In the early to mid 80s, while undertaking a PhD at the Royal Greenwich Observatory and the University of Sussex, I first demonstrated that the dust lane in Cen A was a severely warped, razor-thin disk of gas and dust, rather than a thick band. The nature of the warp is consistent with gas accreting and settling into a triaxial potential well. This was the subject of my PhD with subsequent papers appearing in MNRAS (1987) and ApJ (1992). As Israel's 1998 review on Cen A makes clear, these observations and my interpretation are now widely accepted, supported by many subsequent observations. There remain very few scientists (but they do exist!) who still believe that the projected band of dust and gas arising from a thick torus. The most dramatic confirmation of the thin warped disk model came in 2006 from Alice Quillen's beautiful observations with the Spitzer infrared satellite (see below). The warm dust distribution is shown on the left, the model to the right (see Quillen et al 2006, Astrophysical Journal, for more details).

In the mid to late 80s, I worked as a postdoc at the Institute for Astronomy, Hawaii. I demonstrated that M82 has a colossal bipolar outflow along the minor axis (Nature 1988). This was long suspected but never clearly demonstrated. In fact, there were several observations earlier in the decade that claimed the minor axis material was polar orbit gas although Pat McCarthy had done some nice work showing that an outflow was more likely. I also went on to demonstrate unambiguously that there is a widespread scattering halo throughout the galaxy, the first of its kind. I carried out related work on active galaxies with Gerald Cecil, well known for his work in this field.

In the late 80s, I joined the faculty of Rice University, Texas, a private university with a consistent reputation for having some of the best undergraduates. (I still hear from Rice students who fill the ranks of top schools across the USA!) Here, I began to look at the physics of the ISM in galaxies. With a PhD student (who was older than me by a whisker), we were one of the first to demonstrate the existence of a warm ionized medium in a face on galaxy, something that is now widely observed in spiral galaxies. I continued to work on active and starburst galaxies, including a very comprehensive follow-up study of M82 with PhD student, Pat Shopbell, now at Caltech.

A near miss is worth mentioning at this point. I have had a long and productive working relationship with that master of the emulsion, David Malin. In 1987, he showed me these curious fluctuations in the outer parts of Cen A, and we began to talk about the possibility that these were caused by real intrinsic fluctuations (due to star counts) in the galaxy. I had the plates digitized and began to work on the idea of using these fluctuations to a distance to the galaxy. I took the results with me to the Institute for Astronomy, and gave a lunch-time talk on the results. Brent Tully pulled me aside and said that I should take a look at a preprint that had just arrived on his desk. This was a nice paper by John Tonry on converting surface brightness fluctuations to distances. We did ultimately publish our results, and John has been kind enough to cite the paper, truly an act of magnanimity. In fact, our paper shows that the prospect of turning noise into signal has a very long history, starting in the 1850s with early photographic emulsions! We also suggested a range of other applications that were ultimately exploited years later. Oh well.

In the early 90s, I moved to the Anglo-Australian Observatory, Sydney, in order to be involved in big developments that were under way there, in particular Keith Taylor's masterpiece, the Two Degree Field spectrograph. Within weeks of arriving back in Australia, I visited Matthew Colless at Mount Stromlo to suggest that the Aussies should move quickly to form a team that would be able to exploit this machine, something he'd already been thinking about. In collaboration with a distinguished team from the UK, Matthew ultimately led the Australians in the highly successful 2dF Galaxy Redshift Survey project whose impact on cosmology continues to the present day.

While this highly complex machine suffered teething troubles, I began to think about the possibility of `Fabry-Perot staring' in order to reach to extremely faint emission measures. With this technique, I demonstrated that it should be possible to reach down to the cosmic ionization levels (ApJ 1994). I then made the first H-alpha detection of a fully ionized, extragalactic hydrogen cloud (ApJ 1995), initially identified as a "depolarization silhouette" seen against a giant radio lobe. With Ken Freeman and Peter Quinn, I detected ionized gas beyond the HI edge of a spiral galaxy for the first time and was able to extend the rotation curve further in radius (ApJ 1997).

In the mid 90s, Gerald Cecil and I began a very productive and successful collaboration with the Sylvain Veilleux on active and starburst galaxies. Gerald is an absolute master of reduction, analysis and visualisation of complex 3D data sets, without peer at any wavelength in this arena! The three of us have long shared the belief that the entire spectrum must be grappled with to make progress in this field. This collaboration would ultimately lead to a string of papers and an Annual Reviews article in 2005 (see below).

In the late 90s, I returned to the `staring method' once again. In conjunction with Phil Maloney, I developed a model for the halo radiation field in the Galaxy (ApJ 1999), and showed that this was sufficiently strong to produce detectable H-alpha emission from high velocity clouds, assuming that they were within about 300,000 light years of the Galaxy (MNRAS 1998). This is now known as the "H-alpha distance method." With Mary Putman, I carried out a large survey of high velocity clouds and demonstrated for the first time that most of them must be within the orbit of the Magellanic Stream, contrary to the model of Blitz & Spergel that put them an order of magnitude further away.

By early 2000, the Taurus Tunable Filter (TTF) was beginning to produce a string of exciting new results, including measures of the star formation rate density in the general field and in the vicinity of powerful galaxies. I was fortunate to work with a number of students and collaborators in order to realize a series of key results in this area: J.C. Baker, J. Barr, K. Glazebrook, D.H. Jones, P. Francis, to name a few.

In 2000, Ken Freeman and I were invited to write a review for Science on the formation of the Galaxy, a subject we have discussed off and on since we both overlapped as Visiting Scientists at the Institute for Advanced Study, Princeton in the late 80s. We were both involved in writing the science case for ESA's Gaia astrometric satellite set to launch in the next decade. This incredible mission aims to obtain positional and velocity information for a staggering 1.5 billion stars. This motivated us to think about what one could ultimately learn from so much data, and how these data might be supplemented further by ground-based instruments.

This led ultimately to our Annual Reviews article in 2002, and the suggestion that massive stellar surveys should be carried out from the ground. Here and in the 2000 review, we introduced the idea of a signature hierarchy to describe galaxy formation; "chemical tagging" (fingerprinting) from high resolution spectroscopy; "galactic archaeology" and "near-field cosmology" to describe this new thinking, although the use of the latter dates to my Nature review in 1999. Together, we designed the science cases for the RAVE (with Matthias Steinmetz), WFMOS (with Rosie Wyse), and ARGOS stellar surveys. This work has received a major boost from Gayandhi De Silva's PhD work (2006) that demonstrates the viability of chemical tagging.

In 2002, I was fortunate to hear a talk from Martin Cohen (Berkeley) about new mid-infrared observations of the Galactic Centre. It struck me that the faint bipolar wisps I could see in his data were reminiscent of the degree-scale radio lobes published by Sofue & Handa in 1984. What was immediately obvious to me was the significance of seeing dust (and by implication, dense gas) entrained in the walls of a bipolar outflow: inter alia, the inferred energetics would be much higher than derived from the radio observations. Our ApJ 2003 paper was the first detection of a galactic bipolar outflow at mid-infrared wavelengths, and the first demonstration of a powerful bipolar wind in the Galaxy. What is particularly striking is that this highly energetic outflow would be very difficult to detect in external galaxies -- bipolar winds may be far more common than we currently believe! More recently, I have published evidence (with A. Quillen) for a galactic wind associated with the radio jet in Cen A, and an Annual Reviews article (2005) on "Galactic Winds" in collaboration with S. Veilleux and G. Cecil.

Since 2004, I have begun to look once again at the edges of disk and dwarf galaxies, but this time at the stellar component. This work was motivated in part by the success of the Cambridge group in using the Isaac Newton Telescope to look at the outer reaches of Local Group galaxies. Our long-term study is concentrating on the closest group of disk galaxies, the Sculptor group. Our first results for the outer disk of NGC 300 are spectacular (ApJ 2005) showing that the disk extends over almost a degree on the sky! This work emphasizes the enduring mystery of how stellar disks are able to form over such an extent. Over the same period, Matthew Coleman's PhD (2005) demonstrated the presence of substructure in the Fornax dwarf, although this phenomenon is not evident in the Sculptor dwarf.

Since 2005, I have worked on accretion and feedback processes evolved in galaxy formation and evolution. I will be presenting a review of this topic as part of the Saas Fee lecture series in Switzerland (March 2007).