Oral History Transcript — Dr. Richard Ellis
This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.
This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.
Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.
See the catalog record for this interview and search for other interviews in our collection
Richard Ellis; July 27, 2007
ABSTRACT: Scope of Material: Galaxy cluster cataloging, and toward finding supernovae started in 1984 while professor in Durham, England. Collaboration with Danish astronomers, using telescope at European Southern Observatory in Chile on this search. Thus, with ground-based telescopes obtained the spectra of the supernovae and measured their redshifts. Result: Nature paper of 1989. Collaboration with the Supernova Cosmology Project (SCP) beginning in 1993. Carl Pennypacker first PI of SCP. Camera at Anglo-Australian Telescope (1000 by 2000 pixels, as opposed to 600 by 500 pixels of Danish camera). Euphoria of finding supernova at record-breaking distance, at redshift of 0.45. Turning point in collaboration around 1994, 1995. On skepticism and science. Experiment and theory. On Lambda, the cosmological constant, and the accelerating universe. On SCP and High-z teams, the UK National Meeting (with George F. Stathew and Martin Reese), and talk at Cambridge Mathematics Institute. Use of Hubble Space Telescope.
Pavlish:It is July 27th, 2007, and I am here at Caltech University in Robertson Hall to interview Professor Richard Ellis. I did a bit of background research. I looked at your website. You detailed three projects in which you are primarily involved. The first is gravitational lensing; the second, distant supernovae; the third, galaxy formation and evolution. On the website, you relate the first and second projects through your work in “defining the weak lensing case and survey parameters for SNAP, a major satellite mission in conjunction with the supernova cosmology group.” I excerpted your paragraph on distant supernovae, and that is what I will ask you about today.
Pavlish:My first question is, how did you come to join the SCP in the late 1980s.
Ellis:Yes, so I was in England at the time. I was a professor in Durham, in the North of England, and I was interested in galaxy clusters. A galaxy cluster is an assembly of galaxies. Several thousand that are all at the same distance. I got the idea that if one took images of these clusters repeatedly over a number of months, we might be able to find supernovae in them. Our group at Durham was one of the few that had a catalogue of distant clusters. We teamed up with a group of Danish astronomers in Copenhagen who had regular access to a small telescope in Chile. So we set out to do a project in about 1984 to repeatedly image these clusters that we had access to, for supernovae.
Pavlish:Is the Chile telescope the Cerro Tololo?
Ellis:No, it was the European Southern Observatory. It was a Danish two-meter telescope. Every month, one of us would go down to Chile and observe for a week or more. Then we’d analyze the data and compare the images with an earlier image taken the previous month. We would look very closely to see whether any of the galaxies had changed. It was a very interesting project. It was very hard work. We found the first distant supernovae. We wrote this up in Nature in 1988.  It demonstrated the potential of finding distant supernovae. We had very primitive equipment compared to the equipment that we have today.
Pavlish:But already CCD cameras?
Ellis:They were CCD cameras but they had very small fields of view. In some sense this was the forerunner of the big surveys that we now have — the High-z Survey and the Supernova Cosmology Project. But what it did illustrate is the capability that with ground-based telescopes we could get the spectra of the supernovae and measure their red shifts. The supernova we found was only at a red shift of 0.3 whereas the most distant supernovae that we can now study are beyond a red shift of one — with technology and of course Hubble space telescope having been launched. That is how I got into supernovae. I got very excited about measuring the expansion history of the universe. But I realized it was going to be very very difficult and painful.
Pavlish:I have heard of this Nature paper as the spur to the other groups. From your narration, it sounds like it comes out of galaxy clusters rather than initially looking for supernovae. How did it come together?
Ellis:Vicariously. I had a Postdoc, Warrick Couch. He and I were interested in the evolution of galaxies and clusters. The idea is that you look at clusters at different distances. I hired him as a Postdoc. His thesis was in that area. Then we realized that we could simultaneously get wonderful data on clusters but we could also be searching for supernovae. He and I worked with the Danish group to do this using their telescope. It was a beautiful collaboration. The Danes had access to the telescope. We were interested in the clusters and finding supernovae.
Pavlish:They weren’t the ones who initiated it? Were they also interested in supernovae?
Ellis:Oh, yes. We were equal partners. The paper was published in 1989. The Supernova Cosmology Project then began to survey not clustered areas, but blank areas, with much bigger cameras. The fields of view of their cameras were so big that one could guarantee that there would be a supernova in each exposure. They began this Australia and then later in the Canary Islands. There was a very strong collaboration between the UK and Saul Perlmutter. In the meantime I moved from Durham to Cambridge. A group of us at Cambridge were the UK contingent of the Supernova Cosmology Project. But it took a long time for the Supernova Cosmology Project to get everything working. After the paper we wrote in 1988, the next paper with credible supernovae from the Supernova Cosmology Project wasn’t until something like four or five years later. It took them quite a long time. There was a period when the Supernova Cosmology Project was being reviewed. Obviously, they were spending a lot of money to get this camera working and everything. There was even a moment when their sponsors were somewhat concerned that their progress was quite slow. There was even a risk that this whole project would be cut off and canceled. This was at Berkeley. I was visiting Berkeley in the 1990s when this was happening. I was trying very hard to ensure that the project would continue to be funded.
Pavlish:How did your collaboration with Saul Perlmutter begin? Did he read your paper and contact you?
Yes. There was an earlier principal investigator, Carl Pennypacker. Carl Pennypacker was the original PI. He visited Durham at my request because he was clearly wanting to get into this area. The UK has very strong connections with Australia. At that time we had a joint telescope called the Anglo-Australian telescope. Pennypacker wanted to put his camera in Australia, on the Anglo-Australian telescope. So given that I was in Britain and had demonstrated the method, then he approached me. We agreed to collaborate. He established a very strong connection with Australia with my former Postdoc, Warrick Couch, who moved to the observatory after he worked with me. So essentially, that’s how we got in. Eventually Saul Perlmutter took over as principal investigator. You know, I think Pennypacker was creative, but he did not have the organizational skills of Saul Perlmutter. Saul made it his mission to make this project work successfully so he deserves the credit for the eventual success, although a lot of the concepts were put in place by Carl Pennypacker. The real breakthrough in the survey was the development of a big camera. The Danish camera was very small. It was — let me think — something like 600 pixels by 500 pixels. That meant that if you just pointed at a random direction in the sky, the chances of finding a supernova were quite low.
That is why we went to clusters where of course there are many more galaxies in a small field of view. But Pennypacker had a camera which was initially 1000 pixels and then later 2000 pixels square. The cameras that we are using now are even bigger. The whole subject has been enabled by having bigger images of sky. Just like digital cameras, you know, there are more and more pixels. This allows you to survey larger and larger areas of the sky. Then, the volume of space that you’re sampling is so large that you’re guaranteed even in a single exposure to find a supernova. Nowadays, many. That means you can be very efficient. You can then shed the follow up spectroscopic time when you measure the red shift of the supernovae. You see, in the old days (this was really hilarious when you look back on it) in the Danish days and in the days when Perlmutter was starting up, when we found a supernova it was chaos. We’d have to phone around the world to get anybody on a telescope to give up what they were doing to try and measure the red shift of the supernova before it faded away. This was just so chaotic. Eventually, when the cameras became big enough you could then confidently say that every exposure you took would find five or six supernovae. Then the following night you could allocate a spectroscopic night, confident that you would have targets to look at. You didn’t have to phone around the world and find friends to help you. That is what took time. Saul of course will tell you this in detail.
That is what took five years or so: to get this new method of observing working. I played a role at the Canary Islands in the early 1990s, in measuring the red shift of the next most distant object, the record breaker at the time, which was at a red shift of 0.46 or so. That was exciting. Pennypacker was still in charge of the team at the time. I can remember that was euphoric. This was the first significant supernova at a high red shift from the new team, from the Pennypacker team. I phoned in to California, and he was extremely happy. We wrote that up in 1994, 1995. Then things really took off after that. Once we knew how to do it, it didn’t take that long. By 1999 we had 50 supernovae and we could see that the universe was accelerating. That was a very exciting five years. From say 1995 to 1999 was a truly amazing time.
Pavlish:When did you start actually getting data, or observing, doing analysis for the SCP, rather than with the Danish collaboration.
Ellis:The data probably started really rolling in, around 1996, 1997.
Pavlish:You were in the group that found that 0.46 z supernova?
Ellis:Yes. [Pause for a minute. Let’s find out when that was.] The first significant result from the Supernova Cosmology Project was the supernova that I measured the red shift of in the Canary Islands in August, 1992. That was at a red shift of 0.45. That was more distant than the one we found with the Danes. We wrote a paper on it. For some reason, the paper did not appear until 1995. That was a milestone. In that period from 1992 to 1995 Perlmutter took over from Pennypacker as PI. I would say that the first significant data started rolling in, in 1995, 1996, and then we started constructing what we call the Hubble Diagram to see how fast the universe is slowing down.
Pavlish:That is an interesting historical question. When did you start constructing the Hubble Diagram? Did you do it as the data came in?
Ellis:Yes, oh yes. Everyone was excited to know what the answer was. There is an embarrassing twist to the story, that the first eight supernovae that came in suggested that the universe was slowing down and we wrote a paper we were widely criticized for because we subsequently changed our minds when we wrote the paper based on the 50 supernovae in 1999. That is the one that everybody cites. But there is an earlier paper, in 1997 where we got the wrong answer from the first seven or eight supernovae.
Pavlish:I would not call that embarrassing, though. I will get to that in my final question, about ‘how experiments end.’ My advisor has written a book of that title. In the fourth chapter he analyzes the discovery of neutral currents. There were also two competing teams, and similarly they had a negative result for some time, and then they found the positive result. If one subscribes to a very traditional story of how science works, maybe that would be a bad thing. I think, in fact, it may be good. How many supernovae have been found as of today?
Ellis:At high red shift? Four, five hundred. Maybe more.
Pavlish:Yesterday I went to hear about the new supernova surveys where the scientists are planning to find thousands of supernovae in the coming years. I was amazed.
Ellis:Yes, it has truly exploded.
Pavlish:Let me continue with the prepared questions, so that we do not run out of time. But please feel free to redirect the conversation. That 1992 observation seems very important.
Ellis:It was a demonstration of the Supernova Cosmology Project, yeah.
Pavlish:My question was, did you join with a specific function in mind?
Ellis:Yes, my role was to help with the Canary Islands observations. I was not the only British person on the team. I was still at Durham in 1992 and then I moved to Cambridge in 1993. And so then, I was involved in coordinating a lot of the British activities. We were a very substantial part of the team, because the William Herschel telescope in the Northern Hemisphere was one of the key telescopes for measuring the red shifts of the supernovae, including this first demonstration object.
Pavlish:Who else would you include in that collaboration?
Ellis:In that category? Yes, Mike Irwin was a key person in the team. Richard McMullen, and the other guy who was involved was Nick Walter. That was it. Then there was still the Australian effort: Brian Boyle, and Warrick Couch. And then there was a group in Berkeley of course. That is where a lot of the processing was done. That took a while. The criticism of the project is that Saul works in a government laboratory and so the government laboratories fund these initiatives. There was a review committee which would review the progress. So for a number of years, from about 1989, 1990 up till 1993, they were getting a lot of money, but there were no publications. No supernovae had been published. So there were a lot of questions being asked of the project at that stage. Part of the delay in getting the thing up and running, was processing all the CCD images to find the supernovae quick enough that you could go to a telescope and measure its red shift before it had faded away. Obviously, there is no point in finding supernovae, waiting for two or three months, and then trying to get the red shifts because they would have all disappeared. You would not know whether it was a supernova or not. That is where a lot of the frustration was in the early years, and that is where the criticism was. Probably the turning point was in 1994, 1995 when things started clicking and the team started functioning very effectively.
Pavlish:As a side question, do you regard a supernova as scientific object or phenomenon?
Ellis:It is an event, really. It is a cosmic event. Increasingly, as we learn more and more about them, they become fascinating in their own right. Why should supernovae be so uniform in their properties? To what extent can we trust them not to be changing in their properties at different cosmic times? This is the focus of my work today. Not that we question their use in the past, but for what I call precision work, for using supernovae in future projects, where we use thousands of supernovae, where we are wanting much more accurate measurements of dark energy, are they up to the task? Or, is there some intrinsic dispersion in their properties that is going to plague us in, say, ten years’ time. It is very important to address that now, before we spend large sums of money on future satellites.
Pavlish:That goes to a more philosophical vein so I will skip to a later question, and then get back to some of these factual ones. One way to ask a more sophisticated version of the question, how discoveries are made, is to ask how experiments end. I would like to ask this question of you. I have an exposition detailing different examples where historians have looked at how scientists come to a consensus or a conclusion, where the evidence has built up enough. It is interesting to ask not only how theory influences the end of an experiment but also how an experiment becomes decisive or convincing to the scientists.
Ellis:You mean to the members or to the community?
Pavlish:Well, a publication like Physics World would like to know…
Ellis:When did I believe it?
Pavlish:When did you believe it, when did the team believe it, when did the scientific cosmology community believe it? Mike Turner would say, 1998. Something to brainstorm about.
Ellis:I was originally skeptical. We had, after all, published a paper in 1997 which said that the universe was slowing down. So, when more data came in and we were faced with reversing the conclusions of our earlier paper […] After all, you have to realize, that at this time, we were a pioneering experiment. Firstly, I felt we were ahead of the other team. I had great confidence in Saul as an individual that the project was now in full steam and was doing excellent work. But I was worried that we seemed to be chopping and changing. So I was initially very skeptical in 1998 when the results started pointing to an accelerating universe. I think when I was working on the paper with Saul, I slowly started to realize how careful our analysis was and how many checks we had done. And then, I heard the rumor, in fact when I was visiting here in 1997, I think, I heard a rumor that the other team was puzzled about a similar result. This is mentioned in Kirshner’s book.  So, I started to think, if the other team is finding it too then I should take it seriously. When the paper was written in 1999, I was quite confident that we were onto something. I was willing to be convinced that it was wrong. I felt — as often is the case in science — there are ways of presenting results that are tantalizing, as a challenge to the community to think about them. In other words, astronomers are allowed to change their minds. It is healthy for science to publish results, as long as they are written up in an honest fashion, in which the pros and cons of an argument are clearly laid out. I feel very proud that that Saul Perlmutter paper was extremely careful in arguing what the uncertainties might be. But it nonetheless concluded that the universe was accelerating. So I was happy to take that paper, and for instance, in the UK, giving my own talks promoting this case.
Pavlish:What was that like?
Ellis:Well, it was a great time because the result had a lot of attention. I personally was quite surprised how quickly the theoretical community accepted it. But there were other reasons. The community wanted an accelerating universe, theoretically. It helped explain some of the paradoxes about the ages of stars in the universe. The problem of galaxy formation required what we call the cosmological constant. So when the observations came along supporting it, rather amusingly many theorists grasped it uncritically and accepted it which I thought was naïve. If anything, I would say that although I bought the argument very early on in the mid 1990s, what I was most amused by, was how quickly and uncritically the community accepted it. I expected that we would have a much harder time convincing people of this result. Now, subsequently there have been other indications. In fact, papers I have written in other areas, which have supported the accelerating universe. Now, I think, people are a little critical of the supernova data. They say, well, yes, the supernova data were there early on, but it is the combined evidence from the cosmic microwave background, gravitational lensing, and so forth, that give us the idea that we live in an accelerating universe. I think that is unfair. I think the supernova data were the first observational evidence that we should take seriously the accelerating universe. It got a lot of attention, a lot of publicity, and I think that was correct.
Pavlish:Had you considered a positive Lambda before the results started coming in?
Ellis:No, no. I honestly did not believe, when I set out in the 1980s to try to do this experiment — I was trying to measure what we call the deceleration parameter. We even put some bounds on it in our papers. There were two papers with the Danes. One was our high redshift supernova discovery, and the other on another supernova we had found. We did attempt to put some constraints but it was on the rate at which the universe was slowing down. It wasn’t until 1997, 1998, that we had to swallow hard and say, oh my goodness, the universe might be accelerating. Up until that point, there were people in the community that were advocating a cosmological constant, including people in Cambridge, but I was fairly critical of them.
Pavlish:The steady staters?
Ellis:No, that was much earlier. That was pretty well was demolished in the late 1960s. That was a much earlier generation of cosmology.
Pavlish:Do you see a contrast in the research methods of the High-z Supernova Search Team and the Supernova Cosmology Project?
Yes, yes. I admire both teams, but they are very different. When I was in the UK, I had pretty good connections with Harvard. There was a period when Cambridge and Harvard were collaborating on a possible joint telescope. One of my first tasks when I was director in Cambridge, was to go to Harvard and work with Bob Kirshner who was my equivalent. So I got to know Bob well. Bob is on the other team, and I was working with Saul. As the 1990s progressed, we became rivals, even though we were very close colleagues. Here is my take on this. The SCP was a pretty much hierarchical team, with Saul Perlmutter eventually as the PI. All of the activity on the photometry based in Berkeley. The peripheral groups had quite precise and well defined activities. We had to help with the Canary Islands runs, we had to get the spectra ok. But Saul was a very strong leader who insisted that all of the processing of the data, the first analysis, would have to be done under his supervision. This is very much like a particle physics mentality. The other team, the High-z team, was less coherent. It did not have a senior leader that was as far above the others in coordinating power. You can read in the literature, that Kirshner regards himself as the senior player. Adam Riess, clearly was an energetic young individual who did a lot of the analysis and became very distinguished in his own area later. And Brian Schmidt, who was chosen as a logical, fair-minded guy who could maintain a horizontal group where there were many people of equal stature, where nobody was willing to necessarily be a leader. But everybody trusted Brian, and so Brian became like the father figure who would manage them all. Even though Kirshner was more senior, and Adam Riess did a lot of the work. So it was a very different organization.
You might ask which was the more effective? I think the High-z team had more creative input. There was more encouragement amongst the individual members to contribute new ideas and to think of things in a different way. I have always been impressed that their papers are constantly looking at the data and analyzing. I think a lot of that came from the freedom that each individual member had, that they could make a significant contribution to the project. Whereas I think in the SCP team, everything had to be controlled by Saul. So it was a lot more difficult. Eventually we solved this problem. When I moved here (I moved here in 1999), I managed to get an area of the project that I could lead myself, to look at the properties of supernovae. That has worked out very well. That is still within the umbrella of the Supernova Cosmology Project. As people got more and more involved, they are entrusted with doing their own programs.
Pavlish:You work at observatories, if I understand correctly. What about doing data analysis?
Ellis:Yes. Usually I work with students and Post Docs. One of my former students, Marc Sullivan, was here at Caltech. He and I worked with Saul on quite an important subsequent paper in 2003 which looked at the Hubble diagram for supernovae depending on the kind of galaxy they sit in. This was quite influential in proving that the acceleration is real and not due to effects like dust obscuration. There were a number of objections to the two papers of the two teams. There still are people who would love the result to go away. There are a lot of people who tried to show, this is 2000-2003, there were a series of very influential papers which tried to criticize the two teams’ work. And so the battle continued. Even though a lot of people accepted it, the battle did continue. There are a number of papers from both teams. I worked intensively on one of them with Marc Sullivan and with Saul. So, Marc was here, and the data was analyzed here.
Pavlish:You say that there were people contending the results. That relates to my question about milestones or turning points in the SCP’s research program. One important event was the Aguablava conference in Spain around 1995. That would be before the results came in, but a time when the two teams came to present results together. I hear that there was a supernova team from Italy there as well.
Ellis:Yes, that is right.
Pavlish:The Fermilab workshop in the spring of 1998, where a straw poll was taken.
Ellis:Yes, that is right.
Pavlish:The Great Debate of 1998, between theorists, Turner and Peebles.
Ellis:I was in a difficult position because I was not in the US and so did not go to these meetings in Chicago. But we had equivalent meetings in the UK. There was a debate in Liverpool: is the universe going to expand forever, or not? That was one of the big questions. The players there were George F. Stathew, Martin Reese, and myself. There was a poll taken there.
Pavlish:When was that, around?
Ellis:1996, 1997. I have to look it up. That was terrific. This is what we call the UK National Meeting. It is like the AAS. There was an audience. I was the [astronomer] observer, to answer, ‘what does the observer think of the situation?’ So I reviewed our latest work. And then Martin Reese proposed the open universe, and George F. Stathew proposed the closed universe which slows down. And then there was a vote, I think. People preferred a closed universe. People were not ready for the accelerating universe. There was a lot of theoretical prejudice against what we called the ‘neat and tidy universe’, also known as the Einstein De Sitter Universe where the mass density is very high and the universe just stops expanding at some point in the infinite future. Many theorists loved this model. But it had a fundamental problem. It couldn’t explain the distribution of galaxies we see. And also, its age is quite short. The history from the big bang to the present day would be much shorter than the ages of the oldest stars. There was a problem to be solved there. The supernova data was just opening it up. People were just beginning to see that there might be a solution here. I think at this Liverpool meeting, they weren’t ready for it.
Pavlish:So your observational results at that point should have supported Reese’s argument?
Ellis:No, this was at a time when we had only eight supernovae. We were saying that the universe was closed. We were supporting George F. Stathew. Then, two years later I gave a talk in Cambridge at The Mathematics Institute (I remember working very hard on this talk) to demonstrate why I believed you should believe the new results and not the old results. There was a riot. People were saying, “You know Richard, you told us two years ago to believe this. Now you are telling us a very different story.” The room was packed. In the end, I think I did a pretty good job, but it was hard work. I had to convince people that we now had a result that was robust.
Pavlish:What did you do to convince them?
Ellis:I showed them — this is somewhere where Saul deserves a lot of credit for helping me. I showed that in the first eight supernovae data, the data wasn’t good enough to give the right answer. I went through each individual data point and showed why it was erroneous. Then I showed the new data and showed how the new data was of much higher quality and we had many more of them.
Pavlish:The instruments you used were pretty stable over the duration of the project?
Pavlish:You hadn’t started using the Hubble Space Telescope though.
Ellis:No, the Hubble came later. In fact, that really began as I was moving to the US. In 1999, and this is really the work I did with Marc Sullivan, I proposed to Saul, that we now use Hubble to go back to where the supernovae had exploded and to look at the galaxies in which they occurred, even though they had long faded. Just take a snapshot of a galaxy and determine what kind of galaxy it is. I led that program and brought it to California in 1999. That led to this paper by Sullivan, which showed that the universe is accelerating even if you concentrate on supernovae that explode in dust-free elliptical galaxies. And so we started presenting the work in Washington in 2001, 2002 (I was in the US, so I was much better connected.) That added to the confidence that we had a good result. Then there was a subsequent paper in 2004 by Knop which analyzed excellent data. A smaller number of supernovae, maybe 11 or 12, but all taken with Hubble, very high precision data. Those are the two papers that, I think, the SCP published subsequently, that really just added strength to the result.
Pavlish:I have a series of short questions — they are short as questions but maybe require long answers. One of the questions was how have your colleagues come to accept the accelerating universe, but I think we’ve discussed that.
Ellis:Yes, slowly but surely.
Pavlish:This is not an interview question for the archive. As a historian, I’m interested if you have any documents or images that you think a historian should be aware of. You know, there is the problem of scientific papers and publications. Sometimes there are lab notebooks or personal correspondence.
Ellis:I have all the papers on the discovery. The first paper with the Danes is interesting. We can look at that in a minute. We can see what we can find. I was clearing that out the other day and came across it. It is quite interesting.
Pavlish:I may even have it with me, actually. That would be great. I was going to ask you about technologies of supernova searching.
Ellis:Yes, I think we’ve covered that pretty well.
Pavlish:How does the resurrection of a discarded concept — the cosmological constant — change our view of the universe?
Ellis:Well, that is profound. We are now in a situation where we have a component of the universe, seventy percent that we do not understand at all. The cosmological constant is not a theory. It was introduced by Einstein, you know, without any physical motivation. There is no physical motivation for it. There are theories in physics, which can explain why empty space should push outwards, what we call negative pressure, but not with this kind of value. In fact, they are miles off in predicting the strength of this acceleration. This is why the two papers from the two teams are so highly cited. With no exaggeration, it has redirected the lives of hundreds of theorists who are looking for an explanation for this acceleration. Despite what they may tell you, the progress is very slow. That is why we need more observations, to try to refine. For example, is it a constant, or is it changing its property as the universe ages. That is something I think we can actually over the next five years make real progress in addressing. Do different methods of measuring the acceleration find the same result? If they don’t, then that might tell us that the laws of physics are not exactly what we think they are. That might mean that Einstein’s relativity needs some adjustment. Whatever the outcome, it is exciting. It is an adventure, really. I feel privileged that we lived during this time; especially the second half of the nineties.
Pavlish:A golden age?
Ellis:It was an exciting part of my scientific career, you know. We haven’t solved it yet. We’re still working on it.
Pavlish:Do you think this was a discovery in the conventional sense, if there is such a thing?
Ellis:It was classic. The public loved it: Buffins go in and make a measurement and they come out baffled. You know, it is very much a storybook scientific discovery. Result we do not understand. Overturns over a previous theory. That is why it has been so popular, I think. Also, anyone on the street can understand it. Most intelligent people know that the universe is evolving. They know that the universe began with a bang. Many people know that the universe was expanding. For a long time people pondered what is going to happen. Will the universe expand forever? Will it slow down? This has been a quest of large telescope science. Even the 200-inch here at Palomar was built to answer this question. And it never did. It couldn’t keep up with the supernova technology and the discoveries were made elsewhere. This has been a longstanding question and people are interested in it.
Pavlish:If you had to put a timestamp?
Ellis:1998. For me, it was the momentum within the team that was 1998. That is when the team became confident that we really had a result. And the community, probably two years later.
Pavlish:With the Cosmic Microwave Background results of the year 2000?
Ellis:Yes, a combination of the low cosmic density and the microwave background flatness. That was 2000 and 2001. Boomerang and Max. That was an exciting time. So, this is the Nature paper of 1989 and this is the referees report. Nature sends it to two referees. The first one says, “This is a tremendously interesting report. The detection of a very faint object which is almost certainly a supernova in a distant cluster. As the authors correctly point out this holds promise for global tests in cosmology. Their observations are heroic. Their detection is a tour de force and the results are tantalizing.” That’s great.
Ellis:The second referee says, “It is very impressive, the authors were able to find a supernova at a red shift of 0.3 and even more impressive they were able to find a crude light curve and spectrum. This is an excellent start on a project of possible cosmological importance and the paper deserves publication in Nature.”
Pavlish:You were aware that it was of possible cosmological importance.
Ellis:Oh yeah. That is why we were doing it. We were disappointed it was so tedious. Here’s the problem: one of us had to fly all the way to Chile and then observe all night, and then stay up most of the day reducing the data, because obviously, you need to find a supernova straight away. It is just so exhausting to send somebody to Chile every month, you know, for a week. We’d end up totally exhausted. I have a letter also, where the supernova had been found and I was desperate to get a spectrum. I have a fax here, to Australia, where I say that we have found probably the best case yet, a type 1A supernova at a red shift of 2.3. I had made another measurement myself.
Pavlish:That’s a fax from Chile?
Ellis:No, it is a fax to me in England to Australia where Warrick Couch got the spectrum. Yeah, here are the detections. Here is the spectrum. And here is the analysis of the light curve. The data is terrible when you compare it with the present day quality of data. All the pages are fading, as you can see.
Pavlish:The methods that you put in place were followed by the two teams.
Pavlish:When you say that, what does that mean exactly? What methods?
Ellis:Measuring the red shift is necessary to confirm the distance and velocity. Getting the shape of the spectrum confirms it is a type 1A supernova and not something else. There are various other types of supernovae. And measuring the light curve, and making sure that you get it at peak light.
Pavlish:Beyond peak or at peak?
Ellis:Well, to go through peak so that you can estimate what the peak is.
Pavlish:And measuring the red shift, you do that by comparing it to other distance indicators.
Ellis:No, the red shift comes from the spectrum and the lines are shifted from where they would be in a local supernova. So we have a spectrum of local supernova and then we see what the shift is. We have a press release here which just says, “After the initial discovery on August the 9th, 1988, the Danish astronomer Hans-Jorgard Nielsen, whose turn it was to conduct the monthly search, noticed, and he immediately alerted Richard Ellis of Durham, who followed up the discovery within hours with a more powerful British 2.4 meter telescope. In superb atmospheric conditions (gosh, I don’t know who write this), Ellis was able to resolve the supernova as a star like image separately. This clinched the event as a supernova. Ellis was able to persuade two British astronomers observing in Australia to take a spectrum and that appeared in the Nature paper.” Now, does it say…here we go (I’ll give you a photocopy of this): “for example if the intrinsic brightness is the same, independent and they would determine whether the expansion of the universe is the same or whether it is slowing down and would reverse. On the basis of one event, it is not yet possible to determine this with certainty.” [laughs] So modest, yes. Why don’t I photocopy this? Yes, this is a classic. A press release from twenty years ago.
Ellis:That’s about it, I think.
Pavlish:How great that you have saved the lab notebooks till now.
Ellis:In those days we did not have fax machines. The fax machine sort of came in in the early nineties. There were telexes.
Pavlish:That is how you communicated with the other observatories?
Look at this: “Danish correspondence.” No email. Telexes. There were no fax machines. We did have email but it was not very reliable. It was good fun.
Hans U. Norgaard-Nielson, Leif Hansen, Henning E. Jorgensen, Alfonso Aragon Salamanca, Richard S. Ellis, & Warrick J. Couch, "The Discover of a type Ia supernova at a red shift of 0.31" Nature Vol. 339, 15 June 1989.
Robert Kirshner, The Extravagant Universe, Princeton University Press.