Oral History Transcript — Dr. John P. Huchra
This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.
This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.
Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.
See the catalog record for this interview and search for other interviews in our collection
Interview with Dr. John P. Huchra
John Huchra; February 15, 2002
ABSTRACT: Among the topics discussed: his childhood and early education; his college years at MIT and California Institute of Technology; the effects of the Vietnam war on his young life; his classmates at Caltech; the Apollo moon landing; learning to use a telescope; astronomers poaching off others' work; his skills with a telescope; his discoveries of various asteroids; his early days at Center for Astrophysics; his work in photometry of galaxies and measuring red shifts; his job search after his post-doc position; commissioning the Multiple Mirror Telescope (MMT); the Multi-Conjugate Adaptive Optics system (MCAO); the demographics of astronomers; funding and growth in the field of astronomy. The second session focuses on his tenure as associate director of Cambridge Center for Astrophysics; comments on structure and organization of CFA, its strengths and weaknesses; the relations between CFA and Harvard and the class structure at Harvard among the scientists; challenges facing AURA; challenges facing the US National Observatory and AURA. The interview concludes with his thoughts on research on dark energy, dark matter, cosmology and his preference for a particular cosmological model. Also discusses religion, time, and cosmology. Some of the prominently mentioned names include: Jesse Greenstein, Allan Sandage, Wal Sargent, Fritz Zwicky.
McCray:So yesterday we talked just a very little bit about the beginnings of the redshift survey and the infrared and Tully-Fisher collaboration, which I guess began roughly about this same time. So let's come back to that, and just for some general background how did the redshift survey come into being? Let's just start with that.
Huchra:Well, I mean the first thing to realize is that there wasn't just a single redshift survey. There were a lot of things that were going on at the same time, and lots of people had ideas and lots of people were interested. And I can give you one slant on it, but in fact because there was some competition, there were a variety of other things happening, and in some cases people weren't exactly talking about what it was they were doing. Not everything was absolutely known. First of all, motivation. The motivation, I think, in terms of how I got into the survey and as opposed to how it ended up happening in the rest of the world, was primarily driven by the Princeton group at the time, Jim Peebles and students and whatever. He had gotten into this game in the early part of the 1970s of trying to look at that structure in the universe from a theoretical point of view. There were the competing theories that were being developed, Zeldovich on one side, Peebles on the other. Peebles and folks on one side and Zeldovich and folks on the other. Adiabatic versus gravitational instability, pancakes versus spheres basically, neutrinos versus matter or dark matter or whatever. Other piece of background is that you should realize that in the beginning of the 1970s nobody had any idea what dark matter was. They thought they were missing baryons. We are not talking here about stuff that has properties very different than the physical properties of ordinary matter.
McCray:Were Ostriker ideas out? They were circulated by that point.
Huchra:Well, Ostriker is just a tiny piece of the puzzle. And which ideas do you mean actually?
McCray:The rotation, the halos around galaxies.
Huchra:Interesting question. Ostriker, Peebles and Yahil’s paper was1974 give or take a year. Okay. There were a couple of things that were going on at the same time because the competing Russian group Onosto, Cosack and Szar also published a paper. I think they beat Ostriker actually in one of the Soviet astrophysics journals or the Russian astrophysics — no, better say Soviet, it's safer, journals. But the idea was out, driven by two things, one of which was this business that as you went to larger and larger scales it looked as if there was more and more dark matter. Okay. And this idea that was driven a little bit by Toomre, et al. It could have been Toomre and Toomre, the brothers Toomre, about the stability issues. And the fact that this generally were not stable to bar instabilities if the masses were what they thought they were based on just on the baryon content. So there's a whole, whole string of things that went in there. And I'm not sure I could reconstruct it from the theoretical viewpoint, but the time was right. In terms of redshift surveys, it really wasn't that so much as it was this idea that on larger scales you could begin to see things in the galaxy distribution. And I think the real driver for most of the observational efforts was the map that Peebles made based on the She-Mortannon counts that showed these frothy filament type things. It was Peebles and Soniera, but before that Peebles and a couple of other people, Peebles and Hawser and Peebles and — I forget the whole list, but Peebles and many postdocs and students, and I'd been working on this. And the Peebles and Soniera map based on the She-Mortannon counts, which also goes back to about 1974, really laid this picture out there of a filamentary structure in 2-D, and there was a debate about whether the 2-D was 3-D or not, or whether there was something else that was going on — how uniform was it, what's the real structure, all of this. Then there was also at the same time that the mathematical descriptions that Peebles started to apply to the clustering that could be seen in two dimensions. There was this whole description of clustering in terms of correlation functions. So the 2-D correlation, omega of theta, the 2-D, little omega, the 2-D correlation function for galaxies on the sky.
McCray:Just for the record, what's a correlation function? Because not everyone knows.
Huchra:Ah. Okay. Correlation function is a measure of excess clustering. For example, the way it's measured for galaxies — and you could do the same thing for people on the average — is measure the average density of people at different separations.
Huchra:And because people tend to live in cities, they cluster, people cluster. If you look at different scales of separation you'll find that there is an excess relative to a random distribution and it's usually defined relative to a random distribution on small scales because people live in cities. Well, galaxies cluster because of gravity, and the excess clustering of galaxies on small scales due to gravity is in fact a measure of the overall gravitational potential of the universe.
Huchra:If you can make a measurement of that. That was the hypothesis. If you could make a measurement of that, how galaxies clustered, then you could put in the appropriate scales, you could then go from that to a measure of the integrated mass density of the universe. Those ideas go back to the middle of the 1960s. There's a famous thing called a Layzer-Irvine equation which in fact is the relationship between clustering, and one way of looking at the relationship between clustering and gravity on larger scales in the universe.
Huchra:Peebles came up with the — how can I put it? — the description of that function in terms of the two point correlation function for galaxies. You could measure the angular two point correlation function, i.e., the separation of galaxies on the sky, the excess clustering as measured on the sky. What you really wanted to do with the gravity was to get the two point spatial correlation function, and to get the spatial correlation function you needed distances.
Huchra:And, in those days, distances generally meant redshifts.
McCray:So you want to know the distances between these points.
Huchra:Right. You want the three-dimensional distance; not just the angle on the sky. Because two galaxies, you could have a galaxy over there and a galaxy over there, but along the line of sight they could be sitting on top of one another. It doesn't count. What you want is a real three-dimensional separation of them.
Huchra:You want that measure. So Peebles essentially started egging on his students and postdocs and people he came in contact with to do something about this. Jim is a theorist par excellence. I don't believe he's ever been to a telescope and taken data in anger. I could be wrong, but we could go back and look that up. On the other hand he had a lot of students who did fall into the trap of being experimentalists. A lot of people he interacted with, including J. Richard Gott, who was working with Ed Turner when he came as a postdoc to Caltech. [portion of transcript restricted]
McCray:Okay. Were there any differences in style or approaches between the Princeton group and the Caltech group before the Caltech group went belly up?
Huchra:Might be differences in emphasis, but the point of the matter was that the Princeton group and the Caltech group, there was an awful lot of intermediate vector bosons there. Intermediate vector bozos. Which is another way of saying that — Thuan had been to Princeton. Knapp was now at Princeton. The motivation as I saw it, as primarily coming from Gott and Turner, where Gott had been to Princeton. So you can track everything back. The emphasis on how to do the problem was perhaps a little different, but only a little, because there are only so many ways you can measure redshifts for things and how one goes about doing it. The emphasis on what people wanted to get out might have been different. For example in the Caltech collaboration we wanted to get the radio velocities but we were also interested in a lot of other stuff. In particular I mentioned that we started doing photometry of the galaxies as well.
Huchra:No. The optical photometry.
McCray:Optical photometry. Okay.
Huchra:Optical photometry of the galaxies. Time we got at Kitt Peak, for example, we'd do photometry of the galaxies on the 0.9 meter, the other 0.9 meter, and we'd do redshifts on the number one 0.9 meter. We'd go to Palomar and measure redshifts on the 60" and do photometry at Mt. Wilson.
Huchra:So we were trying to get information about the properties of the galaxies as well as just look at them from the point of view of points in space, you know, test particles. The Princeton guys were primarily interested in test particles. Caltech guys primarily interested were interested in the test particle aspect but also wanted to study the galaxies.
McCray:That's a very interesting contrast in terms of interests
Huchra:Yeah. Some of it is history. Peebles is in the physics department.
Huchra:And in those days I think to him the concept of a galaxy is anything other than test particle, or maybe an ensemble of test particles. It was different. And the Caltechers, driven by the long history of work at Palomar and the like were much more interested in the properties of galaxies — what you do with the spectra. And in fact that kind of attitude has continued for a long stretch of time, in the differences between where the people have come from. So I am still very much interested in the properties of the galaxies that one studies in redshift surveys, not just the use of the galaxies as test particles. In part because the properties of the galaxies can sometimes also tell you something about other things that are cosmologically interesting.
McCray:; Such as.
Huchra:The luminosity function. One of the key factors, or one of the key things that you are trying to, or that any theorist worth their salt is trying to match when they make simulations into both the structure and galaxy formation, is the observed current-day luminosity function or mass function of galaxies. You want a theory which predicts the number of galaxies per unit mass. And if it doesn't give you the right answer then you know you've got a theory that isn't right. So it is a test. It's a detail test that we have a lot of information on — because of studying the properties of galaxies in redshift surveys. Anyway, that's probably the best way of describing the differences. In terms of getting the data, the Caltechers primarily were trying to do this with existing facilities because they had more access to the existing facilities. And as I said, we had Palomar, we had Mt. Wilson — the small telescopes, not the big ones — and we were getting time at Kitt Peak on the small telescopes there, which are plenty good for doing this. At CFA, Marc was extremely interested in starting up a redshift survey.
Huchra:Marc Davis. But frankly, for the first year or so that I was here there was no way to do it. They have made some effort to try to improve the detectors on the spectrograph at Mt. Hopkins on the 60", on the Tillinghast reflector, but there were just a lot of things wrong. And I mentioned yesterday, we literally took the spectrograph apart, redid the optical alignment and the position of the optics to improve the throughput. Part of the problem — and I also mentioned yesterday — I think we were talking about this. You were asking how one goes about doing things like setting up new telescopes and the like, and I said “experience counts.” This is one of those cases where at least I was able to bring from my experience at Palomar and Kitt Peak with spectrographs on similar size telescopes the knowledge that the one we had at Mt. Hopkins wasn't working. So the immediate comparison is you know, hell, on the 60" at Palomar we can get the spectrum of a 12th magnitude galaxy in two minutes and on the 60" at Mt. Hopkins we get the spectrum of a 12th magnitude galaxy in an hour. There must be something wrong. Then that causes you to take things apart — or be allowed to take things apart, and figure out what's wrong and fix it up.
McCray:How long did it take before you had diagnosed the problems with spectrograph?
Huchra:My first observing run. At which point I wasn't interested in using it. Why bother? Marc was interested in trying to fix things, as I said. We had this problem of what was out there wasn't working. And another key piece of information in the redshift survey game was that I was still going back and forth to Palomar to observe and I was there I guess it was in the winter of, I want to say the winter of 1977. That's probably right. I was out observing at Palomar in the winter of 1977. Steve Schectman, who was a friend of mine, had gone back to Carnegie after being at Michigan to help them build instruments and do a variety of things like that. Steve has been the guiding light behind the Magellan telescope projects in Chile and put a huge amount of effort into that. Anyway, Steve when he was at Michigan working on an idea that goes back Jim Gunn, built a photon counting system, a photon counting Reticon. Reticon being a detector that was made by a company that trademarked it. And this was a detector system that could be used on a spectrograph in particular that consisted of a linear diode array or a pair of linear diode arrays — not quite 2-D CCD as we call them, but they were charge coupled devices. Sitting behind an image tube, image intensifier stack, such that a photon would come in. Roughly speaking with an image tube you get 20 percent of photons detected, so one out of five would make it through the image tube and produce a splash of a million photons at the back end. And this little diode array would record the centroid — well, record the splash of photons, and then some hardware, you do this in hardware, the hardware recorded the centroid of the splash of photons as it came in. So that you could actually record the position on the diode array of the incoming photon and at the front end the image tube package. Which allowed you to measure spectra of things — of galaxies or whatever. And I was talking to Steve on the catwalk of the 200" and it was raining, which was why we were on the catwalk with the 200" not observing, and we were talking about redshifts and spectra and whatever, and Steve mentioned that he had built this instrument. And I said, "Oh, how much does it cost? What can it do?" And he said, "Oh, it cost me about $25,000," and I said, "Mmmm, sounds good. Could we copy it?" And he said, "Sure." Now the backdrop to this was that there were a bunch of things going on at the same time in different areas. In fact the leading spectroscopic detector at the time in the community was a semi-commercial one that had been built at San Diego based on a design at Lick that Wampler had produced.
Huchra:Well, no. It was a modification of the Wamplertron. Well, Wampler's was the image dissector scanner, but the San Diego version of this, the Beavers and Harms version of it was a version that essentially had one of these diode arrays that was mounted not behind the image tube stack but inside an image intensifier stack. So it detected the electrons instead of the photons.
McCray:Is there an advantage to doing one versus the other?
Huchra:Yeah. The other. The photon counting Reticon system was better, but that's because it's centroided and a variety of things like that. Anyway, these guys in San Diego were putting together these systems and in fact that detector was the one that was chosen for the original faint object spectrograph on the space telescope. Okay? The one that was the diode array inside the can.
Huchra:The disadvantage of that system is that photocathodes decay. The advantage of the Reticon system is that when your front photocathode or your back photocathode died, you just stuck in a new one and away you'd go. But with this other system you got what you paid for. The other disadvantage was that the cost of building one of these was a quarter million dollars, not twenty-five thousand.
McCray:The factor not being much better.
Huchra:Well, I mean, yeah. Sometimes things are like that. Okay. So there was a name for the company, and I cannot remember it because it was twenty-five years ago, but Marc wanted to buy one of these things from San Diego from the company that had been set up to market them or sell them to astronomers. Trying to get the money from the NSF to do it, and tried to get money from inside the observatory to do it, but at the level of a quarter of a million dollars a year they weren't biting. On the other hand at the level of $25,000 we were able to sell — I came back, talked to Marc — we were able to sell to the folks here, and most notably Herb Gursky, who was a then associate director of the MYR division, and also George Field, who was the director of the observatory. The idea that we could copy this device that was already working on the Michigan telescope that Schectman was building for Palomar and for Las Campanas and get it on the telescope for much less than the quarter million plus bits and pieces. A quarter million bought you the tube. It didn't buy you anything else.
Huchra:Twenty-five thousand bought you the tube but didn't buy you anything else. You still have to get computers and other things like that. So we support from inside the observatory, and I don't know for sure how this worked, but I think George also went to the NSF and called them up and said, "Hey, these guys really need some support for whatever," and we also got some support from the NSF to go on with the program. NSF primarily supplied the science end of it and the observatory here primarily supported the building of the instrument.
McCray:Was this the first instrument that you had worked on hands-on building?
McCray:I mean I know you worked on the Fabry Perot.
Huchra:That was the first instrument I worked on hands-on. Well no. The first instrument I worked on hands-on building were the bloody modulation collimators.
Huchra:That was extremely hands-on [laughs]. But yeah. I mean, this is probably the first one where I had some major role in the conceptualization of the instrument.
McCray:Okay. I guess that's what I was thinking. I mean, from start to finish.
Huchra:Yeah, yeah. But it really was Schectman's instrument at that time. And Marc went to Pasadena and spent two months living in a shack in the back of Schect's house there and copied the electronics. Daryl and I worked on the spectrograph. I remember getting out a diamond saw and sawing one of the mirrors into to get it to fit in the right place. Dave Latham was given the task of working on the camera package, the image tube package for it, so he built up some speed in doing that kind of work. And we connected up with a set of other people who were here. One of Marc's students, who was John Tomry, was a graduate student that came to work with us. He did a lot of work on the software, particularly analysis. Not the analysis software, but the software for turning spectra into redshifts. So it's analysis, but not producing the theoretical final result. Producing a redshift from the spectra.
Huchra:And two systems programmer types who were here, Tom Stevenson and Jim Gettys who worked essentially on the data taking system. Tom is still somewhere in the Boston area. He's a bassoonist and plays in some of the local arias, local orchestras, but has gone off to get paid much more than the software consulting business. I don't know exactly where he is these days, but Tom and Jim, primarily Tom was the leader of that group, wrote the software for the data taking part of it. This was all done on data general Nova IIs which you booted by using toggle switches and then paper tape. You had to load the paper tape reader and then load the paper tape. Thirty-two commands with toggle switches. Quite a hoot. And we essentially got the system going. I went out and helped on the software system for actually taking the data, so if you ever go back and look at any of the ancient and venerable records of the software you can find a bunch of HUCHRA- isms stuck in there. I like bad puns, so the commands are often puns.
McCray:What would an example be?
Huchra:Well, probably my most famous command was the Fribble. If you've ever been to a Friendly's Ice Cream Parlor, a Fribble is a milkshake, and I was really heavily into milkshakes. It was an ice cream milkshake, pretty thick and creamy. And I am also the inventor of the Root Beer Fribble, which is to say I convinced Friendly's to not just make something with vanilla ice cream, but also to add some root beer syrup, and that became the Root Beer Fribble. [Huchra shows interviewer notebook]
Huchra:This is the logbook from the first — this is the first logbook from the photon counting Reticon system, and the first day that we took data was February 23rd, 1978. And this was on the 24-inch telescope at Mt. Hopkins, because we didn't get the telescope time to man it on the big telescope until we could show it could work. People were very conservative. And if you look through here, most of what was being looked at the beginning were bright stars. Lots of comparison lamps and standard stars. And the first galaxy was NGC-2778. The first redshift that we measured with the photon counting Reticon system for the CFA I Redshift Survey was NGC-2778 on February 25th, 1978.
McCray:That's a pretty quick time from starting to collect data, two days.
Huchra:Oh, well the time from starting to build the spectrograph was really a year.
McCray:I mean just to get it on the telescope and start using it and two days later you are getting some data from it.
Huchra:Right. Well, we spent a lot of time in the laboratory here making sure it worked and all that kind of stuff.
McCray:Okay. How long did the building process take?
Huchra:About a year.
McCray:About a year. Okay.
Huchra:That included Marc copying. Well, maybe a little less than a year. Just remember we were duplicating an instrument. So I'd say a year including all of the initial development work getting the software system started and whatever, but we started — I mean as I said, my meeting with Schectman was in the winter of 1977 and this was the winter, the end of the winter of 1978. And Marc went out in the summer of '77 to copy the electronics on it. The advantage was that we were copying an existing instrument. We weren't starting from ground zero.
McCray:How common was it or is it — was it — to be able to copy someone's instrument and have them say, "Here are the plans, and these are the schematics, go ahead and do it"?
Huchra:Fifty-fifty. I mean, what you find is that in the game first of all a fair number of the spectrographs that were floating around were commercially developed, so the answer for those is no. You go to Kitt Peak; they talk about the B and C spectrograph, the Boller and Chivens spectrograph. Or a spectrograph that had been built for Mt. Hopkins, the initial one, the spectrograph itself — you know, remember what we copied from Schectman was not the spectrograph, it was the detector package.
McCray:Okay. For the back end of the spectrograph. Okay.
Huchra:Yeah. But the spectrograph itself was built by a commercial company called Boris Botts. Incorrectly, I will add, it was a commercially built spectrograph. That was before my time, before I came here. I guess they got the spectrograph around '72 or '73. So it was often the case that if somebody built something individually and there were no commercial connections or ties or whatever, you could get the plans and copy it. It was very often the case that people would publish things. The other thing was that — I mean that you should realize, since you are talking to people about projects like Gemini and the like is that in those days you didn't do any engineering on these things. You could put instruments together with sealing wax and duct tape and bailing wire and it would probably work well enough. Because basically instruments were small. You required some care in the building of the instrument, but you didn't for example have to do a structural analysis of the bending modes of the spectrograph on the back end of an 8-meter telescope. But you do now, because the instruments are just so much bigger that if you don't take all of those things into account, if you don't do your finite element analysis of torsion and flexure you can compromise the properties of the spectrograph. But when you're sticking something on a 1.5-meter telescope it wasn't such a big deal.
McCray:How big was this whole thing?
Huchra:Half the size of this desk.
McCray:The whole thing?
McCray:Okay. So about 3X2X2.
Huchra:Well, no. No, no. Lengthwise it was probably about 5 feet — 4½ feet by 2 feet by 3 feet would be a good guess. I'm sure it's hanging somewhere. In fact it's hanging in Washington now.
McCray:Oh, this is at the Air and Space Museum.
Huchra:Right. It's in the Air and Space Museum.
Huchra:Right. Okay, so basically we got it on the telescope in about a year from the time the concept of using that detector package and modifying the spectrograph worked. My job was to — my initial job in the project on the science side was to put together the catalog, and we basically decided to go half a magnitude deeper than the original Caltech group proposal. We would go to 14.5 in the Zwicky catalog. That was a sample in the northern and southern galactic caps available. You know, Zwicky only went down to the Equator, so it was north of the celestial Equator. We didn't go to low galactic latitudes, because the catalogs aren't very good at really low galactic latitudes. We stuck to high galactic latitude north of the Equator, so it was a little less than a quarter of the sky. It had 2400 galaxies. There are 2398 galaxies and one star in the catalog. So he didn't get it up perfect. It's still a pretty good error rate as catalogs go. And we started — you know there were some redshifts that existed already. Sandage and Tammann for example had been putting together their revised Shapley-Ames catalog. They finished about the same time we did, but theirs is all sky, 1300 galaxies. So there is this beastie, the revised Shapley-Ames catalog. The original Shapley-Ames Catalog with Shapley and Ames in 1935 here and what Uncle Allan and Gustav were doing was measuring redshifts using Mt. Stromlo in the south and Carnegie in the south and Palomar and the 200" in the north measuring redshifts for all the galaxies in the Shapley-Ames Catalog. So there was another redshift survey going on at about the same time.
Huchra:These guys were pretty secretive about what they were doing.
McCray:Hold that thought. I have to flip this over.
McCray:Secretive in what way?
Huchra:In the sense that they hadn't let out large quantities of the data.
McCray:Is that common?
Huchra:Uh — yeah. In fact it's normal, unfortunately. And people knew that they were doing it, but there were a couple of analysis papers written in 1979, give or take change, by Sandage, Timon and Yahil. In fact there are three papers with the rotations as appropriate. There's Yahil, Sandage and Tammann; Sandage, Tammann and Yahil; and Yahil, Tammann and Sandage. Everybody gets one shot at one spot in the rotation. Analyzing the Shapley-Ames Catalog. We finished taking data in 1981.
McCray:Just a couple questions about collecting the data. How many galaxies would you study per night?
Huchra:Oh, typically we would get twenty-five or thirty redshifts a night, you know, a few more, a few less.
McCray:Okay. What would a typical observing run be like?
Huchra:Well, when we settled into the game, in the game we hired on a couple of remote observers. We called them remote observers, but in reality it's queue observing, so there would be two people who were working on Mt. Hopkins as technicians. Both of them in fact did work on the Smithsonian Satellite Tracking Program before, but as the satellite tracking program wound down in 1980, you know, and moved over to Bendix, a commercial contractor as opposed to SAO which is a semi-commercial contractor. As that program wound down there were various people out there that were looking for things to do, and we found support for them to work taking data for a variety of different programs, not just the redshift survey. So they became the queue observers on the 1.5-meter, and those guys did a half to 60 percent of the observing for the redshift survey in the last year. I probably did about a third of it, because I liked it. And the other sixth or so was done by sometimes Marc, sometimes John Tomry, sometimes — what was his first name? The last name was Goldberg. What's his first name? Al Goldberg? Oh, I can't remember his first name.
McCray:Okay. We can find it.
McCray:How did you divide up the work? Who said this person is going to do this and this person is going to do that?
Huchra:I don't think anybody ever did.
Huchra:We were a very small group. Basically the people who were interested in pushing on this were Dave and John, being from a scientific point of view. By that point in time I guess I was making up the telescope schedules, because by 1981 I was Smithsonian staff working at Mt. Hopkins a lot. So I was making up the telescope schedules and we just rotate things around. And I liked to fill in for the remote observers as much as possible. So I'd link things to whatever I had running at Kitt Peak or something like that to minimize the travel costs, and the whole game like that. You play those games, right? Anybody could come out when they wanted to if they wanted to. If somebody wanted to take my place, that was possible too.
Huchra:Sometimes we'd all go out every once in a while just to be there and check the instrument, see what was going on. John started working on his thesis, which involved more software work in particular. In fact his thesis ended up not being about redshift survey but rather using the data from the redshift survey to measure velocity dispersions of galaxies to use to do flow fields — to use what was in those days called a Faber-Jackson relation, the relationship between the velocity dispersion of the stars in an elliptical galaxy and its luminosity. The more mass the galaxy is, the brighter it is and also the faster the stars move around in the galaxy.
Huchra:It's like the Tully-Fisher relation, but it applies to elliptical galaxies. So that was John's thesis, so he was back here working on that a lot. Marc was working on the analysis programs. He was doing these long series of papers with Peebles in fact on integrating the BBGKY hierarchy of equations for the growth of structure.
Huchra:KY. Don't ask. I don't know. And I was doing what I did best, and what I did best is sit on the back end of the telescope and take the data. So we were all pretty happy with that game. There was no field marshal. We had had some discussions about how to divide up the responsibility for analyzing the data and I was primarily interested in doing the luminosity function work. We were all interested in looking at the maps, what was the distribution. Marc was interested in the correlation function work. John was given the problem of doing the flow field velocity dispersion work and I guess Marc and I were independently interested in working on groups of galaxies, you know, studying clustering but from different techniques. [portion of transcript restricted] Marc was working with Bill Press on group-finding algorithms and cluster-finding algorithms and redshift data. And we almost by definition decided to go separate ways.
McCray:You and Marc.
Huchra:On this particular problem.
Huchra:[portion of transcript restricted]
McCray:What is a percolation algorithm?
Huchra:You link things together based on separations until you get to the place where the inter-particle separations are too big. So you define — you say,— Okay, we're going to link together any galaxies that are separated from any other galaxies by some distance that generally corresponds to an over density. So you take a sample. Suppose you've got a thousand galaxies and you spread them out over a thousand square miles — well, is that a good number? If you have a thousand square miles and a thousand galaxies, the mean inter-galaxy separation and the random distribution, I'd have to think about it, but it would be something on the order of one mile or two. So you'd expect average separation to be a mile or two. But if things were closer together than a quarter of a mile you could say that they had somehow gotten bound, probably.
McCray:Okay. There is a relationship between them.
Huchra:Yeah, yeah. Probably gravity and well, in our, in everybody's thinking it was gravity. So things that were separated by a hundred miles were probably not related in any way. Things that were separated on small scales – It goes back to this clustering business. I mean, how do you define clustering. And in the case of percolation you've the mean inter-particle separation and then inter-particle separations that are smaller than that can be linked to over-densities. So if you find things that are separated less than the mean inter-particle separation by a factor of ten, that in three-dimensional space that usually means that they are associated with an over density that's a factor of ten cubed, right? Over a density of a factor of a thousand. So you do that. You say, "Okay, we want to look for things in regions that are over-dense by a factor of say twenty. That sets the separation cutoff you want to go to relative to the mean inter-particle separation in a uniformly distributed or randomly distributed sample. And you just link things together. So if you find two galaxies that are closer than that, you link them. If you find another galaxy that's closer than that to any of the two that you have already linked, you link that one, and you keep going until you can no longer link things.
McCray:Okay. Why is this called percolation?
Huchra:Because usually the way you implement this algorithm is by starting in a galaxy and looking at it, and when you stop finding things you pick the first unlinked galaxy and start looking out from around it. And you percolate through the problem until there's nothing left.
Huchra:The concept of percolation was first used by Zeldovich. The concept of “friends of friends” was first used by us, and in fact did the same thing. So I invented the percolation algorithm for linking galaxies together based on discussions with Margaret and Paul Schechter and a couple other people. And it got called the percolation algorithm at about the same time theoretically by Zeldovich and company — even though they never used it to do anything except the theoretical simulation.
McCray:Okay. Whereas you were using it to explain the experimental results that you were doing.
Huchra:Right, right. And we hadn't given it a name. If I had called it percolation I would be even more famous. It's all in the name. What can I tell ya? Anyway, so we did that, and that was an algorithm that in fact did work and it did produce results that were meaningful and that looked like real clusters in groups of galaxies and that you could begin to analyze for this question of what's the mass on different scales.
McCray:Where in time are we?
McCray:Okay. Would this sort of be the end of the first part of the survey then?
Huchra:Yeah, sort of, because it was in 1981 basically just as we were finishing up taking the data that — Well, here's some down and dirty. This department, the Harvard University Department of Astronomy, as usual, managed to get its knickers in a twist, because they had two young people, both of whom were pretty good, coming up for tenure at the same time, and both of whom were not so distantly related in what they did that one could easily consider them separately. There was Marc and there was Josh Grindley. And basically Josh had an incredibly powerful supporter in the form of Ricardo Giacconi. Marc had no incredibly powerful supporters. His strongest supporter was me, and I was not on the faculty — although by that time I had tenure, you see, because I was career on the Smithsonian side.
McCray:Tenure with SAO?
Huchra:Right. As much as it means. I mean, it doesn't mean that much, but it was tenure at SAO. Okay. So I was in essentially a permanent position. It was a hoot. I was the last person to come here in the group and I was the first person to end up in a permanent position. Primarily because of the infrared Tully-Fisher stuff — because I had more than one thing going at the same time. And Marc was up for tenure, he was an associate professor up for tenure, and the department in its infinite wisdom, mostly because of Giacconi, gave the nod to Josh and did not give the nod to Marc. And they were going to try him again next year, but by then signals are signals and all that kind of stuff, so he looked for a job out west and went to Berkeley. And Harvard eventually did offer him a tenured position, but by then it was too late — i.e., the department screwed up. It does that a lot. Again you better seal the tape. It was actually a really silly thing, because I think losing Marc was a mistake.
McCray:How did that affect the survey work?
Huchra:Well, I mean the first CFA Survey was pretty much done. I was still here and interested in working on galaxy spectra and redshifts and the like. [portion of transcript restricted] There were a couple of things that happened sort of at the same time from a scientific perspective. The Einstein satellite had gotten launched in 1979, so there was a fair amount of X-ray data coming back on galaxy clusters. That was one of the things that we did when we were interested in making the connection between X-ray maps of the hot gas and clusters and the optical dynamical maps. We found ourselves competing with Bill Foreman and Christine Jones in the X-ray division here, so an interesting period of time when there were two or three different sets of proposals to work on X-ray clusters with the 16" telescope or with the MMT. And there were fights that went along with that too. Because I found myself in the curious situation of, by virtue of what was happening, by virtue of my job at the MMT I had to go out and observe for people whether or not their observing proposals made any sense. And if I went out and had to observe for somebody whose observing proposal was one that could not be done and I didn't come back with any data I'd get yelled at. Which didn't make a lot of sense. So I developed a rather adversarial relationship with my boss, i.e., the guy who took over after Herb Gursky, which is Dave Latham. [portion of transcript restricted]
McCray:They could not be done because —?
Huchra:Objects were too faint — or whatever. Real classics. This still irks me quite a lot. It still happens. But now I don't have to do the observing for other people, so it's less of a problem.
Huchra:But I was in this rather curious situation, as I said, of having to do the observing for other people on the MMT and had them with their programs or whatever — which often meant going to the telescope and taking the data for them. And I had no say in what was happening. I finally got mad and went to George Field and said, "This has got to stop. You cannot hold me responsible for things that I don't have any say about. I am not going to take responsibility without authority. Either I am going to have some say over these observing proposals or those turkeys are going to have to try to do it themselves." And when George was convinced that — how to put it — that in fact they weren't such good proposals, I got put back on the time allocation committee and got a little bit more control over what actually went through.
McCray:Okay. Were you still though going out and collecting the data?
Huchra:Yeah, yeah, and yeah. That starting dropping down a little bit — first of all as the MMT became more and more of a user instrument as opposed to a development instrument. It was easier for people to go out and actually get the data collected.
Huchra:I still spent a lot of time helping people. I actually enjoy helping people, especially if they've got good ideas that can be done. Anyway, that went on. Margaret and I worked on galaxy clustering for a long stretch of time, like three years or so. Probably you can go back and look through the literature. You'll see a bunch of papers that came out of that work, both using the MMT and the 60". Students like Mark Postman and Tim Beers and various other people, Giles Chapman, JH: Valerie La you know, some of whom are still around in astronomy. And you know later on Ann Zabludoff and company. And at the same time in the backs of our minds — certainly in the back of my mind — was the idea that we had a telescope and once we get finished doing these clusters and doing the galaxies that can be done with clusters, maybe it would be fun to try to start up another big survey. Because the telescope and instrument were capable of going faint or going deeper, seeing more galaxies, doing whatever. And again you'll get different stories from different people, but from my perspective there were two or three different groups of people at the time thinking about the issue. About how to go, you know, what was the next thing to do. There had been the Las Campanas Survey which came out. This is in order of intellectual contribution probably Schectman — Oemler, Schechter and Kirshner, but that's maybe an unfair characterization. They always published their papers in alphabetical order, which always put Kirshner first, but I don't think that was actually how things got done. But that's a different story. I don't know. I'm just guessing. Schectman doesn't get enough credit for the work that he put in on that.
Huchra:In fact the first paper didn't even have his name on it, despite the fact that he built the instrument. So, a little different. Anyway, they had done a survey at higher redshift but very, very sparsely sampled. There were a couple of different ideas that were floating around about how to go deeper, how to sample structures on larger scales, and the joke I tell — and it is a joke, because it's much more complicated than this — was that I had Margaret on one hand advocating something, I had Marc on another hand advocating something else, I had another guy, Simon White, who had been working with Marc and sometimes with me on another hand advocating something else, yet another technique. Marc into sparse sampling, ala Las Campanas; Margaret wanted to do strips; Simon White wanted to do a chunk but do it completely more or less. None of these were absolutely clear, and it was interesting listening to all these folks duking it out over what the strategy was to do the next thing. [whispers:] But I already knew! Because from the point of view of the telescope, from the point of view of taking the data, there is only one. We go back to bad science fiction movies [referring to “Highlander,” sci-fi movie]. There can be only one way in which you do this. The most efficient way to take the data —
McCray:In terms of actually using the telescope?
Huchra:— was in strips.
Huchra:Zwicky, our friendly little catalog over here, came in declamation zones. If you go through this you'll find that things are arranged in declamation zones that are 6 degrees wide by whatever long. Right? In the Northern Hemisphere about 100 degrees long, 120 degrees long, 6 degrees wide. The telescope of course, if you make small moves at the telescope it can move relatively accurately. It doesn't take time to do the slews. Also if you make small moves you don't have to calibrate as often, because it's when you make a big move with the telescope that you have to correct for all the flexure that goes on with the detector.
McCray:So this offered a way of making the most use out of the telescope time that you had.
Huchra:That's right. It probably increased the efficiency of taking the data by 30, maybe 40 percent if you did it in strips as opposed to doing it in one place or doing it randomly around the sky. It made a lot of sense to do that. It also made a lot of sense to do that because you could probably get a result out faster that way, because you could publish an individual strip. So we set out to do that. I guess we first started taking — I'd have to go back and look at the logbooks, but when we first started taking data around 1985, maybe the beginning of 1985 or whatever for the first of the strips surveys that were part of the second CFA Redshift Survey. I always call it CFA II. Not everybody will use that nomenclature, but it was the second CFA Redshift Survey as opposed to CFA I.
McCray:[portion of transcript restricted]
McCray:Right. I understand.
Huchra:Yeah. In your pocket PC. But in those days it was a big deal.
McCray:Well how did you, I mean just physically? I mean you have all of this data. How did you manage it and archive it?
Huchra:In the early days, here, I can show you. Showing you might help.
McCray:Okay. So you're showing me a bunch of black binders.
Huchra:In the binders we've got every paper that's ever been written, at least that we need to have copies of, of old data that's been taken of redshifts.
McCray:A wall of binders.
Huchra:Even today whenever we enter something in the catalog, if it exists in paper form — and now there are new databases that don't.
McCray:Okay. These are all labeled Z-cat, which I am guessing is redshift catalog.
McCray:And how many binders are there roughly? Fifty-five?
Huchra:The integrated total, there are probably about fifty.
Huchra:And these are 2" thick.
McCray:All right. So we were just in the office next door and you were showing me the Polaroid Sky Survey. I'm guessing that's sort of the nickname that was given to it.
Huchra:Right. And now we don't need it because the Palomar Sky Survey has been digitized and in fact the Palomar and the SRC Surveys have been digitized. When we need to make a finding chart we can pull the image off disk as opposed to pull it off using a little piece of Polaroid film. The Polaroid — you notice the Polaroid Company has gone belly up.
Huchra:Well, they lost my business. What can I say? About five years ago. I don't think that was the cause, but yeah.
McCray:What's interesting to me is that the work that you're doing here, this was a very major project and was one of the big projects in astronomy that had multiple authors, multiple people working on it. Was there ever any question about assigning credit when it came to publishing papers?
Huchra:[portion of transcript restricted]
McCray:Okay. When did phase II or the CFA II come to an end?
Huchra:The last data was taken — Well, let's put it this way. There is still some cleanup — there was still some cleanup work being done as late ago as about three or four years. The last major set of data was taken around 1995.
McCray:Okay. So would it be fair to say that the CFA Redshift Survey really had two distinct faces, then, the one that ended around '81, '83 —?
Huchra:Right. '81. Right. The papers on the first one kept coming out until about '83. And that was [inaudible phrase] —
McCray:And the Great Wall papers began to come out.
Huchra:And the great Wal papers were in the second survey. And the first survey was Davis, Huchra, Latham and Tomry, and mixes of that set of people. And the second one was Geller and Huchra. And mixes of us plus students.
McCray:For the second phase of it what were the major results in your view that came out?
Huchra:Very interesting question. The most important result was probably the very first paper, just the picture. At the time if you go back and try to actually get an accurate view of what people were thinking or saying, you can do that in the literature.
McCray:You said the biggest result was the paper. For the record, what do you mean by the first paper?
Huchra:Well, that was the original map, the first slice. And in that first slice you could really see I think for the very first time that the structure, the distribution of galaxies, was not — first of all it wasn't just filaments, and secondly it wasn't beads on strings or anything at all like that, clusters and filaments, and thirdly I think the prevailing view — You know, the majority view at the time was still that this region of galaxies in space was like matzo balls in a soup — clusters of galaxies which comprised about 10 percent of all galaxies, inside a relatively random distribution of things. There were smaller clusters and bigger clusters and there were groups and whatever, but the distribution of those things is primarily random.
McCray:So you didn't have the sheets and the voids.
Huchra:Right. With the publication of that very first picture, it was essentially incontrovertible evidence that on large scales you were seeing things that consisted of sheets, filaments — more to the point, probably bubble-like things, spherical surfaces.
Huchra:They looked awful damn round, or ellipsoidal to be more accurate.
McCray:When did that picture emerge?
Huchra:March 1986. I mean it hit the preprint stands in sort of December '85. You probably have a better record of when it was actually published.
McCray:Yeah, I have it somewhere. As this picture emerged — I mean picture I mean literally, as you began to see this emerge what was your reaction to it?
Huchra:The picture didn't emerge. Since I was so sure that there wasn't going to be anything all that interesting. Actually I was spending all my time at the telescope. We did not plot the data up until the summer of 1985, when essentially it was complete. So it wasn't the case of adding points to the map as you collect the points. It was you collect the data and all of a sudden [makes a sound like pew!]. Right? And frankly I was very surprised. It was not what I expected. And you know I've gone on record as saying that my first words were, "Oh Lord, what did I do wrong?" Those were my first words, "Oh Lord, what did I do wrong? It doesn't look right. There must be something not right."
McCray:Okay, when you say it doesn't look right —
Huchra:It hadn't conformed to my expectations.
McCray:What was your world view of how the universe should look like?
Huchra:Well, as I said, the prevailing world view, the majority — not an overwhelming plurality — was that the distribution of galaxies in 3-D space was primarily clusters, these matzo balls, in a uniform soup, where the soup had some structure but nothing that had patterns. My view of the map as it came out was pattern. Here we have a pattern, a real live pattern, and I expect to see pattern; I expected to see randomness. And the pattern was surprising. And the amount of, the volume of space occupied by voids. Now the first void had been discovered by the KAOS Surveys, the Las Campanas survey, although in those days it wasn't Las Campanas. It was up in the north. The Bootes [pronounced boe-ay'-tees] void. And most people thought that was a fluke. Because the way — You should understand the methodology. The void in Bootes was found by these guys essentially doing a redshift survey a small patch here, redshift survey a small patch here, a redshift survey a small patch here, redshift survey a small patch here and looking at the results from all of those surveys and finding that over some redshift range there were no galaxies or not very many galaxies. So this must indicate that there is a void. There was no picture of it. And in that picture there were voids plural, many voids, and they were round and they were big. And it took me a couple of weeks to really be sure that the data was right and that there wasn't anything really screwy with the catalog or anything at all like that.
McCray:Do you recall going home that evening or whenever you first began to see this picture and you thought you had made a tremendous mistake when going home? Do you recall what your thoughts were at the time?
Huchra:No. I didn't think I'd made a tremendous mistake. There's mistakes and there's mistakes, right? I didn't understand it. I thought it had to be checked. I didn't think I'd have to go back and redo all those radio velocities, but I was worried about the quality of the catalog and the things that went into them. So it's a different kind of mistake. It's a mistake of lack of understanding as opposed to a blunder.
Huchra:Okay. And in those days I never went home. I'd leave work at 5:00, go run ten miles, come back at 6:00 or 6:30 or whatever, take a shower downstairs, and keep working until midnight. So it was that kind of thing. If I didn't have kids I'd probably still do that, but then that's craziness, sheer craziness. Anyway, that was sort the reaction. And I think in many ways there have been more detailed physical analyses of that, but the map was really the result, the startling result that came out. The Great Wall came out a few years later and he was just showing that some of the biggest structures, biggest positive structures, were in fact pretty big positive structures. But at that point, by that point in time I at least was pretty sensitized to finding large things, and it came as less of a surprise and we put it all together.
McCray:How did your colleagues react to it, either here or outside the CFA community?
Huchra:Got a lot of kudos. In part because we did check things. Some people reacted with skepticism, but it went away pretty quickly. In fact it's gone away so quickly that now we don't get any credit for it. It's wonderful. Now it's public knowledge.
McCray:The beauty of having a big discovery and having it accepted is you don't get credit for it?
Huchra:You don't get credit for it. Yeah. Which is okay.
McCray:Were there — were there communities or institutions that had a hard time leaving, abandoning the view that everything was more or less homogenous and uniformly distributed, or did that pass pretty quickly and painfully?
Huchra:There is still a low level debate about this issue, surfaces versus filaments. And in fact it's not really a low level debate; it's a moderately high level debate. The existence of voids everybody caved real fast — You know, it's like playing poker, and you face your full house and everybody has got pairs and trips, and it doesn't take long for them to quit.
Huchra:So on the issue of the existence of voids people folded real fast. You know, binary switch in the field from the space of a couple of months from people thinking uniform to people thinking voids and something. Then the issue has come down to the question of whether it's voids and filaments or voids and surfaces. Okay. And I think that debate hasn't really been answered. Not completely. Because the theoretical predictions still look awful filamentary. In other words, you look at the simulations, the n-body simulations that people do —
McCray:The cosmic web simulations that you see pictures of.
Huchra:Right. Well, they use cosmic web. I don't think it's a web. I think that what we see really are things that are surfaces. They have more — they are not just filaments — that really have more two-dimensionality rather than one-dimensionality in it. And the difficulty at some level is that it's hard to do that, hard to prove that one way or the other unless you have good, well-sampled, deep, large area surveys. And those are coming. 2DF [punctuation?] will sort of be able to do it and the Sloan Digital Sky Survey will be able to do it better.
McCray:2DF is the 2-degree —?
Huchra:The 2-degree field that's being done by the Australians. Sloan will be able to do it probably better than 2DF, although 2DF may get there first, is getting there first. In the end you want a sample; you want surveys that are deep and incredibly complete. You want dense. You want a deep, dense survey. Dense in the sense of having lots of galaxies per unit volume so you get good sampling what the structures look like. 2MASS will do it.
McCray:I have a question about surveys. I was at one of the NGST Science Working Group Meetings and one astronomer referred to surveys as stamp collecting in a very pejorative sort of sense.
Huchra:Butterfly collecting, huh?
McCray:Yeah. It seemed to be implying that surveys were all well and good but there was something to be said for doing, well, whatever the opposite of surveying would be.
Huchra:Yes. The answer is yes.
Huchra:Which is to say they're both good. They do different things. There are times when targeted observations — I mean especially if you've got something that's very clear and succinct that you're trying to test. Or the game you want to play or the thing that you want to do. There are surveys that are done for the purpose of butterfly collecting. So, for example, if you develop a new camera that can work at a wavelength that has never before been observed and far enough away from other wavelengths so that it's really unique, if you go into using that camera knowing what it is that you are going to look at you are pretty dumb, and to some extent you are going to be using that camera to look at new things. And it is butterfly collecting, because you are going to find new kinds of objects. You find a bunch of old ones. You now know something about, in terms of new properties in a different wavelength or in a different way, but you're likely to find, the most interesting things that you're likely to find will be new things that you haven't seen before. So it is butterfly collecting. It's looking for the missing link or chasing the new species down in Amazonia or whatever.
Huchra:On the other hand, there are surveys that get done because you know you are going to find something interesting even though you don't know what it is. And there are also surveys that get done because you are trying to use the data from the survey to make a physical measurement. And in the case of the redshift surveys I think the first ones, the very first ones way back in the seventies might well have been done just to butterfly collect. But in the end, I mean when we did CFA-1 and we did see CFA-2, there was a tremendous amount of theoretical — how would one put it? — intuition, theoretical backing, theoretical groundwork that went into doing the survey. And we were actually doing the surveys to make measurements of physical parameters that had been discussed and that were making a measurement that those physical parameters required a survey. You are not going to measure the three point galaxy correlation function by taking the spectra of two galaxies. It doesn’t work that way. You cannot do it. If you want to measure that you've got to do a survey. And the trick is to design the survey in such a way that you get a good measurement of the parameters that you are trying to measure. So anybody who tells you that surveys are just butterfly collecting is nuts. Anybody that tells you that surveys sometime butterfly collect is right. Anybody who tells you that you should never do a survey, you know, let me meet 'em in a dark alley. We can fix this problem. Right?
Huchra:Anybody that tells you that you should never do pointed observations, I'll take those guys on in a dark alley too.
Huchra:Because sometimes it's the right thing to do. A piece of philosophy which is probably worth writing down somewhere. I've done it in one or two places. But I personally think that one of the changes that's taking place in astronomy — somewhat slowly, but it's taking place — is that we have left the era in which people get into doing astronomy or open up new fields or whatever because they're applying hither before never-used techniques. We have come close to filling up the whole observable electromagnetic spectrum, and there are still some things at the edges. And there are still some places where the Earth's atmosphere really gets in the way. There is still some work to be done in the sub-millimeter for example. But we now have a pretty good idea of what things look like over fifteen octaves of wavelength or frequency. Maybe thirteen, I don't know, but a lot. We could go deeper in some and all this other kind of stuff, yeah, but basically we are not going to be generally opening up new fields by this example of somebody coming in from solid-state physics with a new technique who had no background in astronomy. We are going to push on things by having astronomers talk to the solid-state physicists about building the next generation of detectors. What you are going to see more and more — and in fact most of what's going on these days in the field of astronomy and astrophysics — is that people are becoming problem-oriented rather than technique-oriented. So they will work on observational cosmology and they won't be proud about what they use to solve the problems that they're trying to solve. I am a little proud of myself. If there's one thing I'm proud of myself for, it's not the individual discoveries — yeah, those are good too — but it's the fact that I think I realized early in the game that what counted was the problem. And the object of doing astrophysics was using every technique available, every mathematical method, every observational technique, every theoretical insight to solve the problem. And it was okay to go off and make observations in the infrared or in the radio or talk to the theorists in Germany or England or whatever. If you were doing that to approach a problem. If all you wanted to do is say, "Oh, I built a new camera. Let's see what we can look at with it!" That's the butterfly collecting. Yeah. But, "Oh, we really need this kind of observation to answer this question and Sam over there knows how to do that and Sally over there knows how to do that, let's all work together to solve the problem by approaching it from different ways," good stuff. And I try to get my students to do that too.
McCray:When did that begin? When did you begin to notice that shift taking place?
Huchra:Well, it started — I mean basically it started twenty years ago. It's been happening more and more. Let's put it this way. If you at the decadal survey, which I have a copy of somewhere around here, it's still pretty much organized based on techniques.
Huchra:The next one won't be. The next one will be organized based on scientific problems, you know, areas of research and not techniques. And I — I mean, if I have anything to say about it, and I'm going to keep trying, but I think that will be sort of the key breakpoint when that finally happens.
McCray:You've been on a whole variety of committees throughout your career. I'd like to ask you just about one and that was the 1990 Bahcall, John Bahcall-led decadal survey. I just wanted to get your thoughts. You were on the optical infrared panel.
Huchra:Right. It was chaired by Steve Strom, vice-chaired by Sidney Wolff and I think Wal Sargent were on the committee as the vice chairs. I don't know how much you know about decadal surveys and the structures, but the way they have been structurally set up in the past is to have a panel, a main panel, the members of whom end up being the vice chairs of the sub-discipline or the disciplinary panels, the chairs of which are people selected also from the community. And how would one put it? That seems to have worked reasonably well in terms of an operational structure. And then the vice chairs can report back to the main panel who will try to integrate the results and recommendations of all the individual sub-panels. So I was on the optical one, infrared ground-based panel. There was also an O/IR space panel at the time.
McCray:What were the major issues that that panel was grappling with?
Huchra:Really talked about two or three different things. The biggie was of course big telescopes. And in fact there were a couple things that happened I think that were very key to the development of this area in the next decade. One was that the kick-off meeting for that panel was actually set to coincide with a meeting in Coeur d'Alene, Idaho organized by AURA to talk about 8-meter telescopes. And this is, I think it was the summer of 1989.
Huchra:But basically that was the meeting at which the NOAO was going to try out the idea to the community at large that we should be building two 8-meter telescopes instead of one 15- or 16-meter telescope. That discussion had been going on for a couple of years, but if you look back at the history and you look back for example at the field community report, there they talked about the NNTT, the National New Technology Telescope. And all of the designs for those puppies were 15 meters, give or take change, depending on how you wanted to do it. And there's this wonderful book with sort of Russell Porterish drawings of all the possible designs of 15-telescopes. I don't know if you've seen it.
McCray:Yes, I have.
Huchra:It's really fantastic. On the other hand, when a bunch of us went through the exercise in the middle of the 1980s, there was a committee that was put together by NOAO called the Future Directions Committee, the NOAO Future Directions Committee.
McCray:Strom also chaired that one.
Huchra:Strom chaired that one. He was asked to do it by first John Jefferies and then Sidney Wolff when she transitioned over to be the NOAO director. And that would have been around, again, 1984. You can check the dates.
McCray:'87 is when that —
Huchra:Wow. '87. That's right. Jefferies left around '84-'85, stepped down [1987 actually]. Anyway, the main result that came out of that committee was this idea that perhaps going straight to 15- to 16-meter was biting off more than we could chew and that it might make more sense to build a couple of 8-meter telescopes and it might make sense to try to optimize them in some way as opposed to building everything 8-meter general purpose. And that idea was trotted out at Coeur d'Alene with the O/IR panel in attendance. And it was the big discussion.
McCray:How did people react to it?
Huchra:I think most people reacted positively. Part of it is because there is always this great fear of jumping up a factor of four or so in size. Part of it was that budgetarily it might make more sense. Part of it was that when you approached the idea of building something that was smaller, suddenly other players could possibly come into it. I was in those days — I mean in '89 I was much less connected to the broader community than I feel I am today. Today I would say I'm one of the people the astronomical community at least that's probably closest to the pulse of what's happening in this country, and even worldwide — from serving on all these committees, from having lots of friends in lots of different places. In '89 I was still pretty wet behind the ears in the general political scheme of things. There are a couple of other things that got discussed. Archiving. Never made it very far. There is a reason for that. It's slightly sad, but it's true, archiving never makes it anywhere because it always ends up being competed against something else where the something else is always sexier. Shall we build a new $5 million instrument or shall we put a million dollars a year for five years into archiving the data we're taking? We know the answer to that. It's always going to be built a new $5 million instrument. Astronomers are greedy. "The data will take care itself. Besides, we can stick it in a shoe box and put it in our desk drawers. We don't want anybody to see it anyway." The other thing we talked about was surveys, and apropos to the discussion we were having yesterday, remember that by 1989 we had already posed to do the equivalent of the 2MASS survey from space and had already been turned down. So we had already started percolating the idea of doing 2MASS on the ground. All right? It was already in my mind. So one of the things that I pushed for — I didn't have to push, nobody had to push hard for the 8 meters, but one of the things that I pushed for hard in the 1990 panel was the idea of doing some digital sky surveys from the ground. I guess Gunn already had his ideas about the Sloan survey. I pushed for the infrared. And if you go back and look at the panel reports you will in fact see that. An infrared survey from the ground.
McCray:These are the working papers from the 1990 survey.
Huchra:Yeah. And also the summary. What do we have here? Under small projects on the ground there were three that were singled out. I've got to find it. Sorry. Small programs. Here we go. If you look under small programs, there it is.
Huchra:So I managed to get the 2-micron survey written into the decadal survey. That was a political, or perhaps a scientific triumph for infrared instrumentation. Various things like that. But of the stuff that wasn't technology development, which was an actual real live thing that was the first of them.
McCray:Okay. One of the big recommendations that came out this which didn't seem to get very far — I mean, the major recommendation was actually rebuilding American infrastructure in astronomy.
Huchra:And that never gets very far.
McCray:Is that because of the sexiness aspect that you just mentioned?
Huchra:Yeah. It's like archiving. The other thing is — I mean there's a long story there and boy, this is a long and sad story, and I have unfortunately gotten to know it too closely in the last couple years. The problem is that people think different ways about how to grow something or how to increase something or whatever, and it's not that the different ways are either right or wrong, but they do have different side effects, and sometimes the side effects are big. You hear about the NSF budget growing. Those of us who have been involved in looking carefully at it and also in trying to advocate some of this growth and things like that will tell you that a really large fraction of the growth of the NSF budget has come by adding programs. It's not come by increasing the NSF budget to do more things, comma, in the same range of things that they were already doing; it's come by adding new things. Now you can grow budgets by adding new things. And sometimes the new programs are related to what you've been doing in the past, so there can be some migration. Sometimes the new things are completely orthogonal. Because the NSF back a decade and a half suddenly was given this mandate to work on K-12 education. A really large fraction of the growth of the NSF budget has been due to the growth of the education aspect. It's almost a billion dollars this year. In the 2003 budget we're looking at — well, I don't know what the number is going to end up being, but the thing I remember from reading the news things this morning is $983 million for K-12 education in the NSF budget. Well, in 1980, the NSF wasn't funding anything in K-12 education, and $980 million is a quarter of the NSF budget. So yeah, right. Infrastructure, which is fixing what you've already got, doesn't sell. Or the agency doesn't think it sells. If you can tie infrastructure to very specific things — the Greenbank radio telescope falling down and you've to fix it — then you can get something, but it's often through a process of earmarking, which is not entirely — not considered good form.
McCray:Now it's the Robert C. Byrd telescope.
Huchra:Now it's the Byrd telescope. That's right. That's right. Well, it always was a Robert C. Byrd telescope. Now it has a plaque. So it's that kind of business. The sad part about this is that — I mean how would one say it? You can do, you can look at the demographics, and this where it gets tricky. The size of the budget, sort of the base funding budget for research and development at the NSF has grown a little bit faster than inflation, but not much. I mean if you take away all the extra programs and you just look at what is the basic line or physics or the basic line for astronomy, and you don't count the new nanotechnology initiatives or whatever, those budgets have actually been flat before inflation, maybe even have gone down a little due to inflation. Well in the case of astronomy the number of astronomers in the last decade between 1990 and 2000 grew by 20 to 30 percent. So not only did the cost of keeping those astronomers grow by inflation, but the number of people doing the research grew because there was so much interest in the field. And one would think that in a rational system there would also be, there is not only an adjustment for inflation but also an adjustment that reflects the interest of people working in the field and numbers of people, demographics. And there has never been. So essentially speaking, the funding per astronomer — and especially at the NSF where budgets have been flat — has gone in real terms a lot by this demographic factor. And that has really caused some damage to the field. And that was the kind of thing that we were trying to fix with the recommendations about improving the infrastructure, improving the resources for the grants programs, making sure that the times for the VLA tracks don't fall apart. They're all tied together in that sense.
McCray:I saw yesterday that NOAO's budget is cut by a million for the coming year.
Huchra:Mm-hm [affirmative]. And who knows what's going to happen at NOAO. Don't ask me that question.
McCray:Because you don't know or you don't —?
Huchra:I'm conflicted. I'm chair of the AURA Board right now, and NSF still hasn't announced who is going to operate it next year.
McCray:Okay. So for the Bahcall committee in terms of the major recommendations that came out, I mean the big ones from your panel were the infrared optimized northern telescope and a more general purpose southern telescope. Can you give me some sense of how the community was divided in terms of opposing or supporting those ideas?
Huchra:Ha! Well, I mean you know — the community divided according to lines that were completely different from what you might expect in any sense. It became a political issue, and the division came about not because of north versus south or infrared versus optical; it got divided by virtue of the fact that the NSF — Erich Bloch was the head in those days — the NSF essentially said, "You can have one, and you can have two if you find partners to pay for the other one. Then you are allowed to split them." And the NSF said it, Congress said it. It got backed up in Washington by the legislative branch as well. So then the issue became should we have just one which the U.S. controlled lock, stock and barrel — and various people took up that flag, I think including John Bahcall.
Huchra:And or should we have two and relinquish some control over this in order to cover both hemispheres and perhaps be able to leverage a little bit, because two is likely to be cheaper than two times one. You get the general idea.
Huchra:And that was a debate that went on in the community. I personally didn't care. I wanted something to happen. I thought building an infrared optimized telescope would be a really good thing and if we managed to get a second one that would be okay too. I was a little disappointed three or four years down the line when we did go through this whole business of splitting Gemini off from NOAO and all the other stuff that went with it and when the U.S. essentially did relinquish control of the project.
McCray:Relinquished how? I mean they are still 50 percent partner.
Huchra:[makes a "raspberry" noise as to mean "poppycock"]
Huchra:[repeats the "raspberry" with equal vigor]
McCray:So you were raspberrying the 50 percent partner aspect.
Huchra:Right, right, right. A couple things. One is that the great country of the State of Hawaii for practical purposes is not a member of the United States, and the voting rights associated with that have been lost to the U.S. in this game. The way the administrative structure has been set up, the U.S. has less than 50 percent of the share, so it cannot force an issue one way or the other. And the other side of the coin that also hurts in this game — it's not the other side of the coin, the other side of the issue then also hurts in this game is that the way the NSF has set this up — and I'll be extremely critical of the NSF here — is that essentially no one with any interest in large telescopes and in operating large telescopes and in the national observatories can sit as U.S. representatives on the Gemini Board, which incredibly weakens the organization from the point of view of American representation. The UK representatives on the Gemini Board tend to be people with teeth coming out of their ears; the U.S. representatives on the Gemini Board tend to be people that were pussycats and have never been very assertive when it actually comes to trying to push on things. It's a generalization. I mean there have been strong people, but one strong U.S. rep out of four or five is not, can't fight the tide. And also they rotate them every two years in the U.S. — not in the other countries — so that nobody ever gets to build up the history.
Huchra:And the NSF thinks this is fair. In reality it may be fair, but it's also incredibly weak and the United States has not done well in terms of what it could be getting, it could be doing in leading the Gemini project.
McCray:What's an example of a situation where the U.S. had interests in the project going in one direction and the international partners were able to force it to go another way?
Huchra:I'll give you something worse than that. I'll give you an example of where the U.S. astronomers had an interest in the project going in one direction and the National Science Foundation and the international partners forced it to go in another direction. And the example was a couple years ago — hell, how long ago was it? We must be talking back '97 or so. It looked as if one of the partners in the south, Chili, was going to default and pull out and they were going to open up a share. We could have bought that share. A complicated deal. There was still some left, and instead of the U.S. buying that share, or instead of the consortium closing and stopping and saying, "Okay, we have enough money. We can finish," we let the Australians in. Now the Australians didn't contribute large sums of cash. The Australians bought in for future contributions of instrumentation and a small sum of cash. We could have used that share and we would have gone back to being 50 percent partners. But somehow the agency and the U.S. members of the Gemini Board wimped out. Right?
Huchra:Strictly from a political sense, we could have used the extra observing time.
Huchra:Something that people don't realize is that under the current scheme there is actually a hundred nights of telescope time per telescope per year to be allocated to U.S. astronomers. That's it. That's all she wrote.
McCray:That doesn't sound like half of 365.
Huchra:Well, first of all it's never 365. There is always engineering time that gets taken off the top. But then the Chilean 10 percent in the Southern Hemisphere gets taken off the top. The Hawaiian 5 percent in the Northern Hemisphere gets taken off the top. By the time you're left to what goes to the U.S., it's about 40, 41, 42 percent — and it's 41-42 percent of what's left over after the engineering time.
McCray:And the weather conditions.
Huchra:No, no. This is a hundred nights. This is not a hundred clear nights. This is a hundred nights allocated.
Huchra:A hundred and six I think, or 110.
McCray:So roughly eight hundred hours, nine hundred hours, right?
Huchra:Right. It is an incredibly scarce resource. The number of people who will apply and who are capable of applying for time, in the sense of being optical and infrared ground-based observers. A couple thousand, 2500 say.
Huchra:Yeah. You get the general idea. Most people aren't going to get to use the Gemini telescope and it's a real pity. Whereas if we had a bigger share and perhaps more control over how things are split up, we'd be better off and it wouldn't have cost us that much more money.
McCray:Later on you served — we talked a bit a lunch yesterday about your time on the committee that reviewed the choice of what mirror to use of Gemini and also you were chair of the committee to pick the director of it. Without repeating some of the things that we talked about off tape yesterday was there anything that you would like to add to this oral history about those topics?
Huchra:Well, some of the things. I mean, the mirror review committee was essentially given an almost impossible task, and if you go back and look at the records when the records can be shown, the bottom line, the most important thing that that committee decided to do was not to stop the project. And they could have jumped up and down and yelled and screamed and told the NSF that they should cancel everything and stop.
McCray:It could have stopped the project.
Huchra:Well, it could have tried to stop the project, but it didn't. And if it had tried to stop the project, it might or might not have succeeded, but it would have been messy — really, really, really messy in the community. So it didn't do that. In terms of the decision that the committee made, the recommendation that the committee made, I mean again I can't talk to you about all the gory details, and especially some of the contractual stuff, but from a strictly scientific point of view, engineering and science, NOAO had made the wrong decision. In other words, the honeycomb borosilicate mirrors were possibly a better choice. Whether they are or are not — and I use the word "probably" because ain't nobody had built one at that time, right?
Huchra:Of either kind. Okay. Whether they are or are not will be borne out over the next couple of years as actually both types of mirrors get used to do different things. We'll see how well they work and what the feebles and the fumbles are. From a contractual point of view in terms of where the money was and the controls that were in place — which is not what that committee was asked to do, but we did see some of it and we made the right choice. Okay. Working with the University of Arizona, or working with any university in the way universities are set up is very different than working with industry. With industry you have some assurance of getting what you pay for; with a university you have much, much less assurance of doing that. Time scales, money, contractual agreements, the whole nine yards go that way.
Huchra:I probably at the time would have gone with the U of A anyway. I was on the committee, and that's what we did vote to do, did decide to recommend I should say.
McCray:Why did you favor that?
Huchra:Yeah. Today I'm not sure it matters. It was ten years ago, water over the dam — hopefully it's water over the dam. I've drunk enough Scotch with Jim Oschmann so maybe he doesn't feel so bad about that decision that was made — you know, I should say the recommendation. The committee never decides anything, the committee recommendations things. The recommendation that made at the time. Okay. A lot of the people that were involved in it on either side are now retired or left or whatever, and there were a lot of mistakes that were made that were mistakes in presentation as opposed to mistakes in real concept. As I mentioned yesterday, in some of the cases the review committee would ask question A and we'd get answers to question B which is not what question A was, and that didn't go over well. And some of that was people were wet behind the ears and didn't understand what was going on or what the major questions were, and that made a lot of things seem as if they should be the other way around. On the second issue, on the second committee, the search committee for Gemini’s director, I will say that I'm extremely proud of the choice we made. It was not without its risks. Matt Mountain was definitely young, wet behind the ears, not American — which was not an inconsequential factor in thinking about this.
McCray:Was there a feeling that the British were taking over the project?
Huchra:Yes, Yes, yes. So it was in some sense controversial for essentially a U.S., or I should say a North American-based committee to recommend that yet another Brit come into the project. We had a lot of problems with NOAO itself. Sidney did not want to give up being director, and she at times actively was trying to torpedo the work of the search committee. She was not very pleasant, and you can probably find witnesses to some of the shouting matches she and I had in this process. Because she was at the time the Gemini director and the acting NOAO director, or maybe it was the other way around: the acting Gemini director and the NOAO director. I think that in fact it was right to split Sidney off from that job, but I also think it probably would have been possible to find somebody who could have done both and satisfied all the partners — which she could not have. There was just too much of a conflict of interest.
McCray:What did Matt bring to the project that was attractive?
Huchra:What Matt brought to the project that was attractive was a real appreciation of management in engineering, and that is not something that a lot of people have.
McCray:People or astronomers?
Huchra:People. Come on, now. Right? Yes, astronomers, but astronomers are people too.
McCray:What I meant was, is it common to find many astronomers who have an appreciation of the engineering and project management aspect.
Huchra:No. I think that's true across the sciences. It's less important in some of the others. If you are running a biology lab where you're playing around with tissue cultures you don't really have to worry about project management, but in big physics projects or in NASA for example. So I mean there are lots of astronomers who work the space side of things who are extremely familiar with project management techniques. So it's not astronomers in general, you know, it's those who don't have to play the NASA game that don't have the appreciation for those kinds of things. And Matt did. And that was a real play. That was a real plus in the game. The other thing that he brought was a tremendous amount of enthusiasm. He had a fire in his belly. That's the best way to describe it. What's that guy's name, the writer?
Huchra:Robert Bly. The fire in the belly. That was it. Matt had it. So he had — even though he was extremely young compared to any of the other candidates, and inexperienced in terms of leading projects, he came to us with that background. Had built and instrument, had an appreciation at least that we could see of project management and engineering techniques and a real desire to make it happen. And that was the winning plate. We were still scared. I mean frankly, I as the chair of the committee was still scared that it wasn't the right recommendation, but there was little doubt that — there is little doubt now that that was the right recommendation.
Huchra:I've followed the careers of the other candidates. Trust me, it was the right recommendation.
McCray:Okay. I have one last question and then we can take care of some of the details of finishing this up. Right before I left I saw David DeVorkin and he noted that at the last AAS meeting you chided him for having the Newtonian mirror cage — I guess from the 100"? — on display at the Air and Space Museum. You noted that one could still do good science with it. He'd asked me to get official comments on that.
Huchra:Okay. Well, the 100" at Mt. Wilson is a rather interesting telescope. Okay? First of all, I'll be heretical and I will tell you that of all the telescopes and sites that I've worked at, the absolute best natural seeing I have ever seen is at Mt. Wilson. "Natural seeing" being the image size as left over by the turbulence in the atmosphere. Okay. In that sense it is probably one of the most superb sites that astronomers have ever come across. And it's a real pity that they had to go build that city right next to it. Okay. Two, there's still a heck of a lot of science that could be done in using a telescope like the 100" at Mt. Wilson. I'm not saying that it should be done using the 100", but it's there, it's close to a city, it can be used as a teaching telescope. I did my Ph.D. thesis pretty much on the 100", and most of the data I had came from the 100" at Mt. Wilson. Nobody else wanted to use it, so that was good. And to some extent I was also just chiding David for walking away with valuable but still useful historical artifacts. There are things that could be done with the 100" if people wanted to.
Okay. That seems like a good place to pause. I have some other topics I would like to cover but we will save those for the next time. Thank you.
 “A Slice of the Universe”, deLapparent, Geller, and Huchra. Astrophysical Journal Letters; vol. 302, March 1, 1986; pp. L1-L5