Home  /  Interviews  /  Nicholas Scoville

Nicholas Scoville

Nicholas Scoville

Francis L. Moseley Professor of Astronomy, Emeritus

By David Zierler, Director of the Caltech Heritage Project

February 7 and 17, 2002

DAVID ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Monday, February 7th, 2022. I'm delighted to be here with Professor Nicholas Z. Scoville. Nick, it's great to see you. Thank you for joining me today.

NICHOLAS Z. SCOVILLE: Thank you, David.

ZIERLER: Nick, to start, would you please tell me your title and institutional affiliation here at Caltech?

SCOVILLE: I'm Moseley Professor of Astronomy, and I'm emeritus at this point but I think I still have the named chair.

ZIERLER: Nick, who is or was Francis Moseley, and did Moseley have any connection to your research?

SCOVILLE: The previous recipient of that chair was Maarten Schmidt, but I don't know the history of Francis Moseley, though I am most grateful for the support provided to me and Caltech.

ZIERLER: Nick, when did you go emeritus?

SCOVILLE: When I was probably 72 or 73. It's called early retirement. But at Caltech, that's kind of a joke because they actually give you a year or two off from teaching and responsibilities, when you're still on the faculty, and then you become emeritus after those two years. That's to induce you to retire, which I'm entirely positive about. I think that the idea of having 75- and 80-year-old people teaching young undergraduates and graduate students is not optimum because I know from my own experience that they're simply not as sharp as they used to be, and they're also probably not quite as current. I'm probably a little bit unique and still doing active astronomical research. But a lot of older faculty—not so much at Caltech but at other universities—they keep going on and on, and one wonders whether it's just to collect a salary. But I'm completely against that. I was stimulated a little bit in that direction by Peter Goldreich, but then I also stimulated a couple of other people like Anneila and Tony Readhead to become emeritus also.

ZIERLER: Nick, on that point, given that you're still active in the research, what are you currently working on?

SCOVILLE: My earlier research which we'll get into later has dealt with the physics of gas and dust in the interstellar space. This originally involved observation of the star forming gas in the Milky Way and then as observations became more sensitive, nearby galaxies. I'm now studying the overall cosmic evolution of the interstellar gas in galaxies from the present out to redshift 6 (only a billion yrs after the Big Bang. The interstellar gas and dust is what fuels star formation, and it also feeds active galactic nuclei supermassive black holes in the centers of the galaxies. I recently developed a new technique for measuring the gas content of galaxies at high redshift. I'll explain it later on, but it's using the far-infrared radiation – the so-called Rayleigh-Jeans continuum to measure the total mass of dust. Then, under the assumption that the gas-to-dust ratio is more or less constant, which it seems to be locally, then you can infer the overall gas content of the galaxies. Typically, that ratio is 100 to one, that is, 100 times more mass in the interstellar gas than in the associated dust which is mixed in with the gas uniformly.

ZIERLER: Nick, for this research, what aspects of observation are most relevant for you? In other words, is it space-based astronomy or ground-based astronomy that's most valuable?

SCOVILLE: Actually both. I did an enormous project where I—if you want to measure the rate of star formation in galaxies, you can do it partially in the optical. But if you observe, you know, use rest-frame optical data will suffer extinction from the dust which is associated with the gas. The optical astronomers put in large extinction corrections on these higher-redshifts galaxies of the order of a factor of 5-10 increase in the deduced star-formation rate from what they actually are observing. That's a real deficiency because you really want to know how rapidly these galaxies are evolving, and you get that from understanding what the star-formation rate is in the galaxies. In order to get more reliable star-formation rates, the first part of my project was to go to the infrared observations mainly from the Herschel Observatory, which was a space-based satellite about 10 years ago, launched by the Europeans. It had US partners but, basically, it was a European mission. One of the areas in the sky which they surveyed was the COSMOS field. This field I initiated in 2003 starting with a grant of 600 orbits of the Hubble Space telescope (HST) – the so-called Cosmic Evolution Survey. It's a two-square-degrees field on the celestial equator, and within it there's something like 10,000 Herschel-detected sources. In the optical with HST we detect approximately a million galaxies.

Now the Herschel far infrared data gets us a measure of how much dust-embedded star formation there is in the star forming clouds in the galaxy -- as compared to visible wavelength light which detects only the star formation on the edges of the clouds. For example, in the nearby Orion Molecular Cloud. There, you c (~1300 light years away) the Orion Nebula is seen very easily with visible wavelengths and even the naked eye, but it is clearly only on the front surface of the cloud in our direction – there is an equivalent far infrared luminosity from the dust within the cloud -- seen in the infrared but which isn't seen at all in the optical. We basically had these far infrared measures of star formation activity in 10,000 galaxies in the COSMOS, two-square-degree survey field. When I (and the associated team) started that survey, we located it intentionally on the celestial equator so that it could be observed by telescopes in both hemispheres and, most importantly, so it could be observed by the ALMA Submillimeter Observatory in Chile. Alma is a 64-telescope interferometer at very high frequencies in the far infrared/submillimeter. In my research over the last few years I found that there were something like 2,600 ALMA observations within the field, and I downloaded those from the archive. The earliest observations were ones which I had made. But then most of the observations I downloaded were taken by a large group of other scientists not including me. That was a mammoth task to download all that data because it's an interferometer, and interferometer data sets are very extensive—plus, it's 2,600 separate observations. I downloaded them, checked the calibration of them, and then I had to measure at the locations where I'd found an infrared source. I measured that flux seen by ALMA at long wavelengths, that is, so-called ALMA Band 7 and 6, which are at 345 gigahertz and 230 gigahertz -- a millimeter to the submillimeter wavelength. I then have a flux as observed by ALMA from the long-wavelength dust emission, providing estimates of the dust and gas masses for each galaxy out to a billion years after the Big Bang. I then correlate those masses with the star formation activity as evidenced by the Herschel observations. At the shorter wavelengths, around 100 microns in the rest frame, then we're measuring the dust-embedded star formation.

I end up with, in the case of the COSMOS survey field, 700 objects for which I have good data in both observatories. But then in the COSMOS survey field, we have very high accuracy for what are called photometric redshifts where the optical and near-infrared and UV bands observed by multiple telescopes have been taken, compared with expected special energy distributions in galaxies in order to deduce both whether the galaxy is star forming at all and, secondly and more importantly, what the photometric redshift is for the galaxies. The photometric redshifts for the 700,000 galaxies in COSMOS then tell us their distances (or cosmic lookback time) assuming that they're all taking part in the cosmic expansion.


SCOVILLE: In the overall evolution of the universe, it is well known that the star formation started out slowly but then reached a peak at about redshift 2, which is about half the age of the universe, and then declined to the present day at redshift 0. The physical question is, why did star-formation rate grow early on and then why did it fall off (from the peak at redshift 2) toward the present epoch? The natural assumption would be that it grows because of increasing gas content in galaxies, and that it falls off due to decreasing gas content. Here, for 700 galaxies that now have the gas content estimates as a function of both the stellar mass in the galaxies as determined from the optical, spectral energy distribution fitting, and then also the redshift or cosmic epoch of the galaxies, I can—the other thing which I can do is to look at the star-formation rate as deduced from the Herschel observations and the gas content as determined from the ALMA observations. In addition, the ratio of the star formation rates to the gas mass tells you the efficiency for forming stars per unit mass of gas in the galaxies (and we now know that for a sample of 700 galaxies.

One of the things which has been common in my research has been the recognition that when you have two galaxies which collide, the star-formation rates will vastly increase because during the collision, the interstellar gas clouds get compressed by running into each other. Once you compress the gas, it become higher star formation efficiency would be evidenced in this sample, i.e.it is much easier for the gas clouds to collapse and form stars at a higher rate. For that whole sample of 700 galaxies, we now have both the cosmic epoch of each galaxy, the interstellar gas content, and the star formation efficiency—the star formation efficiency being derived from the ratio star-formation rate to interstellar gas, as I already said. We can then analyze the evolution with redshift and see whether the evolution is due to increasing gas content or changing efficiency for forming stars out of whatever gas there is. I find that going back from the present day to redshift 2, two-thirds of the increase in star-formation rate in galaxies is due to increasing gas content, and one-third is due to changing efficiency of using that increased efficiency for using that gas. That's really pretty cool. Nobody's ever done that before. We separate out whether you're forming stars in a more efficient way. Now, if you look at spiral galaxies in the local universe, you actually see a lot of star formation along the spiral arms, and not much in between the arms. There, the same issue arises because it looks like in the interior regions of galaxies like the Milky Way, the dominant gas component is the dust and molecular hydrogen clouds. If that is a dominant component, then those dust clouds can't just be in the spiral arms. They must exist also in the interim regions because the current astrophysical understanding of spiral galaxies is that during an orbit around the center of the galaxy, material comes into the underside side of the arm, goes along the arm for a little bit, and then comes out the front side of the arm—the front being the convex curvature side.

ZIERLER: Nick, more broadly, I'm curious what some of your—

SCOVILLE: [laugh] David, will you want to get me down to a level that—?

ZIERLER: That's exactly what I want you to do. That's the name of the game, Nick. More broadly, I wonder what some of your key takeaways are from the recently released decadal report from the National Academy, Astro2020. What does it tell you about some of the priorities in the field, and where research and support are headed?

SCOVILLE: I hate to tell you I really haven't read the report more than the superficial summaries of the report. Back when I first started doing research, the way science and science funding proceeded, people would submit proposals, their one proposal. There would then be a bunch of peers, who would review it anonymously. [laugh] I haven't been drinking. Don't worry, David.

ZIERLER: [laugh]

SCOVILLE: Then, each proposal would, one, be a concrete proposal because people wouldn't submit a proposal until they had all the details, what the budget was, what the capabilities would be. But, nowadays, what seems to happen—unfortunately, in my opinion—is that there's a decadal survey committee, which superficially looks at all the fields, and reviews projected plans for various facilities, and then sets priorities. I think it's fine to set priority in terms of what science you want to do. But choosing the scope of the facility on the basis of usually probably a three-hour presentation to some committee without a lot of details—most of it being fairly superficial—I think is not a good way to evaluate future facilities. It cuts out, of course, a lot of smaller projects because people put in a big project because it will generate excitement, but then the actual specifics and the budget are left sort of hanging. Now, a good example of this is JWST, the James Webb Space Telescope. It originally was projected and sold to NASA as a roughly $500 million object. Probably every astronomer in the country knew that it could not be done for $500 million, and yet it was sold that way precisely (I believe probably to a decade survey committee) and the head of NASA. We all know that the budget is now probably over $10 billion, which is a factor of 20 increase. It'll be a great facility. It just underscores the way that decade survey committees proceed, and the irresponsibility on the part of scientists in the field.

ZIERLER: Nick, a little closer to home, how closely have you been following the Thirty Meter Telescope saga?

SCOVILLE: Not terribly closely. But, I must admit, every time I go to my heart doctor [laugh], he's enormously impressed with Mauna Kea, and he keeps telling me he's very depressed with the local opposition to the telescope because, obviously, it's such a fine facility. Mauna Kea already has a full-scale observatory. The mountain is not going to go back to a pristine condition. It's fantastic that there's this tremendous intellectual resource there on Mauna Kea on the Big Island, so Hawaiians should be proud of it. I think probably most of the Hawaiians are proud of it. But, nevertheless, there's enough opposition that it's constantly slowed down. I think eventually it should go there and will go there. But it's been a big distraction to have all the delays going on.

ZIERLER: Nick, some nomenclature questions. Of course, there's astronomy. There's astrophysics. There's cosmology. What would you say is your home discipline, and have those terms changed over the years?

SCOVILLE: I would say in my case, it's definitely astrophysics, which has the connotation of being somewhat theoretical but then astrophysicists also do observations. In my case, probably two-thirds of my research has been observations, but then one-third has been modeling those observations or proposing a new insight in terms of what all the observations mean. I call myself somewhere in that middle place. Obviously, I got a degree in astronomy when I was in graduate school (Columbia), and our department is still called the astronomy department. Now at Caltech majors may choose whether their major is called astrophysics or astronomy.

ZIERLER: Now, this is perhaps as much a generational question as a scientific question, but having each foot in one field, so to speak—in other words, you're involved in observation, but you do modeling as well—is that still done or is that something that came from an earlier generation?

SCOVILLE: It's certainly still done. But if you look at most young astronomers just out of graduate school or postdocs, there's usually not much modeling of what they see, or the modeling can be a simulation program where there are lots of bells and whistles, and you never really know quite what's going on. The kind of modeling I like to do is a simple sort of order of magnitude understanding of what's going on; how things are related. One of the things I find a little disturbing is that a lot of people nowadays just the observations, publish the analysis of the observations in a way that other people have analyzed similar data maybe on different objects, and then not enough of the big physical picture as to what's going on. But you asked also whether I'm a cosmologist, and I've always considered that term philosophically. I think we were all brought into astrophysics by cosmology because what greater question could you ask than the overall evolution of the universe? On the other hand, I consider it now a little bit pretentious to call yourself a cosmologist because cosmology is really a very complex and physical field that, nowadays, requires simulations and minute comparison of the observations, or departures in the observations from what's predicted with the models. In a sense, I feel it's gotten kind of too complicated. The basic ingredients of cosmology, that there's an expanding universe, our age is 12.7 billion years, and that there's a period of decelerating expansion, and now accelerating expansion, all that, it's great but it's not really what I've been doing.

ZIERLER: Nick, how do you see your research contributing more broadly to fundamental questions about the universe?

SCOVILLE: There are two aspects to that question. If you look at the galaxies that we see in the optical and the infrared, and with ALMA, those galaxies are made up mostly of baryonic material, meaning normal matter, it will appear in the periodic table, or fundamental particles. Now, that baryonic matter only consists of about 4% of the overall mass and energy in the universe. On the other hand, it's the part of the universe we can observe most readily, and that part is evolving, galaxies are forming, clusters of galaxies are forming, stars are forming within those galaxies, and the gas is being used up in those galaxies. In a sense, the research which I do has great impact on what people actually observe as astrophysicists and astronomers. But you have to remember that it's really only 4% of the universe. I've always found it wonderful that there are things—I thought when I was in graduate school and a postdoc that our rate of progress was so great that there wasn't going to be anything left to do—

ZIERLER: [laugh]

SCOVILLE: —for the next generation of astronomers. But, amazingly enough, the more you do, the more you find out, and the more intriguing questions that come up. I've forgotten where I was going with that. But it just keeps going on, and getting better and better, so I'm really happy that I chose astrophysics or astronomy as a discipline, rather than physics as a graduate student at Columbia. The natural thing would've been to do physics. But, at that point, physics looked fairly stagnant. This is back the late 1960s. I took a couple of astrophysics courses, and was amazed that there were all these unsolved problems. My department was very much a theoretical astrophysics department, so, at the same time, I learned a fair amount of theory, and I was the first observational graduate student that they had. I was also working with my advisor in a new area of observational astrophysics, that is, looking at these molecular clouds in our galaxy, and then eventually other galaxies. I really came on, maybe stumbled onto a fantastic route to a career with lots of new stuff.

ZIERLER: Nick, I wonder if you can reflect on some of the challenges and opportunities when looking at a galaxy like the Milky Way, which we're in, versus galaxies that are much farther away.

SCOVILLE: Well, the first challenge looking at the Milky Way is that we're sitting in the disc of the Milky Way, which could be imagined as being a plate or a platter, and most of the stars and gas clouds and star formation are in that disc. If you look up out of the disc—we're about two-thirds of the way out from the center to the edge of the Milky Way—you start to see the other galaxies and stars in the halo of our galaxy. Now, as you well know, for a long, long time, nobody knew that there were other galaxies like our galaxy. It was only back in the 1920s that this was settled by a detailed measure of some nearby galaxies, finding that there were nearby galaxies just like the Milky Way. But the big challenge in observing the Milky Way is simply that there are these dust clouds, the same clouds which I observe, which have enough dust that you can't see through them. They are typically extincted by a factor of 100 million at visible wavelengths. If you look at the Milky Way, you say, oh, there's this beautiful band of stars across the sky. But then if you actually look and think quite carefully, you actually notice that there are two bands of stars. The reason why there are two bands of stars is because the dust clouds in between the bands which obscure the light from the more distant stars. Most of the stars you see are somewhat out of that thin disc of gas and dust, which obscures looking to the galactic center. If you look towards our galactic center, there's something like 30 magnitudes of visual extinction, which every 100 magnitudes—sorry—every 5 magnitudes corresponds to a factor of 100 decrease in brightness. It's a big decrease in brightness. Then when you go to other galaxies, of course, many of them you see not edge on but more or less face on or partially tilted to a line of sight. In which case, then you can see the overall structure of the galaxy. In our galaxy, that structure, the spiral structure, is really only deduced from radio observations, both in atomic hydrogen and in a molecular gas emission, which I and colleagues have observed. In other galaxies, you can very easily tell. Is it a spiral galaxy? Is it a barred spiral galaxy? Is it an elliptical galaxy, having no gas and dust, and being fairly spherical in shape? In a sense, observing these distant galaxies is relatively easy. You understand what they're doing. On the other hand, of course, they're much fainter because they're far away, and so you need large telescopes with a lot of collecting area. In addition, those distant galaxies, as you go out, say, half the age of the universe, their light is all redshifted by the cosmic expansion, and so the dominant component of their light is actually shifted into the near-infrared or mid-infrared. You need a large infrared telescope to actually probe the galaxies accurately.

That's one of the things that in the decade survey, one really, really needs is a 20-meter above the atmosphere infrared telescope because in a lot of the infrared, you can't observe due to the absorption by molecules in the Earth's atmosphere, for example, water vapor or CO2 or methane. You really need to get above the atmosphere to have a completely unobstructed view. There are windows in the near infrared. But then when you go into longer than, say, 20-microns wavelengths out to about 300 microns, the atmosphere is more or less completely opaque. That's precisely the regime in which most of the infrared luminosity from galaxies is emitted, particularly those high-redshift galaxies. That's what's sorely needed to understand the evolution of galaxies, a high-redshift, and the Herschel Observatory, which I spoke of earlier, is a first attempt to do that. But it was a very small telescope, and one could easily do a factor of 100 or 1,000 more sensitive nowadays.

ZIERLER: Nick, either directly or indirectly, has having JPL been a part of Caltech, has that been an asset to your research at all over the years?

SCOVILLE: Not so much mine, although certainly it stimulated a lot of the infrared astronomy at Caltech, and a sizable portion of my research is actually in collaboration with Professor Neugebauer and Tom Soifer and Mike Werner. I appreciated a lot having their contact. Even though they were at Caltech, they stimulated a lot of work at JPL, and obviously were helped by JPL. Gerry Neugebauer was the one who basically led the so-called IRAS survey, which was the first all-sky survey in the far-infrared of the sky. That was absolutely critical in highlighting the most interesting objects in the universe, which weren't known at all before that. I mentioned earlier on colliding galaxies. Those were highlighted by the IRAS survey because, all of a sudden, they discovered this collapse of galaxies which were emitting 100 times as much infrared luminosity as was seen in the optical.

Then follow-up observations by one of my colleagues, Dave Sanders, who was a postdoc here at Caltech, and is now a faculty at U of H, did follow-up ground-based imaging and spectroscopy of those galaxies, and discovered that a high fraction of them were actually irregular galaxies which looked like they had two galaxies merging. Hence, the belief nowadays, which is fairly common, that star formation can be enhanced in merging galaxy systems. At the same time, I was in charge of the Owens Valley Radio Observatory, and I started a program of observing those same galaxies. We found that there were enormous gas concentrations in the nuclei of those galaxies, which you don't find in normal galaxies. In the Milky Way, there's perhaps 100 million solar masses of gas and dust in the central few hundred parsecs. For reference, the total mass of gas and dust in the Milky Way, is about two billion solar masses. Now, in these IRAS galaxies, especially the ones made up of two colliding galaxies, you actually find that there is often several billion solar masses of gas within the central 10 or 20 parsecs, meaning 30 - 60 light-years from the nucleus of the galaxy. It's really phenomenal the way that gas has been concentrated in the nuclei of those galaxies, and we wouldn't have known about those galaxies if it hadn't been for the IRAS survey, which was probably the most important infrared survey that's ever been done. On the other hand, it did have low resolution and sensitivity, so it was really a survey of galaxies out to perhaps a gigaparsec in distance.

ZIERLER: Nick, some technical questions. How does one go about measuring the mass of a galaxy?

SCOVILLE: The most reliable technique is simply to measure the rate at which stars and the gas clouds in the galaxies are moving—their velocities—and measure the size of the galaxy, and then use Newton's laws to deduce the mass from the gravity which is required. This is called the virial mass estimate despite the fact that the cloud is not in a steady state equilibrium. We use the same technique to measure the mass of these dust clouds in our Galaxy because there again, we can observe the velocity field in the dust cloud by observing the molecular gas emission doppler shifts. Then you can map how big the dust cloud is by looking at the extent of the emission at each velocity within the cloud. That was one of the first things I did in my thesis work, was to study a few of the molecular clouds which had been analyzed in the Galactic center in the inner disc of our galaxy. Doing precisely what I described, I found that if you measure the velocity and motions in those clouds, and their sizes, you needed a mass exerting gravity of the order a million solar masses to hold the clouds together. It was clear that the clouds were being held together, although it's still not really understood the nature of the motions in those clouds. As I was about to say at one point long, long ago [laugh], one of the things which I love in the field is finding problems which I don't know the answer to, because they stay with me throughout my career. It's not like I think about them all the time, but I stumble upon some idea when having a scotch at night or whatever. I have enjoyed being sufficiently interested in my field to not dissociate from it completely (although it makes me sound like a workaholic). I never view this as ‘work'; I'm simply enjoying trying to develop a new understanding.

ZIERLER: Maybe it's a very fundamental or even elemental question, it's so important for your research, but when you think about mapping a galaxy, relative to what? What are the reference points that you use to understand where galaxies are, how they're formed, the basic questions you have about galaxies?

SCOVILLE: I'm not sure quite I understand the question.

ZIERLER: In other words, do you need reference points that are fixed in order to understand how to go about mapping the galaxy?

SCOVILLE: No. Well, I mean, we have a coordinate system, a celestial coordinate system. I'm probably not answering your question. But we know where the galaxy is in the sky. Galaxies have a diameter beyond which their brightness approaches background brightness of the nearby sky, but this is not usually a sharp cutoff. If you go further and further out in a galaxy, ultimately, the self-gravity of the galaxy also becomes less than the gravity of nearby galaxies. That occurs in the case of the Milky Way roughly, say, a third of the way to Andromeda. That's kind of the sphere of influence of the galaxy. On the other hand, the Milky Way and Andromeda Galaxy and Large Magellanic Cloud were all part of a the same cosmic large scale structure, so this group is called the Local Cluster Galaxies, including other much smaller less massive galaxies – perhaps 100 in total. On the other hand, if you go to the nucleus of the Virgo Galaxy Cluster, there are there probably a couple thousand galaxies, and this cluster extends over a much larger cosmic volume. The galaxy clusters are ultimately joined up into the large-scale structure of the expanding universe. The different galaxy clusters have different sizes and masses.

Now, the big unknown -- I mentioned that only 4% of the mass in the universe is baryonic while roughly 27% is dark matter. We still don't know what that dark matter is, what constitutes that dark matter. So, 27% of the mass of the universe is this material we don't even know the nature of. Then, the expansion of the universe also requires a ‘dark energy component', which also we do not know the nature of. Obviously, there's theoretical speculation about what the ‘dark matter' and dark energy are, but that's fundamentally amazing that 96% of the mass and energy in the universe is not really understood. Now, it's not being an arrogant scientist, but I assume within the next 100 years, the nature of those two components will be understood. It may be very likely sooner, but we'll see.

ZIERLER: If you had to wager, what's the tougher nut to crack, dark matter or dark energy?

SCOVILLE: I think the dark energy because the dark matter, presumably is probably some fundamental particle, and ultimately it may be detected. But the dark energy could be related simply to the cosmology of the universe. That component, I don't really understand myself. Maybe nobody does. It's not a field that I've really spent any time in.

ZIERLER: Nick, really broadly speaking, do you see your research contributing more fundamentally to the age-old problem of developing a theory of quantum gravity, or merging general relativity with quantum mechanics?

SCOVILLE: No, I do not. That was an easy question. [laugh]

ZIERLER: [laugh] Why is that? Why are the things that you study about galaxies not relevant there?

SCOVILLE: Well, because everything I do in my own research involves fairly standard physics of the baryons and the electromagnetic radiation. It's the interaction between one physical process and another which I find so interesting. In a sense, putting it down a little bit, it's applied physics or astrophysics, but it's known physics that's being applied. I like that because at least I can tell when I suggest something, what's reasonable and what isn't reasonable, as opposed to it being a speculation for which there are 10 other opposing speculations. Sorry to be negative.

ZIERLER: No, that's totally fine. Well, Nick, before we go all the way back and develop your narrative personal history, one last sort of broadly conceived retrospective question. You said before, humorously, back in graduate school, you thought it'd all be wrapped up by now. It's sort of reminiscent of when they said at the patent office with the telegram, "We can close down now. There's nothing left to be invented." Right?

SCOVILLE: Right. I had never heard that quote.

ZIERLER: As you reflect over the course of your career, thinking about all of those advances that have happened, where do you see what has really moved the needle? When has it been about advancing new theories? When has it been about manpower, people power, just putting more minds on it? When is it about real big shifts in technology, and advances in instrumentation and even computation?

SCOVILLE: Well, the first easy answer is that major shifts have come about through new telescope facilities and interferometers, which basically didn't exist before I was in graduate school. I mean, they existed in a very simple, three-element scheme. But essentially, as I said, all of the research analysis I've done has been applied physics. There's a second part of your question, which—you mentioned three things, right?

ZIERLER: Well, there's the theoretical advances, there's just more people, more minds on the problem, and then there's the technological advances.

SCOVILLE: The theoretical advances, I think, essentially of the physics which I applied to understanding the structure of galaxies, and the interstellar matter content, and star-formation rates was physics which was known when I was in graduate school. It's not like it required a new theory of quantum mechanics, or a new addition to quantum mechanics, and so there's essentially no new physics involved. But it's applied physics and, of course, not to sound arrogant, but not everybody recognizes and has the right physical interpretation. That's what I sort of am priding myself on now. I mentioned that I was very lucky to have ended up in the field that I did, or subfield that I did – field which was essentially unexplored before then (IR and mm wavelength astronomy). That was driven by technology. Before I went to graduate school, there were no observations at frequencies higher than about 30 gigahertz, that is, 300 times FM radio frequency. On the other hand, the molecules (H2) are the dominant gas components in star forming clouds, Their quantum mechanical spectral transitions are essentially all at higher frequencies, that is, mm and infrared wavelengths. That required development of new technology and more precise radio mixer receivers, amplifiers and spectrometers to analyze the down-converted signal. All of that was developed (as well as the front-end detectors), by my collaborators, who were Arno Penzias, Bob Wilson, and Keith Jefferts at Bell Labs. Bob Wilson was earlier on graduate student at Caltech. They developed these diode mixers or down converters. My thesis advisor, Phil Solomon and myself were very fortunate to work with them on the observation side. The Bell Labs team has taken their radio receiver to a telescope on Kitt Peak in Arizona, a 12-meter telescope, and did the first detections of molecules like CO, carbon monoxide, and I think HCN. This couldn't have been done without their radio receivers. But then, of course, once that had been done, then the National Radio Astronomy Observatory and other groups at radio observatories started developing higher sensitivity detectors, which were cooled to liquid helium temperature. An important figure in the latter was Tom Phillips who was then at Bell Labs but who came to Caltech in PMA.

The sensitivity which we had when we started out was expressed in terms of the equivalent thermal noise, it then was like 2,000-3000 degrees Kelvin. Now, that sensitivity has gotten down to roughly 40 or 50 degrees Kelvin, so this has been an enormous increase in sensitivity by cooling the receivers. That's what's enabled us to go from looking at clouds in our galaxy to nearby galaxies and recently the galaxies out to cosmic age 0.6 – 1 billion years. In addition, of course, the now 10 yr old ALMA array in Chile has tremendous sensitivity by having 64 telescopes, and using these high-sensitivity receivers. Developing dishes which can operate efficiently in the atmospheric windows at 100, 200, 300 gigahertz (wavelengths of 3, 1.5 and 1 mm). Thus, the dish surface has to be parabolic to an accuracy 1/20th of one millimeter over their typical diameter of 10 m. You'd also like the dish to be big to collect as many photons as possible and give you high angular resolution. That, of course, requires clever design of the structure to be stiff and the surface to be smooth. One of the Caltech's professors, Bob Leighton, in the physics department, he was the one who designed the telescopes which were used up in Owens Valley at the Radio Observatory. He also designed and built the telescope which was the Caltech submillimeter observatory on Mauna Kea. This was used by Tom Phillips group for 20 years.

ZIERLER: Well, Nick, let us—

SCOVILLE: Have you done an oral history for Tom Phillips?

ZIERLER: Not yet; on the list.

ZIERLER: Well, Nick, let's go all the way back to the beginning. Tell me about your family, your birthplace.

SCOVILLE: My father was a graduate student in chemistry at Cambridge University in England during the beginning of World War II, and my mother was trained as an artist but didn't really go to an art school. But then when the war broke out, they went to Portugal where my grandmother was living, and then came back to the states, and my father went to the University of Rochester to finish his PhD in chemistry. My mother, at the same time, despite being an artist, she learned to fly, and she trained Army Air Corps pilots to fly during the middle part of World War II. It's amazing to me that this small woman, my mother, would actually fly the Stearman plane which has an enormous radial engine in the front. To start the engine, she had to pull the propeller! I, of course, never saw it because I was born in 1945. [laugh] But they had an interesting life. My father, although I don't know what he was working on in Cambridge, he ended up doing research on gas masks for his PhD at Rochester. I don't know what his intellectual interests were. But then after the war, my father became involved in nuclear weapons testing in both Nevada and in the South Pacific. I think, throughout his career, he was somebody who had an excellent scientific background but who also could be understood by people who did not have the science background (politicians, military and the public). He was involved in the nuclear weapons testing, evaluating the tests, during the '50s, and then went into the Central Intelligence Agency when Kennedy was in office, and I think became a deputy CIA director under Allen Dulles. Ultimately, a couple of years later, he became a deputy director of the Arms Control and Disarmament Agency. Then when the Republicans came in, he left government, and became a well-known liberal advocate for nuclear disarmament on the outside of government. If you go on the Web, he's actually much more famous than myself. He had a huge following and was unimpeachable due to his intelligence background. He did that for probably 25 years before he died.

I remember he came out to visit me at one point during the mid-80s when I was at Caltech. At that point, he had developed cancer, and he probably only had about three more years to live. But he would still travel around the country, giving talks, and I was amazed at how he bore up under the strain of the cancer.

Once my father died in the late '80s, my mother became the glue of the family. She ultimately moved up to Connecticut. Both my parents originated from the Northwestern corner of Connecticut, towns called Salisbury, and then Norfolk. She continued to do her art there after my father died and, in fact, took up steel welding of large sculptures. I think she was stimulated to this a little bit by me because I had worked, when I was at the end of college, building a house with my brother in Vermont. There was an art student from Yale there. He and I bought a cheap welder, and we made junk sculptures in the fields near the house. I still do welding as a hobby and some artistic welding and some furniture. I like that very much. I find, in contrast to doing science, I love the fact that when you do something artistic is entirely your creation. Not from an ego point of view but I like the fact that you don't really know what you're getting into until after you start working on it, physically, and start doing it. By contrast in a lot of the astronomy I've done, data comes in from the sky, you record it, and then you analyze it. But from the start in these observation, you know you are heading in an interesting direction and you are constrained by the physics understanding in your final interpretation. A lot of the time, when I do something artistic but I develop the concept as I go along. Of course, professional artists of develop a scheme for what they're doing, and it may become more mechanical work or like scientific inquiry at that point.

ZIERLER: Nick, where did you grow up mostly? Was it in the Washington area?

SCOVILLE: No. My father and mother lived in McLean, Virginia, a suburb of Washington and Virginia. But I always hated it there, and they sent me away to boarding school in Massachusetts, and then I was there for four years, I went to Columbia for both undergraduate and graduate school.

ZIERLER: Did your dad involve you in his work at all? In other words, did you have a good sense of what he did, even when you were a kid?

SCOVILLE: No, not really. He was quite private about it, although he was involved in nuclear weapons arms treaty. In fact, the whole family went to Geneva several summers, and I got to meet several of the Russian and US negotiators, many of who, at that time were very intellectual academic scientists—for example Hans Bethe was an advisor and close friend of my father. It was a very intellectual group of people. I also remember Geneva being an incredibly beautiful place, and we were there for two summers for a month each time.

ZIERLER: Was it always math and physics that you gravitated towards in school?

SCOVILLE: Yes, because I wasn't good in anything else—

ZIERLER: [laugh]

SCOVILLE: —[laugh] and one might question whether I was good in math also. [laugh] But one of the things I do have is—and it probably comes from thinking a lot in the background of my mind—I have a very good physical intuition about the way things work, and I find that invaluable. I think that's probably the thing that I consider most important to recognizing something which is a problem, and then having the insight to put things together, and figure out an interpretation or explanation for what the observations have revealed.

ZIERLER: When you were thinking about college, it was specifically physics that you wanted to pursue?

SCOVILLE: No, in fact, I had trouble getting into college. I know one is not supposed to say this as a professor at Caltech.

ZIERLER: [laugh]

SCOVILLE: [laugh] My advisor at high school managed to get me in the back door to Columbia University. By "back door," I mean the engineering school, which had just been founded, and so I went there.


SCOVILLE: The first year of engineering school, we had courses like in drafting, and I couldn't believe how boring it was. It seemed like all of the analysis in engineering was kind of by rote techniques, and I didn't like that at all. I liked much more free-form thinking. For the second year, I transferred to Columbia College. Amazingly enough, Columbia had a spectacular physics department, although the professors were getting old by then. But there must've been seven or eight Noble Prizewinners in the physics department there, and it was amazing. But I won't say that I understood everything that they taught. [laugh]

ZIERLER: Who were some of the giants on the faculty at that point? Who stands out in your memory?

SCOVILLE: Well, there was a Professor Polykarp Kusch, who was a Nobel Prizewinner. He was very intimidating as a lecturer because he was so precise, and if you made any noise while he was writing on the board, he'd turn around and scowl at you. T.D. Lee also was there, and I took a course in statistical mechanics from him. He was phenomenal, very quick, and amazing. He also was a Nobel Prizewinner. The astronomers in my department, although they didn't publish prolifically, they were extremely knowledgeable from a theory point of view, so I appreciated doing theoretical analysis from them. I was also equipped by them to do it in astrophysics.

ZIERLER: What was the distinction, as you saw it in undergraduate, between theory and experimental physics?

SCOVILLE: I did not work on research projects with any the theoretical physicists; however most of my friends were physics graduate students; most of them end up doing experimental physics. I haven't maintained regular contact with the other astrophysics graduate students, although one of them I have seen occasionally.

ZIERLER: What about astronomy? Was that a separate program? Did you do any astronomy as an undergraduate?

SCOVILLE: No, although I did, as I said, take the two basic astrophysics courses offered to sophomores and seniors in physics. This was the only astronomy/astrophysics background I had as an undergraduate; however, as a graduate student, I did pick up basic astronomy though being a teaching assistant in astronomy courses. In my opinion, that is the best way to learn the basics which after all are not that difficult and one learns them well.

ZIERLER: Was anybody working on general relativity during those years?

SCOVILLE: Yes, there was Mal Ruderman, who was professor originally at NYU and then Columbia. One of the amazing things to me was when I went back to give an honorary lecture, the Bishop Lecture in 2010. I was amazed to see these professors like Ruderman and Spiegel sitting in my colloquium audience. I had liked both of them very much when I was in school there. It was was also amazing to me to be giving this lecture with 100 people in the audience, many of whom were professors I had been terrified of when I was [laugh] a student there, and to see them all sitting, listening to me, I couldn't believe it. [laugh] It was a trip. [laugh]

ZIERLER: Was observation or astrophysics, was that something that you wanted to pursue as you were thinking about graduate school?

SCOVILLE: I don't think I had any particular idea about which area of astrophysics or astronomy I would pursue. But I ended up working with Phil Solomon who was a Lecturer there at time. His background had been astrophysics theory with a specialization in the physics of the interstellar medium. He was a very aggressive, competitive researcher, so I learned that you need to be fairly outspoken to make an impact. You don't want to be mean to people but you shouldn't be tolerating people advocating ideas you consider wrong. One of the things, I picked up early in my career, was listening to talks and colloquia, and asking questions which either guiding the speaking to communicate better with the audience or revealed some direction of the talk that I believed was wrong. Now adays, the graduate students seldom do this – making the colloquia not nearly as exciting – after we are presenting new ideas and understanding.

ZIERLER: Did you apply more broadly for graduate school? You were happy at Columbia? It was good to just stay put?

SCOVILLE: [laugh] That's a sensitive question [laugh], and you don't know why but let's see if you can guess.

ZIERLER: Well, I know this is a very interesting time to be at Columbia in the late '60s and early '70s.

SCOVILLE: No, let me just tell you because you don't want to embarrass yourself. I actually applied to graduate school, a standard group of graduate schools, and I didn't get into any, and so the astronomy department at Columbia said, "Well, you can stay here."

ZIERLER: That's as good a reason as any.

SCOVILLE: No, one also doesn't expect a Caltech professor to say that they didn't get into graduate school.

ZIERLER: Listen, these discussions indicate that even Caltech professors are people too. [laugh] There's value in that.

SCOVILLE: That's good to know. [laugh]

ZIERLER: Now, on the social side of things, were you politically active at all?

SCOVILLE: Yes, and so in the late '60s, there was the political activity at Columbia, which was the first of the major campuses to become active against the Vietnam war. There had been previous activity at Berkeley, such as the free speech movement, but that had died down by the mid- to late '60s. I was a member of what was called Students for a Democratic Society—not a leader or anything, but I would go to their meetings quite regularly—and they were the ones who initiated the protests against Columbia building a gymnasium in the public park just to the east of the campus, Morningside Park. I found it very illuminating what happened. I was one of the students—I was a graduate student at that point, '68—sitting in the buildings. There were probably 800 students occupying the buildings. In my case, it was a building which had mostly graduate students, so it wasn't like there was anything violent going on. When the campus was cleared by the police, The New York Times had a front-page banner headline "Students cleared from Columbia campus peaceably; about 50 students arrested" or something like that. That article was in fact written before the police entered the campus and took the students out of the building. When the police started grabbing people in front of me in my building, I decided I was going to walk out. I didn't want to be hit over the head with a Billy club—although maybe it would have improved my brain. [laugh]

ZIERLER: [laugh]

SCOVILLE: But, in any case, I at that point didn't get in any fight with the police. Although when my brother was visiting me, I actually did have a struggle with police at that point to free my brother. But to finish this it was just amazing to me that because the publisher of The New York Times, Sulzberger, was a Columbia trustee, and probably a number of the Columbia trustees were of that ilk. The New York Times completely fabricated their coverage. In fact, hundreds of student were arrested and the police were very violent.

ZIERLER: Did you ever talk to your father about—?

SCOVILLE: I should point out, one of the things that the students in SDS had been emphasizing was the interlocking interests at the higher end of our society—example, Columbia trustees and people who ran businesses in New York. It was completely illustrative of that, and it was pretty amazing that this was brought out so clearly in the demonstration. Sorry, you asked about my father.

ZIERLER: Did you ever engage with your father in political issues?

SCOVILLE: Yeah, I certainly talked to him about them, but we didn't get into heated arguments, probably because neither he nor me were argumentative people. He also is very level-headed—was very level-headed—and extremely rational to talk to; not angry in the least, and not dictatorial. We had good conversations. He, of course, didn't want me being arrested at a demonstration, but he never said, "Don't do it."

ZIERLER: Was the draft something you needed to deal with?

SCOVILLE: Yes, and so that was one reason, although it wasn't really my reason for staying in graduate school, but it did protect you a little bit for a few years. But then, at some point, the government decided that they weren't going to exempt automatically students. My department chair, Lodewijk Woltjer, who is European and became ESO director but was chair of our department, he actually wrote a very supporting letter that I was doing critical teaching and whatnot. I don't know whether that influenced them at all. In the end, I think I would've gotten drafted, but I ended up in the draft—they had lottery, and I ended up with a number which was like 175 or something like out of 365.

ZIERLER: Nick, what was the process developing your thesis research?

SCOVILLE: Boy, that's a long question. [laugh] The first observing run I went on was this one with Phil Solomon and Patrick Thaddeus—Pat Thaddeus—and that was to follow up the first detection of formaldehyde molecules. Now, it completely shocked astronomers that you could have complex molecules existing in interstellar space because one would think that the ultraviolet rays would split apart molecules, break them down. It's amazing that there were some molecules as complicated as H2CO, formaldehyde, which we discovered in the inner galaxy in these giant molecular clouds. One of the things that was peculiar about the formaldehyde data which had been taken earlier was, normally, you would expect a molecule to have a base temperature equal to the cosmic background radiation temperature. At the present epoch, this is 2.7 degrees Kelvin. You would think that the lowest formaldehyde excitation temperature, characterizing the populations of the different quantum levels of H2CO, would be equivalent to a Boltzmann distribution at 2.7 degrees Kelvin. In fact, in some of the early observations of formaldehyde, we were seeing that the excitation temperature is like 1.5 degrees kelvin. There were some people who suggested that the cosmic background at the wavelengths which were relevant to excitation of formaldehyde was distorted from a 2.7-degree black body. It hadn't been observed at that point. Then other people suggested that there were strange cross-sections for excitation of formaldehyde with respect to molecular hydrogen. The two professors I was with had developed a theory involving the departure of the cosmic background from a strict blackbody spectrum. At Green Bank, West Virginia, which was where there is a 140-foot radio telescope, where we could observe the formaldehyde, and make more extensive measurements of this cooling below the cosmic background.

It was amusing. We arrived down there. None of us had ever been to a telescope before. (None of us had ever been to a research telescope before.) We walked into the telescope control room in the base of the telescope with our so-called right ascension and declination correct for the standard epoch 1950. However due to the precession of the Earth's spin axis due to the torque of the moon on the Earth, the coordinate system change year to year. The operator said, "We only take in our control system coordinates precessed to the current epoch," which at that point was 1970, not 1950 and there's this significant difference in the two. We went off, and had to figure out how to precess the coordinates forward in time. But there was another thing which would boggle the minds of current young astronomers -- the telescope is 140-feet in diameter built on an equatorial mount. Apparently when the telescope was built in about 1960, the engineers didn't trust computers to be able to calculate from right ascension and declination to altitude and azimuth, the horizon coordinate system -- a most trivial calculation so here is this very awkward telescope which probably cost twice what it needed to cost because people didn't trust computers to do a simple calculation back then.

ZIERLER: What were some of your basic curiosities at this point? What were you looking to find with your dissertation research?

SCOVILLE: That was just the first observing run we went on and, in the end, at the same time, the group at Bell Labs—Arno Penzias and Bob Wilson—they developed these diode amplifiers capable of going to higher frequency where you could observe with more sensitively. They were, I think, stimulated by my advisor, Phil Solomon. They developed the mixers to be able to observe the first carbon monoxide transition which was at 115 gigahertz at 2.6 millimeters. Carbon monoxide is much more plentiful than formaldehyde. At that point, it was obvious that if you really want to map out the interstellar medium in our galaxy or map a molecular cloud in which stars are forming, you really want to use carbon monoxide transition, not formaldehyde, because it's much more widespread and more detectable. I started observing with them plus Phil Solomon. In that case, we went to the telescope on Kitt Peak that was originally I think an 11- or 10-meter telescope—11-meter telescope—and it had been built by Frank Low, an astronomer, in Arizona to do high frequency radio continuum observations. We were using it for heterodyne spectroscopy. The carbon monoxide emission was so widespread in the galaxy, it was obvious that you could map out the galactic structure using it. In my case, I focused on two of the best-known giant molecular clouds, and then also a study of the galactic center region of the inner two degrees around our galactic center. I was probably the first one to do that with CO, and I proposed a model in which the gas clouds were located in a ring around the galactic center, and expanding outwards as perhaps driven by an explosion in the galactic nucleus. That was the main aspect of my thesis. I also was the first person to recognize that these clouds were held together by their own internal gravity, and the mass of material in the clouds, the so-called self-gravity. In my thesis, that was one of the things that I pointed out is that these were self-gravitating clouds. People before us had said that these clouds were probably galactic spiral arm features. They are now generally accepted to be self-gravitating clouds.

ZIERLER: Nick, how much modeling did you do for your graduate research?

SCOVILLE: My first real—well, I modeled that but then the first, you know, the expansion of this ring, but not in a detailed physical way; just describing the kinematics of the ring, and deriving the properties of that ring.

The first real theoretical research that I did was to explain one of the things that puzzled me in the observations. If you look at the carbon monoxide line profiles in the molecular clouds, you see a given velocity of distribution. The carbon monoxide should be excited into emission in relatively low-density molecular gas. But there were other observations of molecules like hydrogen cyanide (HCN), which have a larger dipole moment, and which require a greater rate of collisions with molecular hydrogen in order to bring raise their excitation above the cosmic background temperature. But in the case of the nearby Orion Molecular Cloud, the velocity distribution along the line of sight of the CO line was similar to the HCN and other molecules which typically have 10 times the dipole moment of carbon monoxide. I realized that's peculiar because the overall densities of these clouds can't be such that you would really require a density of 100,000 particles per cubic centimeter in order to excite the hydrogen cyanide significantly. The clouds can't have such high density since that would be inconsistent with the overall mass of the clouds. If the velocity distributions seen in the two emission lines—the CO and the HCN—are really similar, that implies that basically both come from the overall extent of the cloud.

The first real analysis I did was to recognize that there is a way to excite the hydrogen cyanide without having a high-density. When HCN is excited into a higher rotational state, it will spontaneously decay back down to be a ground state fairly rapidly. But, since the HCN has a high dipole moment, the photons which are emitted during that decay get absorbed by other hydrogen cyanide in a nearby gas. The effective decay rate is therefore not that given by the Einstein A co-efficient, spontaneous decay rate, but rather the decay rate reduced by a factor of the optical depth of the transition. The line optical depth is higher for those high-dipole moment molecules than for CO. This elucidates how it is that you can have the same observations which were puzzling to me but other astronomers hadn't recognized the conflict. They didn't really appreciate the fact that there's a contradiction here between simple understanding of the hydrogen cyanide and the CO. That was the first paper which I did on molecular excitations. In the case of that paper, after I did it, I went to an Aspen Institute for Physics summer sessions. These are typically three weeks or so. It turned out that Peter Goldreich was there, and he was interested in the same issue—or, at least, he was stimulated by me and Phil to be interested in the same issue. We actually came up with a solution to that problem, that is, the so-called photon trapping in these optically thick transitions, before he did. We actually got our paper out, beat Peter Goldreich's, which is a little bit of a feat. His paper came out the next year.

ZIERLER: Who was on your thesis committee at Columbia?

SCOVILLE: Pat Thaddeus, Phil Solomon, Woltjer, the department chair, and Leon Lucy.

ZIERLER: Besides the names, anything memorable from the defense?

SCOVILLE: You shouldn't be asking a 77-year-old person a question like that. My memory is not quite what it used to be.

ZIERLER: [laugh] Sometimes, I get the best stories from those questions.

SCOVILLE: I do remember there being a fair amount of interest on the part of the committee –I think especially on the model for the expanding ring of clouds in the Galactic center.

ZIERLER: What were the prospects after the defense? What did you want to do next? What was available?

SCOVILLE: I wanted to get a postdoc job in probably one of the standard radio astronomy places. This is a little bit amusing because I finished up graduate school, and applied for the NRAO, National Radio Astronomy Observatory postdocs, and then nothing. Caltech had two; CSIRO in Australia; and maybe one at Berkeley. I didn't get any of those. [laugh] Then Phil Solomon, who had at that point moved to the University of Minnesota—invited me to come out there and work for another year as a postdoc there, so I went there.

Then what happened was that Peter Goldreich came and gave a talk at the astronomy department there, and I asked a question which impressed him, and he said, "Why don't you come to Caltech?" I came to Caltech, but I didn't really want to do the lower frequency radio observations which was the capability up at Owens Valley Radio Observatory at that point. So, I worked doing observations still at Kitt Peak but then doing theory with Peter. I worked with him on the mass loss from giant stars, red giant stars, and we wrote a couple papers both on the excitation of hydroxyl masers in the mass-loss envelope, but then also the physics, the heating and cooling of the gas as it comes out in a wind from the stars. That was tremendously rewarding because as is the standard with Peter's projects, he sort of starts off in an unknown field, and then develops a thorough understanding of all the things related to it. He was a tremendous person to work with. He had a reputation of being extremely aggressive and competitive at that point, back in 1974 or '75 but I found him a real pleasure to work with -- never competitive, at least with me.

ZIERLER: Do you remember that question that impressed him so much?

SCOVILLE: No [laugh], I don't. He had been working on astrophysical masers, and I think it was related to some model of checking whether the masers were saturated or unsaturated. But I don't remember the question. I'm sure he doesn't either. But, at least, it was nice to hear that he appreciated the question.

ZIERLER: In what ways did collaborating with him launch you into new directions?

SCOVILLE: Well, certainly, the red giant stars, I did find them very interesting, and then also the maser excitation. I wrote one paper on that subsequently. Basically, it sort of triggered me in a kind of a mode of doing research, which is to try to develop a thorough understanding in tackling a problem. I have tried to insist with other people that they develop a complete, and at least, intuitive understanding of what they're doing.

ZIERLER: Nick, tell me about the opportunity at UMass Amherst. How did that come about?

SCOVILLE: That was actually quite amusing because I was at Caltech as a postdoc at the time. I was working with Peter and doing some radio observations. But then the University of Pittsburgh, where there was a Professor Artie Wolfe, who I'd gotten to be friends with, he offered me a faculty job at the University of Pittsburgh. However, I didn't really want to be at Pittsburgh, even though it's a reasonably nice place. It's kind of far away from everything else. I think people wouldn't be visiting you that much there. I found that also in Minnesota. People just don't stop in Minnesota on the way between the two coasts. The centers of scientific research are all along the East and West Coast. But, in any case, I told Artie to give me a deadline of a year, and he was very kind, and said, "Yes, OK," because he wanted me to come there. I had also inquired at UMass where they were building a 14-meter millimeter wave telescope, which was kind of ideal for my research, in the Boston Reservoir Island out in the Quabbin Reservoir.

At that point, I don't think I'd even visited, but I said I was very interested in coming. It was well-enough known that it wasn't like a voice out of the blue. I think it was two days before I had to let Artie Wolfe know my decision. I'm a straightforward person and if I had promised him I would say yes or no in a year, I certainly would honor that. In any case, two days before I had to tell Artie that I was coming or not coming, I got a telephone message from Ted Harrison, who was the cosmologist in the UMass astronomy department. He said, "Call me, your future depends on it." It was just a little note put in my mailbox in Robinson where I was located at the time. Obviously, I called him. He said, "Well, we have an amusing situation here, and that is we have a faculty opening but we've offered the job to somebody else, and we don't know whether he's going to accept it." At that point, the guy who was—his name was Al Chung; he was a postdoc at Berkeley. He had sent them two letters, and one accepting, and one not accepting. He said, "I'll call you and tell you which letter to open,". (Just a little flaky) In any case, he turned them down. So, then, I went and visited there, and they liked me, and I liked them, so they offered me the job right away.

ZIERLER: Was it a big program? Were you joining a group there?

SCOVILLE: Yeah, there were probably maybe eight or nine astronomy professors there. Now, it's a little bit larger too. But Ted Harrison was a theoretician but a really great character, very English but having a good sense of humor but, also, critical of observations. He considered a lot of what the observational astronomers do as no better than ‘collecting postage stamps' as opposed to developing understanding. I kind of sympathize with that a lot. A lot of the astronomy talks you go to are just showing a lot of data, and not really much understanding or new interpretation. You wonder why people are doing it. Anyway …

ZIERLER: Your time at Caltech when you were working with Goldreich, is that an outgrowth from your paper in 1976 on the Physical Properties of Circumstellar Envelopes, or that was a new project?

SCOVILLE: No, no, that was with Goldreich. Sorry, that was the first real project I did with Peter. In the other case, the excitation of molecular emission lines, that was done kind of competition with Peter, although, to be honest, I actually did all of the idea first. But he wrote a larger, more encompassing paper. But, in any case, the first one with Peter was on the physics of the envelopes of outflowing gas and dust from the red giant stars. That paper was entirely with Peter and nobody else. But then there was another paper in the series and notice that this one says number one. There was a paper on the excitation of the OH masers, which was written with Moshe Elitzur, who had been Peter's postdoc, and he was a very smart guy, but Peter got frustrated with him.

ZIERLER: [laugh]

SCOVILLE: Even though we did a lot of the work on that, Moshe was first author on the second paper.

ZIERLER: What does that mean, a circumstellar envelope?

SCOVILLE: These are giant stars. Imagine when the sun evolves more, it will swell up. Its radius will be, say, out to where the Earth is now, and the surface temperature of the sun at that point will have dropped to 2,000 to 3,000 degrees kelvin. Now, it's 6,500 degrees kelvin. At 1800 degrees in the outer atmosphere, some of the heavy trace elements will start to condense into dust grains can condense into dust grains -- primarily silicates or graphite. Those dust grains are extremely important physically because in contrast to the atoms and the molecules, which absorb radiation only in discrete quantum mechanical transitions, the dust absorb absorbs continuously across the whole spectrum. There's much more momentum from the radiation field imparted to the dust grains. The outward moving dust grains then drag the adjacent gas (along for the ride).

In summary, the red giant stars have an envelope of gas and dust surrounding it, which is expanding outwards. This is mass-loss material shed by a star. A typical star may shed a third of its mass in that phase of evolution, until it comes back to being a hot star. Now, once you remove this envelope, then you expose a hot inner core of the star, which then starts the photoionization of any surrounding material. At that point, the mass loss is cut off because it no longer has the dust screens, which can help absorb radiation pushing gas clouds.

ZIERLER: Nick, tell me about some of your administrative work for the Five College Radio Astronomy Observatory.

SCOVILLE: Well, let's see. I was certainly the kind of senior—even though I was young at that point—senior person, the most active person in terms of observational research with the 14m telescope. I didn't build any of the equipment. But, by doing research with the telescope, I certainly promoted the observatory to the point where it could get funding, although the initial funding was gotten before I got there. But the operational funding was certainly aided by myself. I've always been somebody who feels like I should also contribute to some of the dirty work of the operation. So, while I was there, I actually spent probably a third of my time writing

programs for both the analysis and measurement of the observations and also for the operating system in the telescope. I've always done a lot of that, and I did the same thing out at Owens Valley. There, I wrote an analysis package, which was probably a couple of years of work, while at the same time pursuing astronomical research projects.

ZIERLER: Did you have much interaction with Joe Taylor during these years?

SCOVILLE: I shared an office with him. But then he moved to Princeton after a few years of my coming there. Princeton obviously saw a Nobel Prizewinner and wanted to hire him. He was a wonderful person. At NRAO a few years back, I gave the honorary Jansky lecture in Charlottesville. He drove all the way down from Princeton to come to the lecture.

ZIERLER: Right. [laugh]

SCOVILLE: When he got his Nobel Prize, I think he spent a lot of money [laugh] on a ham radio tower and presumably some equipment. The Nobel Prize tends to destroy people's research career given the need to give many talks, but obviously it doesn't have to be that way.

ZIERLER: Nick, tell me about the circumstances leading to you joining the faculty at Caltech.

SCOVILLE: I had been at UMass for probably about nine years; came to Caltech on sabbatical. I spent half my sabbatical at Caltech, and half at the University of Hawaii because I like Hawaii so much. Everybody has got a head on their shoulders. When I was at Caltech, they had at that point built three of the Leighton dishes the 10-meter telescopes up at Owens Valley on the valley floor with a track to move them on. Tom Phillips had started developing receivers for those telescopes. But, at that point, they hadn't really done any—in my opinion—useful science with the telescopes. It always seemed like a tragedy to me that here is this super high-resolution instrument, maybe 10 times the resolution of the telescope which I was using in UMass.

ZIERLER: [laugh]

SCOVILLE: Then when I came there on sabbatical—I've forgotten who initiated it but it was probably Peter Goldreich initiated trying to get me to come to Caltech. But, at the same time, they had a very nice young astronomer, Fred Lo, who was on the faculty as a radio astronomy professor in the same field as myself. He unfortunately did not get tenure, although he became the NRAO director after he left Caltech. Well, I think he went to Taiwan first or Hong Kong—Taiwan first. But then he became NRAO director. Basically, there was an opening to hire somebody, and I had sort of saw that it would be tragic not to have some really good science coming out of the instrument.

ZIERLER: Was the presumption when you joined Caltech that, soon enough, you would assume directorship of Owens Valley?

SCOVILLE: I think so. I don't know what the presumption was. It's funny. When I came to Caltech, I said to myself, I'll only stay there for five years, and then I'll go to the University of Hawaii for five years, and then quite doing astronomy—

ZIERLER: [laugh]

SCOVILLE: —and do sculpture or welding or something because I think it's worthwhile to do more than one thing in life.


SCOVILLE: I've shifted around in astronomy to different subjects and/or sub-disciplines. I've done extensive infrared, millimeter-wavelength single dish and interferometry and visible wavelength observations. In theory, I've worked on spectral line formation of mm emission lines, infrared radiative transfer, molecular cloud structure, evolution of starburst galaxies to AGN (central black holes), models for the AGN emission and absorption lines, maser emission, circumstellar envelopes, and evolution of galaxies in the early universe. I like the idea of exposing oneself to new areas of astrophysics (but then in life) so that you learn anew.

I did not leave Caltech after the 5 years I mentioned earlier; partially because one develops loyalty to the people you've worked with and partly because Caltech and Los Angeles are fantastic. Once Anneila took over directing the OVRO mm-array, it was obvious that she was going to do a great job and I could have left. She was, in fact, better with the politics of joining up the Berkeley array than I would've ever been. The other thing is, of course, I actually like Los Angeles; there is tremendously diverse Asian and S. American culture in Los Angeles, which I really like. At this point in my life, I don't really take advantage of it as much as I used to, but that the way life goes.

ZIERLER: What was the state of play at Owens Valley when you assumed leadership there? What were some of the big issues going on?

SCOVILLE: Well, I think they actually had a fairly bad review—NSF review. Peter, I think, stimulated by Robbie Vogt, was instrumental in developing the first good proposal for operation the Owens Valley Millimeter Array and I came two years after that proposal was funded.

ZIERLER: What were the administrative responsibilities as a faculty member? Were you not teaching during these years?

SCOVILLE: I was teaching the normal load, but at Caltech that is very light compared to most other universities. I'm not somebody who builds equipment, and I respect people who do. Even though I'm running a highly technical operation, in the case of the Owens Valley Millimeter Array, I usually set the direction by asking the more technical staff associated with the instrumentation what they would like to work on. I'd ask them what the next practical development is. I would usually respect input. I wouldn't say, "I have this brilliant idea, you guys do this." I don't like that because these are PhD people who have great technical ability and expertise, and also a sense about the field. They may not want to write astronomical papers, and they may not have the astronomical background. But, still, I respect what they want to do because, quite frankly, somebody working on what they want to do, and taking responsibility for it, works much better than somebody just given orders from the top.

On the other hand, even though you respect their judgment, you do need to keep an eye on the development, that is, that the staff stick to an approximate schedule, and, if the development is delayed, there is a good reason for it. I've seen astronomical colleagues get a pile of funding for a project, and then walk away from it and say, "Well, I've done it. That's all that needs to be done." Two years later, they get re-involved and discover, not surprisingly, their project is way behind schedule. This could have all been avoided by diligent direction starting from day one of the project. One example of this is the SOFIA Airborne Observatory. USRA obtained funding for modifications of a used Boeing 747 and installation of a telescope, etc. Unfortunately (from my point of view) they relaxed too much in the first few years and it got way behind schedule. The same thing has happened with a lot of projects—probably this was also involved in some of the enormous cost and schedule over-runs on the JWST project. But, they had other more severe problems also. Thus, I am advocating that it is critical to keep an eye on schedules and projects from the ‘get-go', not in a bureaucratic way but just making sure things are proceeding.

ZIERLER: Nick, tell me about your work with Tom Soifer and others on the origin of quasars.

SCOVILLE: Really, the lead on that was Dave Sanders, who was a postdoc here, as I have said earlier, now is a faculty person at University of Hawaii. He came funded by the Caltech infrared group (Neugebauer and Soifer). He had been a postdoc at UMass while I was there and he was a Phd student of my thesis advisor Phil Solomon. I think I probably suggested that he should do follow-on observations on these high-luminosity, ultra-luminous infrared galaxies discovered by Soifer and Neugebauer in the IRAS survey. Obviously, he would talk to Gerry and Tom, but I think it was really mostly Dave who initiated most of the follow-on.

Sanders' point was that the equivalent number density of these ultra-luminous infrared galaxies and quasars in the local universe, but the weak link in the hypothesis is that one doesn't know the relevant timescales for each phenomenon-logical phase -- that is the timescales for a quasar being a quasar and for the ultra-luminous infrared galaxies aren't really well determined. Thus even if one sees an equivalent number of the two types of galaxies in the local universe, that doesn't mean that they really evolve from one to another, unless the timescales are also similar.

ZIERLER: [laugh] Now, throughout the 1990s, Arp 220 occupies a significant amount of your time. What was so fascinating about it? What commanded so much need for research?

SCOVILLE: It's the most extreme of the ULIRG objects, it is also one of the nearest and, therefore, easiest to observe and obtain get high-quality data on it. As a member of the NICMOS camera team, which was put on the Hubble Telescope back in the mid-90s, imaging all of the 24 ULIRGS was my main project as a member of the NICMOS science team. With the NICMOS camera, we imaged all of the Northern Hemisphere ULIRGs at two microns. With these high-resolution images, we saw immediately that most had double nuclei -- that verified immediately that these were binary galaxies in the final phase of galactic merging. A similar thing was seen in imaging of the molecular gas we were doing at Owens Valley with the Millimeter Interferometer. There we did follow-on observations of the carbon monoxide line in the Arp 220, and it beautifully showed a double nucleus. From the CO line kinematics, we found that the two nuclei are actually rotating in opposite directions – this is optimum for the rapid merging of two rotating galaxies since the rotational angular momentum can then be cancelled out.

ZIERLER: Nick, how did you get involved in developing software for the OVRO Millimeter Array?

SCOVILLE: Well, I was the director so I could whatever I want.

ZIERLER: Was it a new venture for you? Had you done any software development before?

SCOVILLE: Yes, for the 14-meter UMass telescope, I developed a lot of the analysis software for the telescope and supervised the telescope control software. I'd done a lot of programming and I kind of enjoy building a house of software which hopefully doesn't collapse but which most important save a huge amount of tedious work each observer would have to do if it didn't exist. I hate doing ‘grunge work' when you end up with a huge amount of data.

For example, when I was at UMass, we did a CO line survey of the interior of our Galaxy. The survey had 60,000 observations, that is, 60,000 pointings with the 14-meter telescope, along the galactic disc, and then moving up in latitude along parallel to the galactic disc and covering plus and minus one degree. There were 60,000 observations, and that's not the kind of dataset you can analyze on your laptop in those days -- now you can. It was absolutely critical to develop software which not only removed impurities in the observations but then also cross-correlated the observations at different pointings.

ZIERLER: Nick, an administrative question, when you were executive officer from 2000 to 2004, how much of that was actual administrative authority, and how much of it was simply a title, and it was really the division chair who did those kinds of things?

SCOVILLE: The division chair did important things, like appointing directors of the different observatories and keeping an eye that their budgets were reasonable. On the other hand, the EO is responsible for teaching assignments among the faculty and the day-to-day running of the department. You don't know what our department was like, but it used to be there were a bunch of complainers who were always miserable, particularly in the astronomy department. I don't think it was true in other Caltech departments. But astronomy had a bad reputation with people not only being arrogant but complaining that they didn't have enough financial support for their observing time on the telescopes. When I started as executive officer, I did a little bit of psychological counseling. At our first faculty lunch, I remember starting with: "I'm happy to be the executive officer, and you should all be happy to be in this department. This is the best astronomy department in the country and probably also the world. You have the best financial support, guaranteed time on large telescopes, and the best postdocs and graduate students. If you're feel miserable, you have only yourself to blame." At least for a few months, I didn't get any more complaining!

ZIERLER: Ooh, how was that received?

SCOVILLE: I think people were amused and sort of quiet. [laugh] The situation here was really quite ridiculous. Astronomy departments at other schools and their resources are much, much less. It's not that people at Caltech are that much better, quite frankly. [laugh] They have historical support which has been built up and facilities here which are sitting -- waiting to be used.

ZIERLER: Nick, that's a great place to pick up for next time.

[End of Recording]

ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Thursday, February 17th, 2022. It is great to be back with Professor Nicholas Scoville. Nick, it's good to be with you again.

SCOVILLE: Yes, and thank you for the previous interview. It was enjoyable!

ZIERLER: I want to go back to some of the additional research from your time at Owens Valley. What was your main line of work at that time?

SCOVILLE: The ultra-luminous infrared galaxies were discovered by the IRAS satellite, which Gerry Neugebauer, professor in physics, was PI on. They discovered these objects, which had huge inferred luminosities in the infrared but only maybe 1% as much luminosity coming out in the optical. They were clearly very interesting dust-enshrouded objects. At the time, one of the infrared astronomers in Arizona even suggested that maybe they were a fundamentally new class of object for which he coined the name IRtron. The angular resolution of IR imaging which had been obtained early-on for these objects was only of the order of an arcminute. The aspect I wanted to pursue was where the gas and dust were in the galaxies relative to what you see in the optical and infrared. We had the ideal instrument to do that with -- the Owens Valley Millimeter Array, which originally consisted of three telescopes—three 10-meter Leighton dishes. While I was director, we obtained funding from the NSF and the Norris Foundation for an additional three telescopes. Thus, we had a six-telescope array, which gives you 15 simultaneous baselines, and fairly quick imaging compared to what we had had before with only 3 baselines.

ZIERLER: That's a real abundance of riches right there.

SCOVILLE: Now, the best millimeter array is now the international ALMA array in Chile which was a follow-on to the smaller arrays at Caltech, Berkeley and in Europe. ALMA has 64 telescopes, and when they select your observation proposal, they guarantee 45 operating at the same time for each observation. That's roughly 1,500 baselines to be observed. In addition, the receiver improvements and site at 16,500 ft elevation provide an order of magnitude improvement in sensitivity.

In any case, the project I set out to do at OVRO was to locate where the dust and gas were in the ULIRG galaxies. We already knew that there was a huge amount based upon our single-dish. What I proposed was to get the resolution down from one arcminute (for the our previous single dish observations) to sub-arcsecond using our interferometer. Doing this for the nearby ULIRG Arp 220, what we discovered that there were two galactic nuclei in the galaxy only 1 arcsec apart. Both had gas and dust covering the nuclei. The velocity field of the gas indicated that they were counter-rotating disks of gas. The two galaxies hadn't fully merged but they were only 300 parsecs apart. In the case of the stars associated with the original two galaxies, they were distrusted over a much larger volume. This because the stellar motions don't dissipate kinetic energy and angular momentum nearly as effectively as the gas. Ultimately, I followed this up with observations using the ALMA array once it was operational.

ZIERLER: What year was that, Nick? When did ALMA go online?

SCOVILLE: It came into operation in 2010, and one of the first proposals I submitted, which I think was done in 2011 or so, was to do Arp 220 at high resolution, at roughly 0.1-0.2 arcseconds, five times better than what I was able to get with the Owens Valley array. There, I actually found that not only was the gas centered on the two nuclei, but it was incredibly concentrated – a few billion solar masses of gas with a diameter of just 30 pc.

ZIERLER: Which tells you? What does that tell you, its concentration?

SCOVILLE: Well, the concentration tells you what the gas density is, and also what the opacity to any radiation originating from the two nuclei. Just to give some numbers, in Arp 220, that one nucleus had a dust concentration around the nucleus, which was about three billion solar masses, which is equivalent to taking all the gas in the Milky Way (spread over a diameter of 20 kpc and squeezing it into 30 parsecs. In the Milky Way, most of that gas is out at 5 to 10 kiloparsecs from the sun inward. The fact that you've concentrated the gas so much means that, one, obviously, there'll be a high opacity, and you can't see in either the visible or even the infrared into the center of each nucleus. But the rate of star formation will be that much higher because gravitational collapse is much easier to initiate when you start with gas in the nuclei at a density of a million molecules per cubic centimeter compared to standard giant molecular clouds in our galaxy, which are maybe 300 molecules per cubic centimeter. That was extremely exciting work, and it was a pleasure to take part in. The initial observations were actually done before the paper linking quasars with ultra-luminous infrared colliding galaxies.

ZIERLER: Nick, I'm curious, once ALMA came online, to what extent did it render Owens Valley observational capacities obsolete, or were they more complementary at that point?

SCOVILLE: Well, they could've been a little bit more complementary, but Owens Valley relied upon NSF funding for a major part of the operational funding in for the CARMA array. Once the NSF had put something like $150-200 million into the ALMA array, they basically backed off supporting the university arrays. It was kind of a foregone conclusion that they wouldn't be supporting the university arrays in the future. On the other hand, there were some projects such as large-scale mapping of nearby molecular clouds, which were very successfully done with the CARMA array and those results have been published.

ZIERLER: Nick, I'm curious, what about the distinctions between northern hemisphere versus southern hemisphere observation? What are some of the challenges and opportunities there?

SCOVILLE: Well, one challenge is that there are unique telescopes in the two hemispheres. I think I mentioned last time in our interview that that's particular the case at radio wavelengths where you have the EVLA, which is in New Mexico, which does centimeter-wave radio observations. Then in the southern hemisphere, you have ALMA, which does higher frequency (mm-submm wavelengths) observations. Basically, the latitudes of those two observatories are plus and minus 25 degrees from the equator. With ALMA, it is difficult to do objects in the north by more than +20 degrees Declination. Likewise, something that's more south than 20-30 degrees can't be done well at the EVLA. On the other hand, there's very complementary data from those two telescopes.

This gets into my planning for the COSMOS or Cosmic Evolution Survey I was PI for with HST. We specifically selected a new survey area near the celestial equator so that it would be observable by all large or unique telescopes in both hemispheres, This has worked out incredibly successfully because not only do you have these two radio telescope arrays—the EVLA and ALMA—but then there are large telescopes in the southern hemisphere at the ESO Observatory, having four eight-meter optical/IR telescopes in addition to the ALMA array in Northern Chile. In the northern hemisphere, there are the twin 10-meter telescopes at the Keck Observatory. The COSMOS survey field is only two degrees off the celestial equator, and that has enabled us to collect data from not only unique telescopes but also large telescopes in both hemispheres from which there's a limited amount of time.

ZIERLER: What is celestial equator? What's the difference between that and a terrestrial equator?

SCOVILLE: It's the line in the sky directly above the Earth's equator. You probably know that the Earth's rotation axis is tilted by 23½ degrees to the axis of the ecliptic plane (i.e., the orbital plane of the Earth and the planets around the sun). In any case, I, being particularly interested in the millimeter and submillimeter imaging of objects, insisted that we should have a field that should be observable by ALMA because, obviously, that was going to be the next big and unique future instrument. But then, also, we've accumulated very good data from the EVLA. And a lot of our optical and near-infrared imaging has been obtained at both ESO and at Keck and other telescopes in Hawaii. The strategy has worked out extremely well. I should tell you how COSMOS actually got started.

ZIERLER: Yeah. Just to orient us chronologically, when did those earliest discussions start that ultimately led to COSMOS?

SCOVILLE: The first planning was in 2002, and it was initiated by a kind of fluke on the part of people in the community complaining. Let me explain. The Hubble telescope has a call for proposals once a year. In the previous proposal cycle, it turned out that several of the large projects which were selected to have observing time, for example the so-called GOODS survey— were all PI'ed or had heavy collaboration from people at Space Telescope Institute. There was a complaint on the part of astronomers at the universities and other facilities that somehow there was a bias in the selection or the capability to write proposals. I felt a little bit embarrassed by people complaining, and there were colleagues in Caltech astronomy who did some of the omplaining—not myself. The director of Space Telescope Institute at that point was Steve Beckwith—he had been a graduate student in IR astronomy at Caltech, and then became professor at Cornell, and ultimately director at Space Telescope Institute in Baltimore. To defuse the complaining, he decided to host a workshop to decide what big projects to do in the next HST proposal cycle. These big projects are typically 100 orbits or so of HST time, as opposed to typical small projects which might be a few orbits up to 10 or 20.

He asked me to chair the sub-panel at that workshop, which dealt with cosmic evolution and cosmology. There were maybe 30 people invited to the workshop, who were external galaxy scientists doing surveys of galaxies at high-redshift. I was picked by him to lead this panel, maybe because Steve thought that I would be impartial or maybe had a good physical sense about what was important to do. I'm not quite sure. When I looked at what had been done, compared with the predictions of cosmological simulations, it was glaringly obvious to me that the previous fields—such as the Hubble Deep Field, and then, subsequently, there was one called the Hubble Ultra Deep Field—that these fields of the order of a few arcminutes in size were way too small to map out the large-scale structure at high-redshift in galaxies. At this point, in 2002-2003, it was believed that a lot of galaxy evolution actually determined by their so-called environment, that is, how many galaxies are nearby? Is the galaxy located at the core of a large-scale structure or in a filament or in the intervening space between filaments? If you looked at the size of these previous fields, the scale of the large-scale structure might be 50 megaparsecs or so, and the scale of these fields were perhaps 10 Mpc. We put a figure showing this in the proposal and I think that was one of the prime reasons that the COSMOS survey proposal was selected.

ZIERLER: What were the conceptions of who would be collaborators at the international level of COSMOS?

SCOVILLE: At that workshop, I said anybody who wants to be a collaborator on this proposal, which I will write, I would welcome them to join in, and that included a lot of Europeans who were either attending the workshop, such as Alvio Renzini and Olivier Le Fèvre or others they could suggest. We had tremendous international representation on the project. But backing up just a second, there were other people on that workshop panel who wanted to write their own proposal, and one of them was Sandy Faber from UC Santa Cruz. She was kind of incensed that here was somebody, who had done mostly radio and infrared astronomy, was submitting a proposal for large amount of Hubble time. She submitted a competing proposal. Ultimately, the TAC selected our proposal. Hers was to do a strip in the northern hemisphere, which she'd been focusing in her previous studies. But it was a narrow strip, and wouldn't really get around the issue of how do you map the large-scale structure if you're only looking at a small area of the large-scale structure, in addition to not being observable from ALMA. We actually got selected, and she unbecomingly actually protested to Beckwith, who rejected her complaints. Certainly this was unpleasant since one is supposed to respect the normal proposal evaluation process. Ultimately, she did get time a year or two later to do some of the work that she had originally wanted to do. I think, in the end, she ended up happy, but she didn't like being ‘beaten'. The previous largest proposal, the GOODS survey project, was ~150 orbits or so. It was several pointings, but then in several bands going deep with multiple orbit exposures on each of the pointings. The COSMOS proposal was awarded 600 orbits -- 3-4 times as large as any previous proposal done on HST. With HST there is roughly 2000-3000 orbits per year, and so it's like 10% of the time in each year.

ZIERLER: Nick, what were some of the main research questions that prompted the creation of COSMOS? What was not possible without this new collaboration?

SCOVILLE: The main one, as I said, was the correlation between cosmic large-scale structure and the properties of galaxies. You might expect that in an area of high density large-scale structure, for example in the core regions which will ultimately form galaxy clusters, the galaxies will evolve more rapidly and hence be seen to be more massive at high redshift. In addition, that may be the place where collisions of galaxies leads to the morphological transformation from spiral galaxies with discs and lots of gas to elliptical galaxies which basically are a spherical ball of stars with no remaining gas and an ageing stellar population because they no longer have a significant mass of star-forming gas. You expect that kind of correlation. But it hadn't really been observed, except in a few local clusters where you would see ellipticals in high-density regions, but not at cosmic distances, redshift 1, 2, 3, at the peak of star formation. That was the prime objective. The first exciting project which was really done early was that of detecting gravitational lensing and amplification by foreground galaxies of distant galaxies at high-redshift. This was a project which was led by Jason Rhodes, and Richard Massey, and Richard Ellis and myself were involved in it. This became a cover article for Nature partially because it was the first detection of dark matter large scale structure, correlated with the projected galaxy distribution. That worked out beautifully.

One of the projects which I focused on in COSMOS was mapping the large-scale structure. In the COSMOS survey field, we were doing two square degrees or roughly one and a half by one and a half degrees on the sky, and it's equivalent to something like the area of nine full moons stacked next to each other. It's a huge area on the sky. In order to map large scale structure in the galaxy distribution, you need the redshifts of each galaxy, which infers where that galaxy lies in the overall cosmic expansion which is producing the redshift. At redshift 2, all the light from the galaxies shift about a factor of three to longer wavelengths by the cosmological expansion because the galaxies are separating at a significant fraction of the speed of light. The Doppler shifted radiation from the receding galaxies is seen by us redshift from its original restframe emission wavelength. In the COSMOS field, which as I said was two square degrees across, the imaging that we got with Hubble detects roughly 900,000 galaxies. There's no way that you count on getting spectroscopic redshifts where you typically do 1 at a time, or 10 or 20 galaxies at a time, to cover the whole field, and thereby map out the large-scale structure. If you don't have redshifts, what you see is all these galaxies on the sky—900,000 of them—but you can't tell whether two galaxies projected on the image are close to each other along the line of sight without knowing the redshift, and knowing that they're at similar distances.

One of the techniques which I pushed in COSMOS was to use so-called photometric redshifts instead of relying on spectroscopic redshifts. There were a lot of people in the established high redshift community who said you needed spectroscopic redshifts in order to place the individual galaxies in the large scale structure. For the photometric redshifts, you obtain deep imaging of the field in broadband continuum filters and then fit the distribution of the continuum fluxes to standard galaxy spectral energy distributions to determine the redshift as a free parameter for each galaxy. The Japanese Subaru telescope on Mauna Kea actually built an imaging camera (HSCAM) with field of view matched to the COSMOS field of view (1.5 degrees). Our Japanese and Hawaii collaborators then used that camera—called Hyper Suprime-Cam—to do deep multiband imaging, in the COSMOS field. In the red filter, you see the light which is coming to us at red wavelengths; blue filter, blue wavelengths, and so on, all the way down to the infrared at roughly one micron. There were also a number of other large telescopes (primarily at ESO in chile and on Mauna Kea which also contributed; there are now ~41 photometric bands of data on the COSMOS field – an unprecedented dataset for the characterization of the 900,000 galaxies.

Then there was another survey which was started about 10 years after COSMOS called the UltraVISTA survey at ESO, in the near-infrared, that is, out to two microns wavelengths or four times the visible wavelength. That survey had as a major goal making a near-infrared image of the COSMOS field, which is probably the deepest near-infrared map of any large field in the sky. In COSMOS, we now have 41 bands from ultraviolet down to near-infrared. With that number of bands, you basically get spectra which have a resolution of 40 or 50 to 1. We now have 1% or better accuracy on roughly 600,000 galaxies. To place galaxies in the large-scale structure required only ~5% accuracy. I was able to use those photometric redshifts to map out the large-scale structure in 2009 and 2013 over the whole COSMOS survey field, finding ~100 structures at different redshifts, extending typically over half the area of the field.

Then I looked at the correlation between galaxy properties and their location of the large-scale structure. At low redshift, I found that in the dense areas of the large-scale structure at low redshift, the galaxies were very largely elliptical-type galaxies with red stellar populations, meaning older stars. But when you went out to redshift 2 or 3 at the peak of cosmic star formation, there there was very little spatial segregation in the galaxy types. The implication is that it takes a while for the environmental influences to play a role in the galaxy transformations, say, from spiral to elliptical galaxies. The transition between these two regimes around redshift 1.2 or so.

A more recent study I've been focused on has been to ALMA to measure the gas content of the galaxies, at redshift 1.2 to determine when the gas in galaxies dominates the masses of stars. At our present epoch, the gas content is typically only 5-10% of the stellar mass and the rates of star formation are greatly reduced. That sort of goes along with the picture I was painting of the galaxy transformations. which is that once you've started to use up most of the gas content in the galaxies, they no longer form young stars, and a galaxy can no longer form a spiral galaxy disc which requires dissipative gas to reduce the motions perpendicular to the disc. The Milky Way has escaped the transformation to an elliptical type galaxy simply because it's in a low-density cluster, including the Magellanic Clouds, M31 and M33. The Milky Way hasn't undergone a near galactic encounter YET. But that's projected to happen in the future since M31 is on a collision path towards our Galaxy.

ZIERLER: Nick, what were some of the key funding sources that made COSMOS possible?

SCOVILLE: This gets into the difference between NASA and NSF funded science. In the case of Hubble telescope projects, and with JWST—James Webb Space Telescope—if you get observing time, you also get funding for postdocs or whatever you need. There's a so-called Hubble constant ($ per orbit of granted observing time, typically ~$10k per orbit). (Nobody yet knows what the JWST constant is.) In the case of COSMOS, we had 600 orbits and obtained $2 million of funding, from NASA as funneled through the Space Telescope Science Institute. That gives you a fair amount of discretion. I distributed some of this around to key members of the team around. However, I've seen many scientists get into trouble with their funding, primarily because they over-extended their hiring of postdocs, supporting people, and travel. I didn't, say, take all the US Co-I's, and equally divide the funding. I decided where some of the work could really be done, and divided it accordingly, and then kept a fair amount in reserve in order to help people out when they ran out of money.

ZIERLER: Nick, administratively, how much time did COSMOS take out of your schedule?

SCOVILLE: For administering the COSMOS project, very little time. However, my research over that period of ~15 years was largely focused on the COSMOS field. The ALMA project I just described involve the actual Hubble observations, but it is in the COSMOS field, and so it couldn't have been done without having the COSMOS survey for ancillary redshifts and characterization of the individual galaxies. Whereas most surveys, like these ultra-deep fields—the Hubble Ultra Deep Field and the Hubble Deep Field—they may have 500 to 1,000 galaxies in the field, so that gives you limited statistics to do much with. Once you've accumulated the 41 bands of photometry on every galaxy in the field and their redshifts, you have tremendous ability to make use of all other telescopes. For example to generate a large sample of bright X-ray emission sources which might be active galactic nuclei supermassive or black holes. That kind of project has been done by a lot of the COSMOS Co-I's and external astronomers. All of our data is publicly accessible.

Back to the admin question you asked -- Our initial HST COSMOS proposal had 38 Co-I's, but over the last 20 years of COSMOS, we've had typically 100 people actively participating in the project, and coming to team meetings (once a year, always in a beautiful and interesting place). I ran the project as an open resource for the broader community and amazingly, this avoided my (or my teammates) getting stressed out or anxious about proprietary data rights.

A lot of projects, e.g., the Sloan survey, have rigid rules about who can publish what, and what procedures they have to go through. In COSMOS, I decided that we won't have such a bureaucratic nightmare, but we'll have a website where anybody who wants to do a scientific project can propose that they want to do a project, and anybody else in the collaboration can say, "Please add me to that. I'll help you with such and such." That's worked out incredibly well. Also, a lot of the people who contributed critical data sets like the X-ray imaging—e.g. Günther Hasinger and Martin Elvis were not original investigators on the project—but they were happy to contribute their X-ray images on the COSMOS field to make them available to everybody else because they knew once they became a part of COSMOS, they would have immediate access to all the backup photometry, the redshift and the galaxy stellar masses derived by others. There's a huge of amount of value added to being in a survey where the data is semi-public. I say "semi-public" because when somebody contributes a data set like infrared observations of galaxies in COSMOS, they have sole access to the data for the first year after the data has been obtained. After that time, they all know that this data will be available to other members of the COSMOS survey and the public, so it motivates rapid development of the science. A lot of our data and catalogs are released through the IPAC IRSA database, and then there's a similar database in Europe at ESO and at Marseille.

ZIERLER: Nick, in what ways has the COSMOS collaboration grown over the years?

SCOVILLE: Well, tremendously. Of those 38 original Co-I's on the project, I would say only maybe 20 are now still active using COSMOS because they've aged, or they've gone out to do other projects. But, as I said, we have typically 100 people attending our team COSMOS meetings. The new people contribute datasets in some cases and who also stimulate new research projects with the existing catalogs. The general age of the people within COSMOS has probably shifted downwards and present two leads are woman.

I did not micromanage what people did—although I'd listen to what they wanted to do, and then suggest improvements. However, after I think it was nine years or ten years of being the leader of COSMOS, I want to get somebody else to do this who would bring in fresh energy. I turned the leadership over to Peter Capak. He had come to Caltech as a postdoc in COSMOS and then became a staff member at IPAC in the Spitzer science group. He is a treasure trove of information and remembers every detail of the COSMOS (and other surveys). I don't know how he does it -- he certainly has a different type of mind than mine. He was really great. But then at some point about four years ago, he decided that he wanted to go into industry, and he's now working at Facebook on virtual reality. In turn after 10 years, he turned the leadership over to Jeyhan Kartaltepe (RIT), and Caitlin Casey(Texas).

ZIERLER: Nick, just to bring the conversation on COSMOS to the present, what are some of the key projects going on now?

SCOVILLE: Oh, god, I hadn't anticipated that. [laugh]

ZIERLER: The answer's "a lot" then? [laugh]

SCOVILLE: Well, this project which I related using ALMA is one of the more important projects. Even though I'm making use of data from other people, and they will be coauthors on the paper, most of the work has been done by me in my COVID retirement. [laugh] I'll try to find out what some of the more active projects are. The other thing which is happening is that people like Dave Sanders—the same one who worked on the ULIRGs— has now extended the COSMOS concept to a field which branches off of the COSMOS field, which is 20 degrees across the sky. That's partially because he has access to the telescopes in Hawaii, and also to his collaborators. One of the projects which he is doing with Meg Urry (Yale) is to study AGN properties across their vast field. You need a very large sample for the AGN study and that's the reason why they have initiated such a large survey field, but they'll never have the same quality imaging which we do in COSMOS where there were multiple nights devoted to each filter obtained on the field.

ZIERLER: Clearly, though, the answer is that there's still a lot going on.

SCOVILLE: Well, part of the reason why I have difficulty answering the question is for the last two years, we have not had a team meeting, and the reason being obviously COVID.


SCOVILLE: At team meetings, usually lasting three or four days, people—even myself—are limited to like five-minute talks so that one succinctly gets to the point and without a lot of obvious background information. That allows us to have talks from all the young people who wouldn't normally be so visible in one of these other survey teams. I think that has also served to make the young people appreciated and more visible in the COSMOS survey. But the lack of team meetings has meant that I and probably a lot of other people don't really know what people are doing at this point.

ZIERLER: Well, Nick, now that we've brought the conversation right up to the present, for the last part of our talk, I'd like to ask a few broadly retrospective questions about your career, and then we'll end looking to the future. First, at a very high level, if you survey all of the research you've done, all the collaborations you've been a part of, what have been some of the key contributions to the bigger questions about how the universe works, its origins, and its structure?

SCOVILLE: One of the things I pride myself on is the fact that I've spanned a large number of different areas in survey and observational techniques, and then also I haven't been reticent about doing some theoretical modeling. I love that mix of things. I am happy to have spread myself out rather than focusing in on one well-defined, small area of science.

In each of the areas I've worked in, I've learned a lot which often carries over into later projects. In the case of COSMOS, the main area I've worked on is the connection between the large-scale structure and galaxy properties. But, nowadays, other people have done this also. One of the things I do find a bit annoying is that even when something is done well or adequately, there tend to be people who will come in and repeat it and put a more recent date on it. Then unfortunately, later young people usually refer to the latest paper rather than the one which basically had the initial results. I've never really complained about this to anybody other than you.

ZIERLER: [laugh]

SCOVILLE: —because it would be kind of in bad taste, but that annoys me. I can see myself also when I'm writing a paper, you do go first to the latest Review article or the latest authoritative article, and then kind of work backwards from there. But, given the limited time and the large number of articles, unfortunately, a thorough understanding and appreciation of the development in each area gets lost. Certainly, I am also deficient on this when I write a paper in a new area, although I do try to find the original reference, and then a few of the more later references. Another complaint I might have about how science is done now, compared to early in my career is the narrower focus and specialization of young people. I said that I was particularly proud of the fact or happy with the fact that I didn't have to force myself, but I enjoyed going and working in new areas where I wasn't necessarily familiar. Certainly, the COSMOS field is an example of that. I wasn't doing very high redshift galaxy evolution studies before initiating the COSMOS project, but you meet a whole new cast of characters, and it's been fantastic.

ZIERLER: Nick, on that point, I'm curious, in the way that you have chosen breadth over depth in your research focus, and in contrasting that with what you might see as an over-specialization of younger people in the field, to what extent is that about the way funding sources work nowadays, or how astronomy has changed culturally over the years? What do you ascribe to this trend towards over-specialization?

SCOVILLE: I don't think it's a funding issue. Young people, when they get out of graduate school, they need to write papers -- nowadays, a large number of papers. It's obviously easier to write papers in an area in which you already have prior knowledge and background. But, at some point, when they become a senior postdoc or an assistant professor, I would think that they would want to branch out. I never experienced it but, nowadays, you hear people feeling good about the fact that they got tenure. I actually think tenure is a bad and un-necessary thing -- astronomers don't need to be protected from political issues. I think that was the main reason for tenure. Tenure can only lead to people eventually getting too relaxed, and not actively doing new research. I can see that not so much at Caltech but at a lot of universities where the people above age 50, 55, end up spending a lifetime only teaching. Good teaching is undoubtedly needed but it should be the only challenge given our low teaching loads. At Caltech, we obviously have an advantage of fantastic facilities, on the other hand, most astronomical facilities do have public time, and there also are national facilities where you can get time.

ZIERLER: Nick, in the decisions you've made to move from subfield to subfield, intellectually, what have been some of the assets of being a newcomer at a certain stage in your career, and having a different perspective than somebody who might've been working in that area for decades?

SCOVILLE: I think a general response is, simply, you come in with a more naïve/open mind, and not dictated by what's been done in that subfield before. I think astronomy and astrophysics is not really difficult or new physics; so there's really no excuse for others coming in with fresh ideas that you want to test, and then using your scientific background to fully understand what's going on.

ZIERLER: You're saying once you're at a certain level, there's a low bar of entry into moving into other subfields?

SCOVILLE: Yeah. The funny thing was that starting out in radio astronomy, I always used to look up to these optical astronomers (many of them at Caltech), who I had respected, and then think that they do high-class,, precise astronomy. But when I started doing optical observations myself, and working with some of these people, I found that they were really no better [laugh] than those in my old subfield of radio astronomy. They used similar approximations in what they did, and had similar limitations in their techniques. Radio astronomers used computers extensively before digitized optical astronomy; however so in some sense their data sets were better calibrated. For a modern data sets, you really need to be proficient in computer usage, or have a young student or postdoc who's going to do the work for you. As I said, I sort of appreciate doing everything myself as much as can be tolerated. At the same time I have greatly enjoyed supporting and working with the younger scientists. It's just that I enjoy the process of figuring stuff out for myself.

In astronomy, there are two major fields -- studying galaxy evolution and star formation, and those who study the central black holes of galaxies, the active galactic nuclei. Peter Goldreich, one of my mentors, once asked me, "Well, what's new in the AGN field?" Thirty, forty years ago, they hadn't figured out really how the accretion happens, how the mass loss from the AGN develops, and some of the phenomena in the AGN. There are these very high-density emission lines, which still nobody has really explained how they exist so close to an AGN". The gist being they haven't made much progress in real understanding over the last 30 years.

I actually wrote a paper working with a theoretician Colin Norman in Baltimore on the high-density emission line regions in only a few parsecs away from a central black holes in AGN galaxies. We suggested that the source of this emission was actually the photoionized atmospheres of red giant stars. On this project, I could make use of some of my understanding from that earlier epoch working with Peter Goldreich on the circumstellar envelopes. Do they really have the properties that you would expect as deduced from the optical emission lines? It turns out that you actually do get the right parameters. That is not to say that my model with Colin has been accepted. Still, no one really understands how these high-density clouds can exist so close to a supermassive black hole. In our model, which I just described, they can exist because there's a star inside the cloud which provides gravity to hold the cloud together despite the fact that it has a lot of radiation coming from the black hole anyway. That's the kind of realization, an idea which I find lots of fun, even though it may or may not be correct.

ZIERLER: Nick, is there a particular project that you worked on that really exemplifies the benefits of the approach you've taken over the course of your career, moving around from subfield to subfield? What sticks out in your memory in that regard?

SCOVILLE: Well, this recent project I've been doing with ALMA sort of brings everything together. It's built upon observations of the COSMOS field with ALMA and with the so-called Herschel infrared satellite, which the Europeans launched. I couldn't have done it without, one, understanding the millimeter wave emission from galaxies, either the molecules or the dust emission. I couldn't have done it without understanding the infrared emission from rapidly star-forming molecular clouds, infrared emission, which was emitted by dust but, ultimately, is a measure of the luminosity of all the stars embedded in that dust cloud.

ZIERLER: Nick, on the technology and engineering side, what have been some of the real game-changers over the years that have allowed astronomers to do things that were simply not possible previously?

SCOVILLE: Well, when I started out doing astronomy, all astronomy was done—and that was back in the 1960s—essentially all astronomy was done with photographic plates. My first project was measuring the spectrum of some comets on photographic spectral plates. God forbid that anybody should have to do that because not only is the sensitivity not very good but measuring these photographic plates is extremely tedious. Then, also, the photographic emulsion has a very nonlinear response to light, so it requires very good calibration. The major optical transformation has been obviously to electronic rays of detectors, charge-coupled devices, and at the optical wavelengths. In the infrared, when I came to Caltech in the mid-80s, one of the projects—I think it was even before when I was on sabbatical—one of the projects I wanted to do was to map the infrared emission, the near-infrared emission from a nearby spiral galaxies like M51. At that point, infrared detectors of those wavelengths were just single-element detectors. I talked to Gerry Neugebauer and Keith Matthews. Tom Soifer was probably involved. I talked to them about my interest in mapping the spiral arms of M51 to see where the mass concentrations were as opposed to just the recently formed stars dominate the visible wavelength images. So, we took a single-element detector down to the Palomar 5 m telescope. Then, the issue was that there was no integration between the data-taking computer, which was probably some small laptop computer, and the telescope control system. Our detector was the single-element detector, which we wanted to scan across M51 to accumulate a complete image of the galaxy. Gerry Neugebauer suggested that since the detector output was being plotted on the strip chart, we could tell when ‘turn around' occurred in the telescope scanning and in the data stream simply by pulling the plug on the detector at the end of each scan, and putting it back in again at the beginning of the scan – a kind of kind of hands-on approach!

ZIERLER: Nick, on the institutional side, to go back to something very interesting that you said in our previous discussion, when you had that conversation with your colleagues, and you talked about the complainers, and you exhorted everybody to remember that Caltech has one of if not the best astronomy program in the country—perhaps even the world. You also said that it's not that the people are so much better at other top-flight institutions. That prompts me to ask, if not the people—? What has been Caltech's secret for the excellence that it has demonstrated in astronomy over the years?

SCOVILLE: Extremely good students, the best crop of students who—Caltech has a good reputation in astronomy—an excellent reputation—and so it does attract sort of the top tier of incoming graduate students each year. But the thing which Caltech really has is guaranteed access to several unique facilities, and certainly the Keck telescope is a major advantage that Caltech professors have. Then having your own radio observatory, Caltech was not unique in this because Berkeley had a similar observatory, and Illinois used to have a radio observatory. But the Caltech radio observatory has actually survived up to the present, and it is going strong now.

ZIERLER: I hope I do. [laugh] Well,

SCOVILLE: The other thing is, of course, at Caltech, there's a fantastic mix of disciplines. I could work with physicists like the infrared group. I don't mean to imply that the Caltech professors are not good. They are good! It's just that I think they maybe have a little too easy life, and they aren't pushed enough. In the case of being a radio astronomer, like myself, we always had to go out and get funding from National Science Foundation to support the Owens Valley radio telescopes, whereas in the case of Palomar and even Keck—nowadays, for instruments, they do go to the NSF to fund new instruments, but the bulk of the support for Keck is provided by the Keck Foundation and by Caltech and UC out of their operating budgets. For some reason, the radio astronomers have always been in the habit of having to raise their own funds, and that, to some extent, is true of the infrared astronomers. But they've certainly made good use of Palomar and Keck.

ZIERLER: Well, Nick, one last question looking to the future. Obviously, your career demonstrates right now emeritus does not mean retired. For however long you want to remain active in the field, what's most important to you? What do you still want to accomplish?

SCOVILLE: That's a slightly delicate question because I do sculptural welding as a hobby. I put that off for many years, just doing, say, one artistic project every couple of years, and I really want to get back to doing that as long as I still have the physical stamina to do it. This ALMA project which I've described, I think that may be my last big science project. I hope it's my last big project because I'm now 77, and I really want to be able to do another few projects which would be artistic. I've found a great deal of intellectual joy from discovering from starting from raw metal where you don't really know what you're going to do. Then when the final result actually liked by your friends is an added treat. One of these days, I'll show you some pictures. [laugh]

ZIERLER: What's going to become of ALMA? What do you hope to do for as long as you'll be a collaborator on it?

SCOVILLE: At this point, I have been making use of data in their public archive although initially I did have extensive project of my own initiation. I'm a senior person in the field, and the work that I did and which Anneila did initiated the support for building ALMA neither of us really active in augmenting ALMA at this point.

ZIERLER: Do you see your artistic sensibilities as an outgrowth of your career in astronomy, or do you tend to think of them as separate endeavors?

SCOVILLE: I tend to think of them as separate. I don't do astronomically oriented sculptures!

ZIERLER: [laugh]

SCOVILLE: —For sure.

ZIERLER: Well, I'd love to see them sometime.

SCOVILLE: [laugh] I do some furniture and then also a mobile of fish with screens on both sides for the two sides of the fish—

ZIERLER: Wonderful.

SCOVILLE: —which generate moiré patterns as the mobile rotates.

ZIERLER: Always something to look forward to.

SCOVILLE: The AAS Russell prize lecture I give in June. That'll be a challenge – 1000 people in the audience (unless COVID peaks again) and I have to make it intelligible and interesting to all levels of astronomical people.

ZIERLER: I would love to. Well, Nick, this has been a fantastic set of discussions. I'm so glad we were able to do this and capture all of your research and contributions. I'd like to thank you so much.