).

He was responsible for the instrumentation of the British telescopes installed at the Roque de los Muchachos Observatory (1984-87), La Palma.

Ellis is of the Royal Astronomical Society (1974), the American Astronomical Society (1985) and the Astronomical Society of the Pacific (1985) and has published more than 200 scientific articles.