Martin Harwit - Session III

DeVorkin:

This is Tape No. Three, Side No. One. The date is now June 22, 1983, but the conditions are as before. This is David DeVorkin interviewing Martin Harwit in his office at the Air and Space Museum. When we left off last time, we were beginning to talk about the NASA airborne program. Could you give me a summary of how you became involved in the program, how you first heard about it and how you gained access

Harwit:

I don't remember where I first heard about it. I knew that Frank Low had flown on the Lear jet quite a while, and of course around 1970 there was talk about converting a C-141 aircraft. Gerard P. Kuiper was quite involved in that and pushed it quite hard.

DeVorkin:

Hadn't he died by then?

Harwit:

No. He certainly still was alive at a meeting I attended in 1973 or maybe 1974. It was named after him only after his death. At the same time also I had been told that if we wanted to apply for flights on aircraft, we would be welcome to propose, even though NASA was no longer interested in supporting any rocket work .n the infrared. So both Jim Houck and I wrote proposals to do spectroscopy in the infrared from aircraft. He was going to work at shorter wavelengths, I was going to work at longer ones. He was going to work out to about 40 microns, and I was going to try to work from 40 microns onward, out to about 120 microns with Gallium-doped germanium photoconductors. We started building equipment. The beginning was really quite disorderly. There were a number of false starts. One of them involved Carl Frederick, a former student of Bill Hoffmann, who came and worked with us as a post-doc. That didn't work out too well. He eventually worked on a balloon project with us which also had its difficulties. But the first real break in that came in my having a student by the name of Dennis B. Ward, a Canadian student, one of the most promising whom I have ever had. Unfortunately, after he got his degree, he decided that he really wanted to work in Canada, and he eventually ended up going to a library science school. I think he now is active in library science, although I only hear from him from time to time. But he really was a very promising student. He could do the instrumentation. He understood the theoretical part as much as one needs it for observational work and he had very good common sense and also a lot of energy, a very strong combination. He really built up this effort almost single-handedly by just being very energetic. We then started flying on the Lear jet, together with George Gull, who had been an undergraduate at Cornell and then stayed on to become a mixture of engineer-technician.

DeVorkin:

Let me ask you about the process of gaining access to the Lear jet. Would you submit a proposal?

Harwit:

Yes, one had to submit a proposal and then one was funded. Then every year thereafter one had to submit a renewal proposal, and this has gone on now for 10 years or so.

DeVorkin:

So, this is a programmatic type grant that is continuing?

Harwit:

Yes.

DeVorkin:

Did you prepare your own detectors for the telescope?

Harwit:

We did build our own detectors; not everybody had to. If we had been able to purchase them we would have, presumably. But we built our own detectors. We built our own spectrometers and we built the electronics that we used to detect and analyze the signals. NASA Ames provided the 12-inch telescope, gyros to keep it pointed, and then as time went on increasingly sophisticated tracking equipment, which had the advantage of making it easier to track on a source, but at the same time had the disadvantage of adding bulk to the plane, and decreasing the length of time that one could fly at altitudes.

DeVorkin:

You did start then with the 12-inch on the Lear jet?

Harwit:

That's right.

DeVorkin:

Was it the one that we have downstairs, the prototype?

Harwit:

I don't think so, no. By the time we started I think that NASA Ames had already built its own telescope.

DeVorkin:

So the history of that program was that, if I am correct, Frank Low must have built that original prototype.

Harwit:

I think that's right. I wouldn't swear to that. By the time I got into flying on the Lear jet he already was out of it. I asked him once why he had quit. And his comment was, "too damn fat". Evidently he had gained a lot of weight and maybe his doctors told him that he shouldn't fly. I didn't want to probe him further, because it must have been painful for him to quit this type of research which he had originated after all.

DeVorkin:

For health concerns?

Harwit:

I don't know whether it was health constraints or what. There is also a limited amount of weight that can be flown on the plane.

DeVorkin:

Is he an especially big man?

Harwit:

Yes, he must weigh about 300 pounds, I think. He is quite tall. You hardly notice it really, but maybe it is only 250. It must be more than that, because George Gull flew and he's about 250 pounds. He's 6 foot 4. Frank Low is not quite as tall perhaps, but certainly heavier. Yes.

DeVorkin:

Since you have been here this semester, your life is very much constrained, or seems to be constrained by the flying schedules. Has it always been that way?

Harwit:

Not quite as much perhaps. When we started out we were only flying on the Lear jet and not also on the Kuiper Observatory, and so there were only two sessions a year perhaps that we would go out for. Although at that time I also had ground-based observations at Kitt Peak, that we were running.

DeVorkin:

What were you doing at Kitt Peak?

Harwit:

Well, at Kitt Peak we were doing Hadamard Transform spectroscopy? I can come to that later. But what both Jim Houck and I decided to do was to do spectroscopic work. The airplane seemed to us to be an ideal place to get that started. Other people were doing photometry. I had, as I mentioned last time, become interested in spectroscopy from the days that I had worked at Michigan as a graduate student, so I knew something about it and I had always felt that was the way to try to go in astronomy. Of course, I thought that was how one would do chemical analysis ultimately because I had seen that the laboratory chemists identified compounds by means of infrared spectroscopy. Then it came as a surprise actually that the radio people beat us to it with microwave techniques. They really had an edge on us, a big lead, in the molecular range, except in such things as diatomic hydrides; that is, diatomic molecules which have one hydrogen atom in them. Those molecules tend to not have one hydrogen atom in them. Those molecules tend to not have radio spectra, but have their dominant spectra in the far infrared, HC1 and a number of others. So that seemed like a promising thing to do. Also the atomic and ionic species have fine structure lines. There was a paper that had been published around 1969-1970 by Vahe Petrosian, who had been a former graduate student at Cornell of Ed Salpeter's, but now was at Stanford as a faculty member; he has been at Stanford ever since then. But he first showed that fine structure transitions of all kinds should provide very big signals. This wasn't entirely new. People like Stuart Pottash and also, I think, Burbidge, Gould and Ramsay had talked about fine structure transitions. But Petrosian really did the first comprehensive survey of different fine structure transitions that might be expected for the Orion Nebula, and actually calculated fluxes which later on, when one started measuring them, turned out to be quite good estimates. So here was something where the theorists had really done a very good job in preparing the field for observers. Spectroscopy just seemed like the right sort of thing to get started in. Nothing had been done in for infrared spectroscopy. So both Jim Houck, who was a solid state physicist by training and understood the importance of such things, and I, because of my earlier background, decided we would like to do that sort of work. So we started on that. It turned out that Jim Houck had a liquid-helium-cooled spectrometer design that he had started on a small grant that he had had at one point. He very much wanted to see that put into use, so he persuaded me to make use of that instrument and build it up.

DeVorkin:

What kind of dispersing agent did it have?

Harwit:

It was a grating instrument, where all the optical components, including the detectors, mirrors, and grating, were to be cooled down to liquid helium temperature. This was in a way a follow-on to the fact that we already knew how to cool an infrared telescope, so none of us was that shy about cooling other types of optical systems. A little bit later Jim also was starting to cool down a grating instrument for use in the rocket payloads that he was then flying with Air Force Cambridge research support. So we got started on that, and initially we had very low spectral resolution, something like eight microns resolution at 100 microns, so only one part in 12. It was very low. We were really only just getting what in the visible part of the spectrum, one would call spectrophotometry.

DeVorkin:

Was that mainly to work at it in a stepwise continuous way, or did you not have the technology for going in at higher resolution?

Harwit:

I think the main point was that initially we probably didn't have sufficiently sensitive detectors, and so we wanted to make sure that we got enough flux on the detectors to be able to make decent measurements. We got some spectra of Jupiter and some other planets, and then also M42, the Orion Nebula, and M17, the Omega Nebula. Of course one of the problems was that you didn't really know whether you should trust any of the planetary sources as calibrators. You had to use some kind of circular argument, where you tried to compare the various planets and the moon, and if they all agreed at some point, you figured that they must behave as black bodies pretty much, because the compositions of these atmospheres were rather different. For the moon you figured it ought to be blackbody, although one does see patchiness in the visible. We did that until about 1975; I think by that time the program was a couple of years old. At that point Dennis Ward said that he really wanted to look for the 88-micron fine structure transition that had been predicted for doubly ionized oxygen in Orion. I was quite skeptical of whether we would be able to see it or not.

DeVorkin:

Because of the predicted strength, or your sensitivity in that area?

Harwit:

Well, because of the predicted strength and our sensitivity. I think eventually when we found it, it was somewhat brIgnter than we had expected, maybe by a factor of 2; and also because we had never seen anything so far that looked at all indicative of anything. We had looked broad band at the Orion Nebula before. He thought that he saw a little bit of an enhancement maybe at 88 microns, and possibly at 63, for atomic oxygen. There were lots of ups and downs, and we thought it might just be noise, because it didn't look that different. This was really something that he pushed very hard. Usually I feed the students problems. I spend a lot of time thinking about what might be interesting to do, and I suggest these things to them. They generally carry them out, although of course with an awful lot of input of their own when they are good students. I can't claim any credit for it at all; this was Dennis Ward's pushing very hard to persuade me that was the right thing to do. I didn't resist it very much, but I was too pessimistic, I think, at the time to think that we were at the point of being able to see spectral features. Then when we did find it, of course we were very pleased, because this was the first spectral line in this sort of whole general range. I think the next shorter wavelength that had been seen was at 12.8 microns, the neon line, and then on the longer wavelengths side people were working in the millimeter range. So for us this was really a big deal. Just as, I think, we were putting this in for publication — or maybe it was in the works, I don't know which — I got a call saying that our program was being discontinued, because the spectroscopy was quite uninteresting and we were really not seeing anything much more than the general blackbody curves that other people had seen. By that time we had looked at the galactic center and with this 8-micron resolution we were sort of seeing a blackbody.

DeVorkin:

Who wrote you?

Harwit:

This was Nancy Boggess. She was in charge of infrared work all the time. I was of course terribly upset. I think it must have been just after the proposals had been judged that we found this line. I don't know. It was just about that time, and I wrote back and said, look, we have just found this line. I think this is really very exciting, to turn us off now would be the height of injustice. I wrote her boss, also, Jesse Mitchell — I don't know whether he was at the level that later became Noel Hinners' office or what later became Frank Martin's office. Eventually they said, okay, they would support me, but they could only give me half the amount of money we had had the previous year.

DeVorkin:

This was in the early 1970s?

Harwit:

No, in fact it was 1976. I was in Bonn at the time. I was trying to do this by long distance phone calls which was rather expensive.

DeVorkin:

Could that have been part of the problem? That you were in Bonn?

Harwit:

It's not unlikely. Usually when I've been away on sabbatical I've lost contracts. People think you are on vacation. On a previous sabbatic leave in Czechoslovakia I had lost all my contracts, and I had to start building up again when I came back. It's bad because you have built up a group and you have to try to see how to support them. Sometimes you have a little bit of money left over from previous years. But at any rate, that was a very tough time for bridging over, and fortunately Jim Houck at Cornell helped out by sharing costs of technicians and things like that. But it lasted long enough so that internal strains actually developed within the group, which then took a number of years to heal over again. When one person depends on another one, it just is too tight quarters, and that's not very good. But at any rate, the next year we proposed again, and I think within two years they had doubled the amount of money they had given us originally. There was this fluctuation up and down.

DeVorkin:

Why was that?

Harwit:

I don't know. After about seven or eight years of that I really got fed up, and a year or two ago, I forget which, in the proposal for the Lear jet I told them they really should make up their minds. One of the problems was that some years they would decide they ought to retire the Lear jet, and then they would decide, well, maybe they shouldn't depending on what the philosophy was that year of the advisory committee, the MOWG, the Mission Operations Working Group. They advise Nancy Boggess once a year. They would change their mind, and one of my proposals then included a bargraph showing the ups and downs of the funding that we had had for the previous 10 years, and complaining that it would be nice if they could make up their minds, because it was just too much of a strain. It was easier to work on a level budget that was low than to work on the same average budget, or even somewhat higher budget when it fluctuates up and down. You just can't work that way easily.

DeVorkin:

Was it about that time that you started thinking about writing COSMIC DISCOVERY, in 1976?

Harwit:

Yes, I started then.

DeVorkin:

Is there any relation?

Harwit:

To that? Possibly, yes. I don't know. When ycu have worked in space work for quite a long time, you see people coming in and saying, do this, don't do that. You try to get whatever money you can get, and then both satisfy the funding agency and do interesting things, which often are quite different from each other. That is tough. I wanted to see at what point in the planning and granting procedures of these funding agencies some of these fluctuations and more blatant difficulties arose. That was one thing. I think the real seed for COSMIC DISCOVERY was quite different, though; namely, at one point I had started thinking about whether one could decide how much remained to be done in astronomy. I had given some popular talks on that, because I thought it was an interesting question. The first talk of that kind that I gave, I think, was directed at Cornell alumni. That must have been in the late 1960s.

DeVorkin:

That early?

Harwit:

Yes. Then I started writing some of that up, and at some point in there I realized two things. First, I started drawing up diagrams that illustrated the various capabilities of instrumentation to show what had been done and what had not been done.

DeVorkin:

These were similar diagrams to the ones that are in your book?

Harwit:

That's right. Those are in fact the exact same ones, except at that time I just made them refer to something like 1970; I hadn't drawn in the historical development. The diagrams I had in COSMIC DISCOVERY showed the state of the art in 1939, 1959, 1979, and also superposes the different discoveries that were made possible by these techniques. That wasn't in the earliest diagrams. The earliest diagrams appeared in a 1975 article in QUARTERLY JOURNAL OF THE ROYAL ASTRONOMICAL SOCIETY. That article was actually written sometime in 1972 or 1973. It took me about two years to get that published.

DeVorkin:

Why was that?

Harwit:

Nobody wanted to accept it. Eventually I got to the point where I would send it to an editor and ask him not to send it out to referees, if he was pretty sure that he wasn't going to accept it anyway. I told him I'd be willing to waive that, but I didn't want them to go through the posturing of fairness only to reject it after the refereeing. It was interesting. I gave a talk at Harvard — I think it was 1974 sometime — in which George Field, who is somewhat of a wag, introduced me and gave some of the things away. He said, I bet you he's going to say such and such, and of course I did. So at the end I said, I'm really impressed. How did you know I was going to say that? He said, oh, it's easy. I refereed your article for such and such a journal (laughs).

DeVorkin:

Did you ask him that publicly?

Harwit:

No, no, that was afterwards. We were walking down the hall, and he said, I refereed it; I gave it a very good critique. I know I sent it to PASP. They sometimes have lead articles which are also of a more general nature. I had also sent it in, I think, to Foundations of Cosmic Physics, the Journal that Al Cameron edits, at Harvard, and various journals that I thought might take it but none of them would. Eventually somebody suggested — I forget who — that Quarterly Journal might be the right place to send it. That was a very good suggestion, because they accepted it right away. But I don't know whether I've spent a paper to as many different places; and yet it's turned out to be one of the few best sellers of my own work.

DeVorkin:

Do you know that through citation studies, or from what people say?

Harwit:

No, actually, I don't think very many people cite it.

DeVorkin:

It must be through what people say?

Harwit:

It's just that people have read it and talked about it, but not actually used it in their work. The first time I ever found anybody who knew about that was in 1977, I think, after I had come back from a sabbatical in Germany. I ran into Frank Martin at a party that Nancy Boggess had given for some committee of hers that I was on. He said, oh yes, he had read that paper, and all about the baseball cards, which illustrate the statistical process. So I asked him how he knew about it. He had heard about it, I think, from Harlan Smith, who had something to do with the Space Science Board. It turned out that Frank Martin actually was using the book. At the time he was in charge of advanced projects at NASA. He had asked Harlan Smith whether there was anything around that would guide one towards what would be an advanced project in astronomy, what sort of instrumentation NASA might put up for the next generation of astronomy experiments.

DeVorkin:

Beyond space telescope or something?

Harwit:

Yes, Frank Martin used to publish each year, or every two years an updated version of instruments that NASA was thinking of bringing on line in the future. Each of these had a different color. You might have seen than. They look like the rainbow. I don't have any of these here. There may have been more than 15 different such instruments that had interferometers flying in space, gravitational wave experiments, all kinds of equipment. He wanted to know which things to put into that class, what kind of instruments. I guess Harlan Smith must have suggested to him that he look in the book, because the blank spaces on the plots that I had on these charts were the areas where one hadn't done any work yet and hadn't made any discoveries.

DeVorkin:

Yes, I know Frank Martin was very excited about it. Harwit. He really spread the idea then. I think one needs to have someone who peddles such things, and doesn't have a vested interest. I think with Frank Martin being interested, and the people at Headquarters hearing his presentations, that it got quite a lot of coverage. That really started the book. I think I got encouraged, by giving talks on the subject myself, to then become more interested in the history of these discoveries. People would occasionally come up after talks and be very upset, or get up in discussions and claim that I completely short-changed theory, that theoretical predictions were what led to advances, and also to instrumentation that we were using. I said that that was nonsense, that it was simply military hand-me-down equipment that was being used for these discoveries. And unless the generals were theorists in disguise, there was no way that theoretical influence could somehow enter all this.

DeVorkin:

Who were the people who came up to you? Do you remember anyone in particular?

Harwit:

I remember one talk that I gave at Groningen in Holland where the community as a whole was almost screaming at me, especially one graduate student, by the name of M. Nepveu whom I later met in Bonn. I think the Dutch are very orderly people; and it bothered them that one should approach this problem in a kind of dirty way by just throwing, at random, instrumentation into the pot, and making discoveries that way. It wasn't the kind of thing, for example, Jan Oort would suggest. Of course they had made one of the few good predictions in astronomy that did lead to an important discovery; namely the 21 centimeter work that Van de Hulst and Oort had nurtured very early on. So they had the idea that discoveries come as the result of theory. They gave a number of examples of that.

DeVorkin:

Did it influence you?

Harwit:

Yes, very much; I then decided that if I came up with my set of examples and other people came up with counter-examples, then what I ought to try was to do an exhaustive analysis of the whole thing. I then selected all the major phenomena that I knew about, the 43 that I have in this book of mine, and then started looking systematically at how the discoveries had been made, who was involved, what role theory had played, what role instrumentation had played, and what role military equipment had played, things of that kind. That took me about two years. So from 1976 to about 1978 I was just working on Chapter 2 in my book, which is the historical part. At the same time also, I started looking at the Greenstein report and finding that a lot of things that they did really just didn't sit well at all, not only with me, but also with the book by Edge and Mulkay, which I then started looking at, and found not only fascinating, but in radio astronomy at least, agreeing with everything that seemed to me to be true on a much broader scale, across all of modern astronomy. Or maybe I should put it the other way around; what they found to be true for radio astronomy I could also discern in all of these other disciplines, which I had now looked at by virtue of my having tried to understand how these various discoveries had come about, and doing these thumbnail histories on them. A lot of the things that I did then, in COSMIC DISCOVERY really were generalizations of things that Edge and Mulkay treated. For example, they have a beautiful quotation from one of the people that they interviewed, who says, "the nice thing about radio astronomy in the early days was that the people who built the instruments and the people who used them were at one and the same place; and in fact were even the same people."

DeVorkin:

They were talking about Britain alone at that point.

Harwit:

Well, no, I think most of the early radio astronomers built their own equipment, receivers at least, and things like that.

DeVorkin:

Even in Holland?

Harwit:

Yes, well, I think so, yes. They had C.A. Muller, who built up their technology. In contrast, Edge and Mulkay's anonymous scientist says, "The optical astronomers will take a piece of instrumentation, carry it up on a mountain, and then years later come down and say, now we need one just like it, only larger". Sure enough, when you looked at the Greenstein committee recommendations, they were all one after another asking for a larger telescope, each discipline for itself.

DeVorkin:

Each sub-discipline in astronomy?

Harwit:

Yes. And so it became clear, if you were a Greenstein or a Field, and you chose your membership on the committee, one member to represent each part of the astronomical community, then you didn't have to go through all of the fancy meetings and everything. All you needed to do was to look to see what community this man represented, and then write down, larger telescope for infrared, radio, x-ray, whatever it was. It seemed to me that was really so much at variance with what my thumbnail histories had shown; namely, that the biggest instrument of its kind or the most expensive of its kind rarely have a.ny bit influence on what discoveries take place. It's more high spectral or angular resolution, or time resolution that makes the big difference.

DeVorkin:

Both of those come from collecting more light.

Harwit:

That is true, but it just turns out when you look at things, that the biggest telescope in the world at the time — unless it is the only telescope, such as the first x-ray telescope or Galileo's telescope — for one reason or another, doesn't seem to be the one that's used for these discoveries. It may just be that people, once they have such a big telescope, want to be very careful in how they allocate time, and so they only do it for sensible things, and sensible things never lead to discoveries. They do lead to good analyses, but they don't necessarily lead to the really big fundamental discoveries. At that point I decided I really ought to go into the policy issues and do a fifth chapter in the book that would deal with all of that. So I had a second chapter written and a fifth. The third chapter dealt with these spaces which showed what the instrumental capabilities were, and showed that that was the full space that one needed to consider; that because of the physics of things and astrophysics of things, there wasn't an infinite number of possible observations one could make, but a finite enclosed set.

DeVorkin:

Had you read information theory by this point?

Harwit:

Oh yes. I knew a lot of information theory, which I had learned through the Hadamard transform spectroscopy I was doing at the same time. In fact, I started writing the book on Hadamard transform spectroscopy at exactly the same time as COSMIC DISCOVERY. I was doing the Hadamard transform work together with Neil J.A. Sloane. Hadamard transform spectroscopy is named after Jacques Hadamard, a French mathematician who did for discrete transforms what Fourier had done for continuous transforms. He was interested in matrices of numbers all of whose elements were either plus 1 or minus -1. It turns out that those matrices allow you to transform a digitized array of values into a transformed set. One was able to make use of this in spectroscopy. If one had a grating instrument and, say, a single entrance slit and one displayed the spectrum in the exit plane of this grating instrument, much like one would in a spectrograph — but one didn't have a photographic plate as one doesn't have one in the infrared — then what one could instead do was to place a mask in the exit focal plane half of whose slots were open and half of which were closed in some sophisticated array.

DeVorkin:

What determined that array?

Harwit:

That array was determined by the sequences of plus ones and minus ones in one row of these Hadamard matrices. A Hadamard matrix would be a matrix of n rows and n columns, and each element in each row, and each element in each column was either a plus one or a minus one.

DeVorkin:

These were not symmetrical?

Harwit:

These were not symmetrical. what one tried to do, what Hadamard did was to make each row orthogonal to each other row. If you took row one and row two of this matrix and you multiplied the elements by the second element in that row, and so forth; and you summed all the products, then the sum of that product would always be zero. This is a dot product, a scalar product essentially, and it acted as though all the different rows in this matrix were orthogonal to each other, because if you take two orthogonal vectors and you take the scalar product, you also get zero. Now, in the Hadamard matrices you get the scalar product of any two rows or any two columns being zero. It turns out that that particular sequencing of ones and minus-ones corresponds to a useful sequencing of open and closed slots in a mask that you would place in the exit plane of the spectrometer. Now, of course, what you need to do, as in a Fourier transform, is if you want to look at a spectrum that's got N elements in it and colors and wavelengths, you have to have N slits that are either open or closed. In addition to that, you have to have N different masks. What you would do normally would be to put in the first mask and collect all the light that went through the open slits. Then you would put in the second mask and collect all the light that went through its open slits, and so forth. When you then looked at the sequence of signals that your detector recorded you could then go through an inverse transform, again, which involves a matrix multiplication. You could transform back to obtain the original spectrum of radiation that had been displayed in that plane.

DeVorkin:

Are you talking about actual mechanical masks?

Harwit:

Yes.

DeVorkin:

This was nothing that was digitized, and then you had various filters?

Harwit:

No. These are actual physical masks.

DeVorkin:

So your observation would be to place sequentially these masks in the spectrometer?

Harwit:

Well, this is what Marcel Golay, Swiss-born, I believe, a very ingenius physicist, suggested around 1949. He actually built such an instrument at the time. But one didn't have sensitive detectors then. One also didn't have the computers that would make a big mask possible without a great deal of hand calculation.

DeVorkin:

By a big mask, what do you mean?

Harwit:

Maybe 200 elements, something like that. We eventually built instruments that took 255 spectral elements and 255 slots in the mask.

DeVorkin:

Okay, let me turn the tape over.

Harwit:

I will show you what these look like.

DeVorkin:

In a sense, the positions of the little masks and the filters were determined by whether there was a plus-one or minus-one in there.

Harwit:

That's right.

DeVorkin:

In no way were they a direct analog of the positions of the spectral lines or anything like that.

Harwit:

No, none at all. In fact, it had nothing to do with the spectral lines. So you want to see what one of these might look like?

DeVorkin:

Sure. I would like a reference to it.

Harwit:

Well, we did a whole series of papers on these things, and I'm trying to see if there is a mask in here some place.

DeVorkin:

This paper is "Historical Background of Multiplexing by Means of Masks".

Harwit:

Yes, this was actually not the best article, perhaps, although, it's all right. This is a review article.

DeVorkin:

"Hadamard Transform Analytical Systems," by Martin Harwit.

Harwit:

I'll tell you what the reference to that is. It was in a book called, TRANSFORM TECHNIQUES IN CHEMISTRY, edited by P.R. Griffith, and put out by Plenum Press in 1978.

DeVorkin:

This is Paper No. 106 on your vita.

Harwit:

Yes. At any rate, I thought that this would be a nice way of perhaps eventually increasing the sensitivity of our spectrometers, because what you could do then with that kind of technique was to look at all of the light half of the time; that is, the slits were either opened or closed. I think last time I told you that Larry Mertz, had already started influencing me in that direction. He had worked with this kind of an instrument. We made two contributions to the subject. The first contribution came from Neil Sloan who had been in the Electrical Engineering Department at Cornell, and then eventually went to Bell Labs. He is an information theorist, and information theorists are very good at the mathematics of ones and minus-ones. He knew a lot of theorems and derived a lot of optimization theorems. A lot of what we did was really applied mathematics. He did most of the calculations, and put upper and lower bounds on the amount of improvement one would get. He was able to show that a Hadamard transform and a Hadamard transform mask gave the optimum results that you could ever achieve with any kind of a mask system, and in fact, we showed that in theory one can do better with those than one can do with Fourier transforms. It turns out that the Fourier series is not completely ortho normal. That is, you need both sine and cosine series which are the same frequencies. If you look at N-spectral elements, you really need 2-N points in order to analyze those instead of N-points. At any rate, my side of things was to try to build these instruments. Eventually, we also built imaging devices that made use of these masks. Instead of having slits now, you had two-dimensional arrays of open and closed slots. I think the main contribution that we made on the instrumental side was that we showed that rather than having to have N separate masks, which would have difficult and expensive to construct, one could use a single mask. Eventually we got up to the point where we had 1,023 elements in a mask, open or closed. Now, you couldn't have 1,023 different masks. But many Hadamard matrices have the property of having two to the n power of slots, (2, 4, 8, 16, . .1024). It turns out that if you take one of these off and make it 1023, or 511 instead of 512 – 2n-1 — it is then possible to make up a cyclic array where you just make a mask which is twice the size of the minimum length. Instead of having N elements in a string of slots, you have 2N elements, or rather 2N 1. If you take the sequence from 1 to N, you find that that is almost orthogonal to that from 2 to N+2. And then, in fact, when you take the scalar of product of these, the answer is always 1 rather than 0. But 1 is every small compared to N still. So you had almost orthogonal functions there; but they allowed you to make something which was cheap to construct. Instead of having to make up masks containing a million elements — if you attempted 1,023 elements and you needed 1,023 of those, you would have had to fabricate a million different elements or more — we could get by with 2,000.

DeVorkin:

How carefully did these things have to be fabricated?

Harwit:

They had to be done pretty well. In fact, we spent a lot of time looking at the errors that were introduced by fabrication. Again, it was a collaborative effort between Neil Sloan and myself. We published papers called "Errors in Hadamard Transform Techniques" and tried to trace all the various different problems.

DeVorkin:

What was the opinion of the technicians who had to make these slits? Did they ever ask what they were for?

Harwit:

We actually went to a professional outfit to get those made. There were a number of people who knew how to make those, because they were making them for television systems.

DeVorkin:

The same kind of masks?

Harwit:

No, not the same kind of masks, but they used the same fineness of slit. We were making up masks where the openings were a 10th of a mm. in size.

DeVorkin:

Yes, that's pretty small.

Harwit:

You can see in this one paper here, that these are the sorts of things that we were worried about, a lot of ones. Instead of having minus-ones there we had zeros or blank spaces. You would get little bits of noise in the spectra. This is in Paper No. 92, which was called: "Practical Multi Spectrum Hadamard Transform Spectrometer." This was done by some students at Cornell and myself. In this particular one, Neil Sloan wasn't involved.

DeVorkin:

Your interest particularly was in infrared spectroscopy at this point?

Harwit:

That's right, but also to some extent in imaging. It seemed to me that this could also be a useful technique there.

DeVorkin:

Can you identify any particular person or event that got you interested in this?

Harwit:

Well, I mentioned before that Larry Mertz had interested me.

DeVorkin:

So that was it.

Harwit:

That was it, yes. What had actually happened was that in between 1967 and 1969, I had been a consultant at an industrial consulting firm in Boston, a place called Comstock and Wescott. They were interested in various space ventures. I was supposed to suggest things that they might do. I think at one point I suggested this kind of a technique might be useful to them for some purpose, I don't remember what.

DeVorkin:

Highly efficient low light level spectroscopy.

Harwit:

That's possible.

DeVorkin:

Would it be limited to the infrared? It would be easier in the infrared because the tolerances would be less?

Harwit:

These instruments work any place where you don't have an imaging device; that is, you don't have something like a photographic plate, or a CCD array, or those kinds of things. And also, where the primary source of noise is noise that is intrinsic to the detector itself, rather than photon noise. When you have photon noise, the more light you have falling on the detector the noisier it gets. In that case, these schemes which allow you to see more light at any given time become self-defeating. But whenever it's detector noise you are limited by, or a number of other classes of noise, these schemes work well. I worked on that at Comstock and Wescott then with one of their physicist, John Decker. Later on in fact, John got so interested in this that he and I started a small company and tried to produce them.

DeVorkin:

Yes, what was the name of that company?

Harwit:

It was called Spectral Imaging, Incorporated, and it was in Concord, Massachusetts. John was president, and was full time, and I used to consult one day a week that Cornell allowed one to consult. I was vice president. We did that for about five-six years. We had some interesting projects. Together with American Science and Engineering Corporation, we got a contract from NASA to build an instrument that would both do spectroscopy and imaging simultaneously in the infrared. One could analyze a source or a scene simultaneously for 1,023 pixels, picture elements, and 63 different colors in the infrared. We actually built that machine for NASA and it worked. That was the culmination of the work.

DeVorkin:

What did they do with it?

Harwit:

Well, NASA thought that they might be able to use it for down-ward looking multispectral analyses.

DeVorkin:

Like a lens?

Harwit:

No, more like a set of separate colored pictures. I had actually invented that, or at least, to the extent that I had a patent for it. That hadn't been suggested before. It was a patentable idea. Do you want to buy a patent? (chuckles)

DeVorkin:

Buy a patent?

Harwit:

I still have the patent. Nobody's building any of these. (laugh). It's sort of an interesting thing, because it turned out that there was so much information you could gain, that I think it just wasn't matched to any of the kinds of problems that people had been thinking about.

DeVorkin:

What's the efficiency of it?

Harwit:

It's efficient. You have two masks. At each mask you lose one-half of the light, but other than that, you are looking at all of the colors all of the time. It's a pretty good system. It does all kinds of things. I learned a lot from that. Inventions which on paper sound best are not always the things that actually get used, because this is a highly general approach to things. It usually turns out that people don't reed the most general instrument available for what they want to do; and that they can build something far cheaper within the constraints of the limited objectives that they have. In addition to that, there were no applications at the time, it seemed, for something that gave you 63 different colors and 1,023 picture elements. I had thought initially that maybe in things like flame spectroscopy it might be something that might be useful.

DeVorkin:

Laboratory analysis of things.

Harwit:

Yes, one of the things that we had done with this instrument was to have it look at a flame, and then go through a carbon dioxide spectral line.

DeVorkin:

To modulate it?

Harwit:

To see at what wavelength you could see carbon dioxide emission in the flame. (Here we have it, page 98; I'm sure the picture of this is here). Here's the mask, the spatial mask. You can see how fine these elements were. It was bigger, but the whole disc was about this big.

DeVorkin:

This is page 1597 in APPLIED OPTICS, Volume 15, 1976, "Hadamard Transform Imager and Imaging Spectrometer," Swift, Watson, Decker, Paganetti and Martin Harwit.

DeVorkin:

How much did this thing cost?

Harwit:

It cost about $350,000.

DeVorkin:

That was in the production of the masks?

Harwit:

No, in the overall engineering. The masks themselves, I think, were only about $15,000, something like that. But you see here now, we go through a number of different successive wavelengths. Out of those 63 we have chosen 16 here as shown on page 1605. You see a small dot here in the upper lefthand corner which was a soldering iron, and you see that that is quite bright here, and you see no flame at all at this wavelength. Then the whole thing is normalized — well, I don't know if it is normalized or not — but you see that the black body emission from the iron is fairly constant and rather faint. I think these must be normalized, because it gets fainter wherever the flame is bright. As you step through the carbon dioxide spectral line, the flame gets brighter and brighter, and then diminshes again. Presumably one could do similar things with CH vibration emissions. The idea was that it would be nice to see, for example, at what point certain molecules went through combustion in a flame, because these things don't burn uniformly throughout a flame. So we thought this was one potential application. 'But it never really materialized, and after a while both John Decker and I decided that this was not a way to make a lot of money and closed down the company. Actually, it was much to my relief, because we were doing some Hadamard Transform work at Cornell also at the time. I was continually leaning over backwards to make sure that somehow the students wouldn't feel that they were doing things only because I was interested in using them then in the company. So what we tried to do was that if there was any transfer of equipment, or even ideas to some extent, that it tended to go from the company to Cornell, rather than the other way around, which also then put a constraint on the company which was uncomfortable. But it was just a lesson that it is difficult to do something commercially that you are also doing through a nonprofit organization at the same time.

DeVorkin:

Did you get a sense of this from the students, or was this purely inside?

Harwit:

No, but I had been at MIT as a grad student, and there were a lot of people who had their own companies. There was always this scuttlebutt amongst the students at MIT, oh yes, old so-an-so is just having me do this so that he can make his million on it. I just wanted to steer away from that. The whole things was cleared up with the university well ahead of time; and so legally it was alright, but these things tend to go beyond the legal, in a way. Your relationship with your students is at stake.

DeVorkin:

Yes. I agree. It's an interesting thing. It's obviously a very exotic device. How many were actually built?

Harwit:

Oh, I don't know. At Cornell we built three, four, five different types of things. In the company perhaps a similar number, not very many. But I was really pleased earlier this year; I read in the paper that the Japanese had put up an x-ray satellite where they were making use of a Hadamard Transform Mask. So here is a completely different area of application. We had suggested in our book that for x-ray techniques — where again you are using single detectors; you don't have any imaging device in many cases — that this might be a useful sort of a scheme. I don't know to what extent that influenced them. I do know that I talked to some French people one time who said they were working on some gamma ray detectors, and when I mentioned the book, they said, oh yes, well, they used that all the time. Well, we had worked on a lot of general theorems that could be used by anybody who wanted to make use of the technique. So once we had more or less worked up this technique, at least in principle, and had built all the various possible things, an imager, a spectrometer, a spectral imager, we really kind of exhausted that particular technique, and since nobody was willing to jump in and adopt that technique for medical purposes or whatever, we decided we would write it up as a book. Then after that I got out of it.

DeVorkin:

Yes. So it's interesting that you got into that because you were interested in solving astronomical instrumentation problems, but then it was a general instrument that you ended up looking for commercial uses for, or for wider research R and D uses.

Harwit:

Yes, yes.

DeVorkin:

It sort of bridges a question I'd like to ask you a little later about what drives what in your experience and your work, technology driving science, or science driving technology. We can get back to that.

Harwit:

Okay.

DeVorkin:

Let me go back at this point. I assume you have pretty well covered your work with Hadamard transforms?

Harwit:

Yes.

DeVorkin:

Let me ask you to review the history of space science at Cornell, the Center for Radio Physics.

Harwit:

Let me just add one thing still first about the spectroscopy that we then went on to do from the airplane. Because I just mentioned this first fine structure line that we found at 88 microns of oxygen. After that we went on and we detected a second fine structure line from the same doubly ionized oxygen atom. We then were able to get all kinds of physical insight on number densities in the regions where the emissions were taking place. And then later on we did neutral oxygen. We discovered the line at 63 microns. And I think it was only after that that somebody else made a discovery of a new line, a double ionized nitrogen line at 57 microns, a European group with five-six co-authors.

DeVorkin:

They made this as an observational discovery.

Harwit:

As an observational discovery, Balluteau (I don't know whether it is a double or single "1") and Moorewood were the main people in that group. Later on also, the Berkeley group of Charles Townes got involved in the same wavelength region spectroscopy, and they detected the first far infrared molecular transitions, transitions of CO and CH.

DeVorkin:

Tell me, were they, again, predicting where these were?

Harwit:

Well, for the molecular lines, one knew exactly where the transitions would occur.

DeVorkin:

Where they should be, yes.

Harwit:

Not for all the molecular lines; for some of these species — for example, we recently looked at CH — the wavelengths were not all that well known. Just before looking for them we contacted Ken Evenson, at the National Bureau of Standards in Boulder, Colorado. He gave me the exact wavelengths that we would have to look for. We looked for them a day or two later and got signals then. (background noise) We have to work the data up but we think we just detected those lines.

DeVorkin:

Was that very recently?

Harwit:

That's very recently, yes. That was just in March, a few months ago. So we have now also discovered a number of molecular lines. The whole point is that one is now getting to the point where we can start analyzing interstellar gas phase chemistry. If you look at a region where shocks are occurring, you want to know what are the chemical ingredients that are there before the shock, and what are the ones that are there after the shock, and during the shock. That is very important because that may in turn affect how that region is able to radiate away its energy, cool itself down afterwards, and potentially go into stellar collapse, on its way to becoming a protostar — this kind of thing.

DeVorkin:

The whole problem of radiating away the energy?

Harwit:

That's right, yes. And also, one wants to know how the molecules that you see in the molecular clouds are formed. The shocks may be where some of these molecules are formed, and so forth. So one really would like to understand this chemistry. The chemistry involves to a large extent hydrogen-containing compounds. The basic molecules will start out as diatomic hydrides. So you will have CH, CH+, OH, and maybe OH+ coming in, as intermediate steps towards forming the larger molecules. So if you want to understand the formation of the larger molecules that the radio astronomers see, this gas phase chemistry is one means by which they might be constructed. Unless they are formed, and perhaps even if they are formed, in the atmospheres of stars. They are either formed there or perhaps in the shocks. But ether way, one might hope to be able to discover the sequence of steps by means of which these larger molecules are formed.

DeVorkin:

Now, you're talking about the formation of the smaller molecules which may be the constituents of the large molecules which may be formed in the atmospheres of stars.

Harwit:

Yes, and also in the shocks. But there has always been this puzzle, how do you form large molecules? One doesn't understand that. And maybe by understanding how you form the smaller ones, and the reactions that they undergo you might understand how they became bigger ones; if you could see each of the successive steps as the molecule grows, you would have a better understanding of the physics that is going on. So this is really the thrust of the work that we are carrying out, and I believe, also the intent of the Townes group, it is a fairly hot subject right now. Over the years we've increased our sensitivity, I would say, maybe a factor of two every two years, maybe faster than that, and the spectral resolution by a factor of two every two years also.

DeVorkin:

With better detectors?

Harwit:

Mainly better detectors and also better spectral equipment. For example, in our case, we added onto our liquid helium-cooled grating instrument an interferometer of a novel kind that we invented, and that allowed us to jump from what had been a spectral-resolving power of about 1,000. And next year we expect to be at about 3,000. So we have already increased the spectral resolution by a factor of 100, and are going on to 300 now. The Townes group actually is somewhat ahead of us in spectral-resolving power at the moment, a factor of two or three maybe.

DeVorkin:

You mentioned a few weeks ago that given the limited space on the Lear jet and limited funding that you were in competition with the Townes group.

Harwit:

On the Lear jet we initially were in competition with them, but they dropped out. I think they had a bad graduate student and that project just petered out. But on the Kuiper Observatory, we are in very hot competition with them. In fact, at the moment, I think their instrument would have to be said to be somewhat better. But our instrument, I feel, has the stronger potential, because it's a multiplex instrument. Again, it will give one the capability of looking at 32 or 64 spectral elements all simultaneously. Groups tend to leapfrog each other. When you start a new thing like adding an interferometer onto your existing instrument, then initially you are going to have some losses; there are reflections, you lose half the light in the modulation process and so forth. You have to step by step bring that up on line; you go a step backwards, and then hope to take two or three steps forward.

DeVorkin:

Right. Yes, that's very common.

Harwit:

So what happens then, and I think what I complained to you about a week ago about, is that because sensitivity of the Townes group is high now, we were chided by NASA and threatened with losing our funding next year unless things get better. This is the uncomfortable part of a highly pragmatic approach, which says, what have you done for us this year? In any one year you just can't really bring all of this sort of thing up on line and have it working at its ultimate level of performance. Some of these things take two or three years to really get going.

DeVorkin:

Has this yearly oversight always been the case at NASA?

Harwit:

There were times when they had what was called step funding, where you got three-year contracts and you were guaranteed the first year you would have the full amount of money. The second year you would at least have two-thirds, and the third year you would have at least one-third of the original amount. In 1970 or so when NASA started to have the post-Apollo crunch, they just threw that out the window. They didn't have the money and they just sort of ate up their babies, so to speak, by saying, well, in the second year we have to cut you out. Some people got hurt by that. Sometimes NASA honored it, but it did have the original intent of going on for many, many years, with a built-in kitty there that guaranteed that they could always pay the last two years at the two-thirds and one-third level respectively. They just didn't have that reserve so that they could guarantee that any more.

DeVorkin:

Has new instrumentation suffered as a result of this?

Harwit:

I think so, yes, to some extent. I am somewhat ambivalent about this. I go to Germany, where I think I mentioned to you I've been a number of years in a row, for extended periods.

DeVorkin:

You are a consultant at the Max Planck Institute?

Harwit:

I'm not a consultant. I'm what's called an external member of the Institute, and with no fixed duties, but with complementing interests. There are only a few such people whom they elect to that. You have to be elected by the whole Max Planck Society for that.

DeVorkin:

How did you become a member?

Harwit:

I had been on sabbatic leave, because I had been interested in understanding radio measurements better. I found that in the far infrared we were looking at regions for which the best data came from radio observations. I really wanted to work for a year some place where radio techniques were being employed; in particular, where the shorter wavelength radio work was being done, which dealt with molecular constituents with ionized regions and so forth. The Max Planck Institute for Radio Astronomy in Bonn was just a very good place to go for that. I guess Peter Mezger was one of the three directors there, and now is the so-called business-conducting director. That places him a notch above the other director or directors. It's potentially a rotating job, but he's had it now for about six years. He and I hit it off together very well. We had a lot of common interests. We were trying to go out to longer wavelengths in the infrared towards the submillimeter. He wanted to come from the centimeter and millimeter region into the submillimeter; and so there was sort of an agreement that we would somehow try to meet in the middle and in a sense that I would help his people come down in wavelength, and perhaps they could show me how to get up in wavelength. Their people have flown with us on the Lear jet, using our instrumentation. He wanted people to get familiar with the techniques and also do the actual observations, which are of interest to augment their radio work. So as a result of that I've gone there, and I act as a consultant for them. They have been trying to build up a European airplane for observations.

DeVorkin:

Are you discussing this in contrast to the NASA yearly funding policy for instrumentation?

Harwit:

Yes. Now, over there, they tend to have three to five-year grants. Peter Mezger is always saying people take things much too leisurely; and perhaps there is that element. You do see in the European laboratories much more polishing (technical refinement) of instrumentation, and you keep thinking, gee, these guys have such great instruments, why don't they use them on something astronomical? On an instrument, you can always make things better. You can automate it better. You can think up neat programs that you put into your computers to display all kinds of stuff that probably you'll never use, and lovely visual displays, multicolor and so on. You can do that for years without running out of new things that you could be doing. One tends to see more of that in Europe; a heck of a lot more than in this country, where a lot of stuff is dirty-looking but it works. You have to know what you can have taped up, and what has to be absolutely correct — you can't fool with them at all; they really have to be done right. Other things are not that important. Townes group, for example, which has been very successful, has always been fairly lean on their computational displays. They haven't gone into much effort there; whereas, I've had a couple of undergraduates in my group who have been real hotshots on doing things with computers. They have actually brought our stuff up to a point where it is most sophisticated. I don't think it has helped us that much, or didn't help us that much until we got into the Fourier analysis, the multiplexing techniques. Then it was indispensible. You can't do it by hand. You have to do that in flight in real time. You have to know what you are getting during flight, and so you have to perform a Fourier transform right in flight. So to that extent the computer was necessary there; but other than that you can overdo the computer automation. But anyway, in this country where we have this year-to-year funding, we always have to come up with a minor miracle each year in order to impress the people who judge these proposals. If in one year you just come up with nice astronomical results, they yell at you for not having advanced the instrumentation, or falling behind other people in instrumentation.

DeVorkin:

Aren't these astronomers who are doing the evaluation?

Harwit:

Yes.

DeVorkin:

Don't they appreciate the difficulties and the impracticality of a miracle a year?

Harwit:

I just don't know. I've never sat in on that meeting, except for a number of years, when I was chairman of the Users of the Airborne Observatories, I would make a presentation to these people. So I could tell by the questions that they had afterwards what was on their minds.

DeVorkin:

Do these people include Nancy Boggess?

Harwit:

Yes, Nancy Boggess is of course the person whom they advise. They are good scientists. But what happens to a good scientist when he sits on a committee is that he wants to bring his analytical powers to bear on what is a complex situation. So he will pick up key signs, like noise equivalent power, the sensitivity of detectors. If he hones in only on that, he will say, well, this person claims such and such a sensitivity, while that person only claims one-third that sensitivity. We really can't afford to have two instruments on the plane. Let's throw this one off. Yet it may very well be that the second group has really done much nicer astronomy. But that's something that will only appear 10 or 15 years later.

DeVorkin:

Do they only, then, look at the instruments and the designs?

Harwit:

Well, it's the simplest thing. It's the most simple minded thing that you can put a number on, and these people want to be analytical in some way, so they don't want to be confused by the facts, I suppose (laughs).

DeVorkin:

There are all of these recent attempts by NASA and other funding agencies in the Government to fund on the basis of past astronomical, or past scientific performance, past scientific products.

Harwit:

I have never encountered that.

DeVorkin:

It really doesn't happen then?

Harwit:

I don't think so. It may happen some places; there are so many peer review committees that it occasionally may happen.

DeVorkin:

But in your area, the miracle you're talking about is a technological miracle.

Harwit:

Well, I try to do both. Usually what I've tried to do each year is to come up with some sort of a really striking observational result and also a striking technological result. Last year, for example, we did a first lunar occultation observation, where we were looking for G spectral line across a very diffused region. We chopped against the moon where we knew there wouldn't be any spectral line, and used that as a reference point, because there was no other part of the sky where we could be sure that there was no spectral line emission. But the moon wasn't going to emit that particular line; we knew that and so we chopped against it. We found the line and find what the flux was that we were seeing from the galactic plane; and I thought that was quite a neat experiment and that it would impress those people. This was done on the Lear jet. As you know, they have now decided to cut out the Lear jet next year altogether.

DeVorkin:

Is this definite?

Harwit:

Yes, it's pretty definite, but it's had nothing to do with us. I think it is mainly that they don't have enough money for both the Kuiper Observatory and the Lear jet for next year. They have major repairs coming up on the Kuiper Observatory. There are so many people who fly on that plane and relatively few people fly on the Lear jet, so they just decided they would put all the money into the Kuiper for this next year. But I think it may well be the death of the Lear jet then for astronomy at least.

DeVorkin:

That's too bad.

Harwit:

Again, it's considerations like these that led me to write COSMIC DISCOVERY, because every time one of these things happened I felt there must be a more rational way of dealing with these things. I don't think you are ever going to get people to be completely rational. In fact, it's not a rational discipline, deciding what to do next in science. But at any rate, I wanted to see whether there were any kinds of major guidelines that you could perhaps indicate that would be useful to committees that were sitting in judgment of other scientists.

DeVorkin:

This is just another way to collect data and interpret it.

Harwit:

Yes. There is this aspect of the work where we actually do things, the aspect where I get upset because of the way that I see committees looking at things, and which then leads me to the historical studies, trying to understand what goes on. Then there are the information theoretical aspect which came out of the spectroscopy (they are still involved in what we do) which slant my particular approach to the policy studies; as for example in the third and fourth chapters of my book, which are really information theoretical types of considerations.

DeVorkin:

Yes. Okay, let me change the tape.

DeVorkin:

Tape No. Four, Side No. One. I had just mentioned before I was changing the tape that I was sorry to hear the demise of the Lear jet, and you had an unexpected reaction.

Harwit:

(laughs) Well, I said I wasn't that sorry. The reason I'm not that sorry is because that Lear jet has been allowed to fall apart for the last five, six, seven years, not in a way that it would be unsafe to fly in, but the telescope, the gyros on that are slowly wearing out. There has been one part-time technician to the job. Personnel has changed. No documentation has been kept. New people come in and out. We have to show them how to operate their own equipment at NASA Ames. We've got to repair it for them. This last time around I was glad to see at least they had installed the telescope on the airplane before we came out. But when I looked at it I found that the primary mirror was installed wrong on the telescope. Instead of having three mounting pads there were only two. The guy who had put it in there had just not given a hoot, I guess, or hadn't noticed. We took the mirror out to install it properly. We found that an O-ing that is supposed to seal the telescope had just not been put in. People had been wondering why the airplane couldn't be pressurized. Then we got that back in. I had to make up an O-ring; cut up a bigger one to make one the right size, and then put it all back together. I looked through the telescope and I saw that the secondary had the aluminum coating flaking off. I think that was because the plane had been down in Panama, and some of the salt air had probably gotten at the mirror. So then I tried to find another secondary that was around. They were all mislabeled. One had to sort of go through the optics of trying to figure out what was what. It took me a day to assemble that whole telescope, one whole Saturday, of course, swearing all the time, as you can imagine, in my best Army-taught language. It's that kind of thing where it is a continual hassle.

DeVorkin:

Yes. It also sounds like a small time operation.

Harwit:

It's a very small-time operation.

DeVorkin:

Yes. Does this explain why I was not able to get a 12-inch mirror from them?

Harwit:

It could be. I don't know; I mean nobody knows where anything is. For example, there had been a log book that had been there in the previous session in January. We hunted all over the place trying to find out where that was, so we could see what the technician had done that time, so we could reinstall things in the same way. Then on the guidescope we had had a reticle that I had glued in, myself, and made a holder for and everything on a couple of previous flights. People had been using the guide telescope imaging tube for some test for the Kuiper Observatory and had just torn that out. I had to make up a new one and put it in. It's that kind of stuff, where between flight series things get left lying on the floor of the hangar. They don't get put back right. One loses them. Everything gets thrown in a closet, and then you can't find anything because everything is lying on top of everything else.

DeVorkin:

Is Al Harper still flying the Lear jet?

Harwit:

No. He gave up, maybe for those reasons, years ago. We were the last group the last three or four years to fly the Lear jet, and then a former student, Gary Melnick, and a former post-doc, Ray Russell, decided also to go back into that. They started flying in the plane with me, and they could see that they could handle it, but that was only just this very last year, now.

DeVorkin:

How would you compare service, even on the Kuiper, to what you normally get at Kitt Peak as a visiting astronomer?

Harwit:

The Kuiper is absolutely great. They are very good people, very helpful. It's the contrast with the Lear jet which is the difference between the Rolls Royce and a model-T that's had 250,000 miles on it. There is no mechanic any more who knows how to fix it.

DeVorkin:

Was the Lear jet conceived of initially as a permanent part of the program, or just as a step?

Harwit:

I don't think so. Most of these things have always been conceived as steps. For years now, there has been talk of going to a short version of the 747, but that's so expensive that when we made a presentation at NASA Headquarters about two and a half years ago, Frank Martin said, this is the last time I want to hear about this plane. It turned out that over a period of the life of the plane it would have cost, initial investment plus annual contribution, something like $300-million, and he felt he could run a pretty respectable space mission for that sort of money, and that was what NASA was really all about. It made sense.

DeVorkin:

So NASA doesn't see the Kuiper program as a central interest.

Harwit:

No. I think the sub-orbital programs which comprise the rockets, the airplanes and balloons are there really only in a supporting role. It's a place where you invent and test out new instruments and demonstrate the performance of your instruments that you would want ultimately to fly in space.

DeVorkin:

Is this then what people at Goddard see as their role, like Werner Neupert, the good tinkerers who try different things. They send things up on Black Brandts and sounding rockets.

Harwit:

Well, Neupert also has had satellites.

DeVorkin:

Sure, he has, but not in the present.

Harwit:

Yes, I think that's true, whenever you have these things, there's a double role. One is to tinker. This is why, I think, the MOWG, which I mentioned before, for airplanes has this great interest in sensitivity and instrumental performance, because they see this as the testing ground for instruments that ultimately go into space. So I think they have this in mind. There are two sides. There is also the additional side of getting enough astronomical background, so that the spacecraft instrumentation that you design actually do make astronomical sense later on. If you never did anything on the Kuiper and you wanted to start out doing far infrared spectroscopy from the Shuttle, for example, you would probably do a lot of things wrong. You wouldn't know where to look. You wouldn't know what spectral resolution and sensitivity you needed and so forth.

DeVorkin:

Let's talk about, if you don't mind, NASA ground based, while we're talking about NASA in a large capacity. Were you ever involved in consulting with NASA, or for the IRTF?

Harwit:

No, although I sat in on some discussions. There was a presentation of that at some committee that Nancy Boggess had asked me to just sit in on for reasons that I don't remember exactly. I've also had Frank Martin, when he was the head of that office over at NASA Headquarters, saying that he really felt the IRTF was not what NASA missions were all about, that it was purely a planetary support scheme initially. Bill Brunk, I think, had justified that on the basis of getting what's called, "ground truth," a horrible word, but meaning essentially verification.

DeVorkin:

That sounds awful.

Harwit:

You've never heard it before?

DeVorkin:

But "ground truth"? Who coined that term?

Harwit:

Oh, I don't know. That's been there for a long time. You always have that kind of verification in almost anything, even in the military. If you, say, map from a satellite, you then have a low-flying aircraft to take the pictures of the same area to compare and see. One of them is what you consider to be the truth, and so if you make an observation with a telescope from the ground of something where you have overlapping information from a spacecraft, you can verify that the spacecraft is doing things correctly.

DeVorkin:

So it's still that people trust ground-based observations?

Harwit:

Yes, I think so, as a cross check. Then also, they wanted to have a lot of these things to follow up on observations that you could make from the satellites; and again, probably develop instrumentation also that ultimately would go into satellites and give more information on what you should be doing from spacecraft, like the Voyager, all kinds of things like that.

DeVorkin:

But Bill Brunk was perceived as supporting ground-based planetary work in pushing for the IRTF? Do I have the right?

Harwit:

I think so, yes. I once had a grant from him for this Hadamard transform work and he was quite adamant that we had better do some planetary things with that. He wasn't supporting galactic astronomy. I can see that; I mean he works for a single part of that organization. The mission aspect of NASA comes to do this or that. What NASA is supposed to be doing isn't — they don't mind if you go off a little bit and do something else, as long as you deliver on what they've got to show their bosses, and also fulfill the role that they have been assigned.

DeVorkin:

Yes. So this was between 9-73 and 3-77 when you did 10-micron spectra, and that was supported by Brunk at NASA?

Harwit:

Part of the time; also, I think I had some Air Force support for that initially from ADM, at that time AFCRL.

DeVorkin:

Right. Now that was definitely planetary work.

Harwit:

Well, Brunk funding was planetary. The Air Force funding was not. That was primarily optical.

DeVorkin:

What was your perception of the attitudes of others during the earlier period with respect to planetary work? This is something that Joe Tatarewicz of course is working on. Is it appropriate now to talk about that, the perception of the support for planetary work by NASA, and the perception of its efficacy by the astronomical community?

Harwit:

I don't understand the question. I'm sorry.

DeVorkin:

How did classical astronomers regard planetary work as NASA began supporting it, both in space and on the ground?

Harwit:

I really don't know. That dye was cast by the time that I got into doing spectroscopy, and before that I had been doing rocket work throughout the 1960s. The only spectroscopy I had done was for this molecular hydrogen work much earlier, and didn't involve any planetary work, and was NSF supported.

DeVorkin:

So you were pretty much independent of planetary work?

Harwit:

That's right, yes. The only time I did planetary work was when I had instruments that were insufficiently sensitive to do galactic work. But that's not quite right either. To the extent that I got involved in planetary work eventually some of the things did become interesting to me for their own sake. For example, just two months ago we decided we would look at some phosphene bands in the atmosphere of Saturn that had been predicted there in the far infrared. We thought it would make an interesting observation that was very easy for us to do actually. If it had been difficult to do, I might not have decided to do it. I weld have done something else galactic that was equally difficult, but because it was easy, and because the papers I read suggested that it would give interesting results, we did it. Peter Gierasch at Cornell and our group had just decided to join forces on that.

DeVorkin:

Who's he?

Harwit:

He is a planetary atmospheric dynamics person, but he's also been interested in planetary atmospheric spectra to understand the dynamics properly. He decided that there were a couple of interesting things that our observations could tell which differed from those of the original papers that I had read. The original papers just weren't right any more. I didn't know enough planetary astronomy to judge that as being wrong.

DeVorkin:

Yes. If it is all right, could we move to your Cornell association? That will involve a little bit of planetary, too, quite likely.

Harwit:

All right.

DeVorkin:

You've already told me a number of interesting things about how Tommy Gold wrested control of Arecibo from the engineers; but this must have happened after the DOD transfer.

Harwit:

No. That actually happened at the time that DOD was still there. ARPA was running it for a while. Tommy Gold then was director of Arecibo. Around 1968, I think, the Observatory got transferred to the National Science Foundation, and the National Science Foundation – I don't know what the details were — didn't want Tommy as director of that.

DeVorkin:

Oh, they didn't? That's interesting.

Harwit:

Well, there were a number of reasons, I think. In the meantime Cornell had gotten in some new ionospheric people to replace the group that had gone to the University of California at La Jolla.

DeVorkin:

The engineering group that was kicked out.

Harwit:

Yes, and Don Farley had been brought in, an excellent ionospheric person, and also Neil Brice, who died in a tragic Pan Am crash in the South Pacific, a very bright guy.

DeVorkin:

What I think Joe is interested in, and some others, is Frank Drake's developing role in Arecibo. Could you clarify it?

Harwit:

I'll bring that in in just a second. These people were complaining, and the community also was complaining that they were not getting enough ionospheric time at Arecibo; and Tommy Gold felt that astronomy was really the strongest role there. He, and later on also Frank, when he became director, suppressed the ionospheric guys, and didn't want to give them time at Arecibo on the telescope. Then Farley and Brice wanted to go out and get their own grants, and, then Tommy and Drake worked on the vice president for research at Cornell to not permit that policy either. I think Don Cooke already was vice president then, but it might have been Frank Long. I don't know whether it was Don Cooke or Frank Long. But at any rate, Brice and Farley couldn't make any headway with the vice president either; and they took their case to the ionospheric community, which in turn went to the National Science Foundation. I think somewhere along the line all hell broke loose. There just had been too many enemies made. Bill Gordon, who had lost control, and Booker, all the strongest people in ionospheric work were very upset at the way this was being done. This is all third-hand by the way.

DeVorkin:

I appreciate it.

Harwit:

There is a history of Arecibo being written by Joanna Rankin.

DeVorkin:

That's right. Do you have any idea how that stands?

Harwit:

No, I haven't seen her in years. I hope it will come out. It will be a best seller, I think, if it comes at all close to being open.

DeVorkin:

Well, (for the tape, because I want to remember this) She initially wanted to do a history of the study of black holes.

Harwit:

Is that right?

DeVorkin:

A radio study. And only became interested in the institutional history when she realized she knew that a lot better. And after talking to a number of historians. That was much more interesting.

Harwit:

Yes, I see, good. At any rate, Frank was made director.

DeVorkin:

Even though Frank Drake was working in combination with Tommy Gold, was he seen as less of an antagonist to the ionospheric guys?

Harwit:

Well, I don't know, maybe so. I think I've jumped ahead too fast. I think the ionospheric people were antagonistic to Tommy, but that may not have been the major thing at that time. Whatever it was, his control was taken away and Frank Drake who had been site director out at Arecibo, the resident director there, became director of Arecibo then in 1970, I think it was. Now, he continued his dispute with the ionospheric people, and eventually the ionospheric work was forced on Arecibo. I never was a participant in this whole business, so I don't really know just what the disputes were, but I know that at some point the ionospheric people did get quite upset, and that was actually even after Frank Drake became director. Eventually that somehow got resolved. The people at Cornell in ionospheric work were allowed to apply, and eventually did receive NSF support for their work, but it was always an uneasy relationship, as long as Frank Drake was director there.

DeVorkin:

How was he perceived by you or by other people at Cornell as an administrator?

Harwit:

Well, Frank likes to play things very close to the vest, and so we in the department really never knew anything about what was going on at the Arecibo advisory board meetings, except through what Frank would report to us at faculty meetings.

DeVorkin:

Was Frank Drake accountable then to the faculty?

Harwit:

Not for his work at Arecibo, no. He was accountable only to the vice president for research, in whom officially the National Science Foundation had vested its authority. In fact, I think that was somewhat of a trick that initially NSF didn't want to deal with Tommy, I understand, and so they decided that it ought to be handled at the vice president's level. And then the vice president made Tommy associate vice president for Arecibo. I was away in Czechoslovakia that year, so all of that happened while I was away. I can't give you any reliable information of any of that, but the whole thing sort of blew up.

DeVorkin:

You were starting to say something, before I interrupted you.

Harwit:

Yes. Around 1980 we on the faculty felt that we really wanted to have more direct access to the advisory board, because we felt very ill at ease at getting just what we heard of as criticisms of the faculty from them through what Frank Drake was saying to us. So we petitioned the university to allow us to have some member present at relevant parts of the advisory board meetings. The university didn't want to do that, and Frank objected very strongly also, and the whole affair became quite unpleasant. Eventually then, after about a year or year and a half of this, Frank was coming to a completion of his tenure as director, and the university now has Tor Hagfors as director of Arecibo.

DeVorkin:

In the Center for Radio Physics and Space Research, I'm curious as to the origin of the name, and then to the relative roles in importance off radio and space at Cornell?

Harwit:

The center was all set up before I got there. I think it was set up with that name because the electrical engineers who were doing ionospheric work considered themselves radio physicists. They were people like Booker who were really doing physics of the ionosphere. So that is where the word, radio physics, I imagine, came about. Space research was the good word to use instead of astronomy at the time that NASA was founded, because then you could get money from NASA.

DeVorkin:

In this case, was the study of the ionosphere or the study of the upper atmosphere thought to be space research? Was this the implication?

Harwit:

No, that wasn't the implication. The implication was that somehow we were going to go into space research. In fact, NASA provided money for us to build the building that we are in now. I think that was solely the purpose.

DeVorkin:

Is it true that you were the fist one, though, to go into space research literally?

Harwit:

I think that's correct, certainly the first one to build hardware. I don't know whether Tommy at any time might have been co-investigator on, say, one of the early lunar shots or something like that. But in terms of actually building equipment and starting up a laboratory and so forth, yes.

DeVorkin:

You were the one.

Harwit:

I think that's right. I don't think anybody over on the electrical engineering side was doing anything like that.

DeVorkin:

How many space groups within the space research part of the center or on the faculty existed, or exist now? Or have existed in the past?

Harwit:

We really don't divide things up in that way. There are four research groups that we have, and sometimes people work outside of their particular area. For example, when I do some of these historical-sociological type studies, I'm not acting as an infrared astronomer. But I'm considered part of the infrared group. There is a planetary group, a radio astronomy group and a theory group. Now radio astronomy is not in the Center for Radio Physics and Space Research. Ever since the National Science Foundation formed NAIC, the National Astronomy and Ionosphere Center, the radio astronomy group is directly responsible to the vice president of research at Cornell. The director for the Center for Radio Physics and Research also is directly responsible to the vice president for research. So they are really parallel organizations.

DeVorkin:

Yes. Are you all in the same building.

Harwit:

We are all in the same building, although the NAIC people of course have this very big establishment in Arecibo. Then they also have an overflow of engineers and technicians in a building near the airport in Ithaca, because we just couldn't accomodate the whole group. The building that we have on campus isn't big enough. We felt that that ought to be the place where students and faculty would be, because it seemed best to try to keep all of those people in one building. Actually, even that isn't quite right. There is a neighboring building in which we have two or thre rooms, because we just couldn't make it.

DeVorkin:

When did the planetary group start?

Harwit:

There had always been some planetary work, in the sense that Tommy Gold had been interested in the moon, in the planets, and in the solar system as a whole. Frank Drake had also been interested in planets, but in a proper way it started with Carl Sagan's being brought on. I think that was around 1967, perhaps. He had been denied tenure at Harvard for purely political reasons, I am sure. It wasn't a matter of ability.

DeVorkin:

Do you know the political issue at all?

Harwit:

I think there were one or two people who just didn't want to give him tenure.

DeVorkin:

Astronomers.

Harwit:

Yes, astronomers. I think, and again, this is scuttlebutt, that somebody on the faculty, I forget who, didn't want Carl. Whipple was for him. He mentioned that recently; but there was some sort of a story that I think Whipple told that Urey had written a letter in which he didn't think Carl was very good. Years later when he had seen some of the things Carl had done on television, he wrote Carl a letter in which he apologized, although obliquely. Again, I don't know how useful that sort of information is; these are just things that one hears. Dave Layzer, I think, was not very pro-Carl Sagan. I just don't know what the juxtaposition was; but it may very well be that, if there were one or two people on the faculty who were lukewarm — and Urey certainly had a very strong influence on planetary research — and if he had a negative letter, then it is quite possible that would have made a difference. I just don't know, there were so many stories around that I don't want you to take any of what I'm saying here very seriously.

DeVorkin:

What was your perception of him when he first came?

Harwit:

He was quite well known already. It was to some extent a surprise that he hadn't been able to receive tenure, and he came in as an associate professor, and at the time I also was an associate professor. I got promoted the next year and he got promoted the year after that. It was all fairly straightforward; within two years he was full professor. He then started to try to attract younger people. He had brought some post-docs along from Harvard, and former students as well. Joe Virverka came with him. Joe had been a student of Fred Whipple's. But he had started working with Carl at Harvard before coming to Cornell. Then Dave Morrison came. He had been a student of Carl's. He was a post-doc at Cornell for a couple of years, I think. His wife, Nancy Morrison, started in as a graduate student, and then they both went on to Hawaii. I think she finished up at Hawaii.

DeVorkin:

She's at Toledo now.

Harwit:

That's right. I think they are divorced, yes.

DeVorkin:

Was Carl Sagan perceived as a real intellectual scholarly shot in the arm for Cornell to build up the planetary program? Why did Cornell bring him? Were you privy at all to that?

Harwit:

No, I wasn't. Tommy Gold really ran things without ever consulting anybody. In the whole time he was chairman I don't think we ever had a faculty meeting. Although he also was director of the Center for Radio Physics and Space Research, and there were a few meetings of that. There was also another organization. For a while Cornell had what was called the Cornell Sydney University Astronomy Center. Tommy and Harry Messel at Sydney University had decided that what one needed was an intercontinental relationship for things which could be done by radio work in the Northern and Southern hemispheres to augment each other. They had good people there, like B.O. Mills and Hanbury -Brown, and a whole number of other people. So that was inaugurated with some fanfare around 1963 or 1964. A few of the senior people went back and forth from time to time. But nothing ever really solid came out of it. I don't think that it ever lived up to its intentions. They had a few meetings where the Australians came up to Cornell. Then for awhile there were some annual meetings of the CSUAC; but eventually those petered out. There was just very little communication. Tommy just wanted to run things by himself.

DeVorkin:

So you don't really know how or why Carl Sagan was brought?

Harwit:

Frank Drake was a good friend of his, and they both enjoyed the extraterrestrial intelligence thing. I think that must have played a big role. Between them they represented the strongest group in the world on that subject. We also needed somebody in the planetary area, so it made a lot of good sense to pick him up. As far as raising the intellectual level goes, I can't imagine that Tommy would have thought he needed some one to do that when we had him and Ed Salpeter and peripherally also Hans Bethe. That wasn't the problem, I don't think.

DeVorkin:

Did public popularity, or notoriety play a role at the time? Carl Sagan was well known, but wasn't the public figure he is now.

Harwit:

Notoriety might have played a role. Tommy Gold always had an admiration for success of that type. So it might have, yes.

DeVorkin:

Did Sagan's coming at the time change the atmosphere of the place?

Harwit:

No, I don't think so. It enlarged the repertory that we had, but relations had always been fairly cordial between most of us in the astronomy department. In fact, I would say better than. cordial. And that didn't change.

DeVorkin:

Has it changed any time since?

Harwit:

There have been times when there have been disputes between individuals that have lasted for a year or two and then have been forgotten. Considering the individuality of the people in the department, it's surprising that we've gotten along as well as we have. It's never been the kind of thing that spills out and is known all over campuses from coast to coast; you know, so-and-so never talks to so-and-so, or the students of so-and-so are on the black list of his rivals, this kind of thing.

DeVorkin:

Yes. Has anything ever developed, though, that would have affected the course of your research, vis-a-vis space work, say, trying to get a major berth on a scientific satellite and not getting institutional support because somebody else objected, or something like that?

Harwit:

No, that's never happened. When I first got into rocket work, Tommy Gold took me aside in his office, and I was standing there like a little school boy, and he was lecturing me on how one had to take money seriously and stuff like that. It just seemed rather ridiculous, but what are you going to do.

DeVorkin:

What did he mean you have to take money seriously?

Harwit:

He was telling me about all the responsibilities that come with money of this kind. That was a lot of money. I think, since he had other types of projects that were not quite that expensive, it may have been more money than he had for most other things.

DeVorkin:

Was he expecting more from you because there was more money involved?

Harwit:

No, I think it might have been that he was afraid I might be manipulated on the outside, or that maybe I would get too cocky because I had my own money and I didn't need him — or god knows what.. I'm not sure what the sense of it was.

DeVorkin:

He never took you aside and said, why are you getting into the space business?

Harwit:

No, I don't think so. Initially when he really had hoped that I would work on his problems there was some of that. But once it was clear that I was going to work on my own things, I think he must have decided it was okay; not everybody had to work for him and that was fine.

DeVorkin:

Okay. Is there any major element of the history of astronomy at Cornell, space science at Cornell, that we have not covered that you feel we should?

Harwit:

I think there is one; namely, I think — maybe this is a faculty member's position — we have had exceptionally good relationships with the students we have had. I think the students we have produced have also been exceptionally good. There was a recent evaluation of science in the Physics Department. There wasn't one of Astronomy. Among the physics departments in the nation, in terms of stature (by which they meant numbers of Nobel prize winning physicists, that sort of thing) the faculty at Cornell came out around fifth in the country. But in terms of the best place for a student to study physics, Cornell was listed at the top. It's a fairly fluid system where our students and the physics students take somewhat similar courses, work on similar things, or can transfer from one field to the other; Cornell is fairly good about that. I have a feeling that the students that we have also are in a good situation where they don't get involved in any kind of faculty disputes. There are not very many disputes in the first place, and when there are some we do our best to keep them off the students. The students are encouraged to take a lot of responsibility.

DeVorkin:

But not get involved in politics.

Harwit:

They take a lot of responsibility in the sense, for example, that my student, Gordon Starey, now and also Paul Viscuso, who is a year behind Gordon in his studies at Cornell, have been out at Ames Research Center now for three weeks running the Lear jet series. I talk to them on the phone maybe five times a week. We discuss how things are going, what should be done and so on. But they can do all that and I trust them. That's not done everywhere. They are slowly brought into the position of taking responsibility. You have to make it clear that you trust them and that that nobody is going to yell at them if things don't go just right. That works out very well, and I think there's quite a lot more of that at Cornell than you would find at a lot of institutions. I'm very proud of the students that I have had over the years, who have worked their way and have become better known people in the astronomical community. There is Mike Werner, who is at NASA Ames and was at Caltech before; Judy Pipher who I think is just going to be made a full professor at Rochester; and Ted Gull who is at NASA Goddard, Jim Houck had good students, also. Tom Soifer (Baruch Thomas Soifer) is perhaps the best-known one. He is at Caltech now.

DeVorkin:

So really, what you're talking about is the teaching aspect at Cornell has been very strong.

Harwit:

Graduate teaching, yes.

DeVorkin:

Graduate teaching. Let me turn this tape over and ask you, why.

DeVorkin:

This is Tape No. Four, Side No. Two. Why is Cornell strong, if you can give me a brief.

Harwit:

There are a couple of things. One is that all of the faculty people, with relatively few exceptions, came in through physics or engineering. Ed Salpeter is a physicist; I was a physicist. Tommy Gold studied electrical engineering. Jim Houck was a physicist. A number of people did have astronomy degrees; Frank Drake, Yervant Terzian, and Carl Sagan did. But even among those people there was a good understanding of physics. One of the things that we insisted on was that the graduate students in their first year take just about the same curriculum as the physicists, even in the second year also, taking usually only one astronomy course. They take quantum mechanics, electromagnetic theory, and statistical mechanics with the physicists.

DeVorkin:

So you have more than two years of graduate courses?

Harwit:

We usually have two years almost full time. A third year graduate student will take one course in astronomy, usually a seminar each semester. In his first year he might take radio astronomy, which is fairly popular, and which is a full-year course. The second year he might take a course in one of the specialities, or one of his specialities, or take stellar structure or gravitational theory.

DeVorkin:

Whereas, this produces a good physicist, it makes me wonder how broadly based the astronomical training really is at the end?

Harwit:

I think the most difficult part to learn in astronomy is the theoretical basis. Then by getting involved with an actual problem, you also learn the phenomenological basis. Or at least, that's been the approach we've taken, I think, pretty much by common consent. You get interested in a problem, you learn the phenomenology, possibly you go some place for a short time where people are expert at that to make sure that you haven't missed anything, and then by doing you learn that discipline. At first you make some mistakes, but eventually ycu are able to master the phenomenology as well as the practicioners in that field. It takes a half a year, usually to come up to speed, and knowing what other people are interested in, what the main facts are. After that you still make mistakes, but at least you have other talents, the physics and mathematical depths, so that you can make up for it with your own particular approach.

DeVorkin:

I see. Let me ask you one other question. We have covered it in a way, but this is again from Joe Tatarewicz: Were you involved or aware of events around the resurfacing of the Arecibo dish at all?

Harwit:

All I knew was that it was going on. Frank Drake was in charge of that and handled that all by himself.

DeVorkin:

Was there any DOD involvement in that at all?

Harwit:

I wouldn't be surprised, but I don't know.

DeVorkin:

Do you know DOD relinquished the dish to NSF initially?

Harwit:

I don't know whether that wasn't part of the Mansfield Amendment.

DeVorkin:

Okay, fine. Now, let's move on to two final questions. The final major area of questioning is on Space Telescope. In any of your advisory roles at NASA, have you been involved with deliberations concerning Space Telescope?

Harwit:

No. I haven't had all that many advisory roles. I have never been that much of a committee person.

DeVorkin:

What are the roles you've had in NASA?

Harwit:

In the 1960's I was usually on an advisory panel that dealt with infrared astronomy. In the 1970s when Jim Houck, who initially had been a post-doc with me after doing physics at Cornell, got to be a professor, I asked him whether he would like to sit in on these things; because I used to get pretty upset by them. Having been listed at the bottom of the rocket list, of the list of priorities, for as long as I could remember, I thought it might be quite nice if somebody else was in on that. He didn't seem to mind it as much as I did. By that time we were out of rockets anyway, pretty much. So he did more of that. The most heavy involvement I've had recently has just been when I was chairman of the Users Group for the airplanes, and that was just for about the last two and a half years. Occasionally I've been asked to sit on one or another committee, but I've never been that much involved. Not out of principle or anything. I just wasn't asked.

DeVorkin:

Did you support the early concept of including infrared instrumentation on Space Telescope.

Harwit:

Yes, of sure. I was for that.

DeVorkin:

And were you involved in any way?

Harwit:

Judy Pipher submitted a proposal, which in fact got the highest grades of any of the infrared proposals. Unfortunately, on the first go-round they decided that they would not have any infrared.

DeVorkin:

Why was that?

Harwit:

I think partly it was that the optical people felt that they hadn't really had a satellite of their own, and that they always had been dumped on, especially considering the size of the community. Then also partly, I think, because it would have required some kind of cryogenics on there most likely. That would have complicated matters. But I think it was mainly the wishes of the optical ultraviolet community. They felt that it just was overkill for the infared. Perhaps there is something to that.

DeVorkin:

In looking back at all the work you have done on instrumentation, what do you feel is its relative role in doing science? Does the science become possible on balance because the technology exists.' Or does the science drive the production of new technology?

Harwit:

I think it's the first. Most of the things that we have done have been possible because we were able to get detectors that had been pioneered by the military. A lot of the techniques that we have used have been pioneered by the military. But once we use them, we often are able to do things which are much more sensitive than what the military had done. Not more sensitive work, maybe, but work, for example, of higher spectral resolution than the military has done, as far as I know. Generally that's in an area where the military feels, either rightly or wrongly, that it doesn't need that sort of expertise or capability. Sometimes I felt that we had instruments that were really a lot better. For example, when we built the first liquid helium-cooled telescopes. That was something that the military, as far as I know at least, never had. It is possible that they did, but I never heard about them. Later on the Air Force Cambridge people also flew cryogenically-cooled telescopes. They did it also for astronomical purposes. IRAS has solved the problems of doing cryogenic work in space for long periods of time with liquid helium. Eventually it might very well be that that kind of thing would have practical applications, for example, when you get to the point where you want to keep computer components at liquid helium temperatures, or if you wantt to have a computer in space that possibly, digests meteorological data before sending down an otherwise overwhelming stream of data that you couldn't handle with the data links you have, this kind of stuff. I think the time lag may not have been long enough to judge whether there will be a sufficient flow from astronomy technologically into other areas. But certainly there has been a very strong flow from the military, and also from the communications industry into astronomy, computerization and all that.

DeVorkin:

Getting back to IRAs, I know you said that you had written proposals for developing an infrared satellite for quite sometime.

Harwit:

Yes.

DeVorkin:

But in this particular case you were not involved in the initial proposals?

Harwit:

No, that's not true at all.

DeVorkin:

Right, can you clarify that.

Harwit:

I was head of one of the proposing teams for IRAs.

DeVorkin:

Do you mean there were competing teams?

Harwit:

There were competing teams, yes. And I formed a team of maybe six, eight people who had different areas of expertise. These were all young people who were really very good, I thought. I tried to get in also people like Gerry Neugebauer and Frank Low, but they wanted to have their own team.

DeVorkin:

Yes. So there were at least two other identifiable teams?

Harwit:

No, they joined forces.

DeVorkin:

Oh, that's pretty insurmountable.

Harwit:

No, it wouldn't have been insurmountable, because they had not done any rocket work. They had never built a cryogenic telescope. It was clear this was going to be one. But they lured Russell Walker away from Air Force Cambridge Research Laboratories to become head of their proposing team.

DeVorkin:

When was this approximately?

Harwit:

This was in 1975, 1976. Actually a lot of infrared astronomers whom I approached to join our team said, oh well, the Low and Neugebauer team can't be overcome; Ed Ney, for example, and Wayne Stein. Jim Houck and I flew up to Minneapolis to try to persuade them to come in with us. They didn't want to. But Low and Neugebauer's team then went around the astronomical commuity and got the support of people like George Field and a lot of others, and added something like 40 names, I think, to their proposal, mostly people who were not infrared astronomers, but people who applauded this proposal. They wrote a proposal which I thought wasn't a proposal at all. They said, well, we might do this or we might do that. We will look at these three techniques of solving this problem, or we might try this orbit or that orbit, or we might try cooling with liquid helium or with pumped liquid helium. They were only going to go out to 30 microns, because they felt going out to 100 microns was impractical. We exchanged proposals after they had been submitted, and I looked at all the various alternatives that they had, two possibilities here, three there. You could have put together 200 different payloads with these different permutations. We wrote up one.

DeVorkin:

This is primarily you and Houck?

Harwit:

Jim Houck, yes, and Ghertz and Hackwell from Wyoming were there, Mel Dyck from Hawaii. Richard Jennings from England was interested, but then the British got into IRAS also, and I don't know whether he ultimately withdrew his name because they were going to come in as a third partner with the U.S. and the Netherlands. Ball Brothers was helping us to prepare the proposal, and they were working also to help the Dutch team; namely, the Dutch team that then eventually prepared IRAS.

DeVorkin:

So the original definition wasn't for IRAS to be multinational?

Harwit:

It could have been. It was going to perhaps be multinational, and the Dutch were suggesting various things. Ball Brothers sort of acted as intermediary between what I wanted to do and what the Dutch wanted to do also.

DeVorkin:

Were they a pretty fair-minded intermediary, or were they loaded?

Harwit:

I don't know. I got along with them very well up to about two days before we were supposed to make our presentation, which was done here at Goddard on a Monday. Either Friday or Saturday I got a call from Dick Herring, who had been the man there in charge of helping on this proposal, and he said that he could not come down for the presentation on Monday, because A1 Schardt, who was at that time at NASA Headquarters, told him that there should be no involvement by an industrial team that would sway things one way or another relative to this collaboration with the Dutch; something like that. This of course was a real blow, because they had given a price. They had made up the pricing on it which they were responsible for, and I wanted somebody from there who would be willing to answer questions on the pricing, if there were any on the peer review committee that wanted to dig into those numbers. But fortunately, there was somebody who had just gone to American Science and Engineering from Ball Brothers, and who had been involved with this proposal, and since American Science and Engineering and I were involved on a Hadamard transform instrument at the time, I called them up, and as a favor to me they sent him down to answer those questions. But I knew that day that we had lost. It was quite clear that Ball Brothers, I think, for one reason or other, might have wanted to disassociate themselves from us. At about the same time when I got the proposal from the Low-Neugebauer group I also was pretty sure we had lost. I knew that NASA never could resist having a lot of names on one side. The moment they had a lot of names on one side they would go with that.

DeVorkin:

So this was a political thing?

Harwit:

So this was really a political thing, and so when I went to present the proposal that Monday I went in and told them what our background was. We were the first people who had flown liquid helium-cooled instruments and had seen astronomical sources. We had gotten total fluxes from a variety of sources in galactic astronomy. It wasn't a complete success in the sense that we had successful rocket payload after successful rocket payload; but there was one time, for example, during one flight were we saw five sources, four of them completely new, in four different colors from 5 microns to 100, twice each, in 25 seconds. It was a data rate which simply had not been matched anywhere. This was a 6-inch telescope. So that demonstrated what you could do, and I felt that we had the greatest expertise in that. NASA had also said that they would reserve the right to pick their own team from the various teams that would be proposing. Now, I figured that the way things would go was that they would pick most of the people from the Neugebauer-Low team, and then that I would probably be a token member from our team, having been the principal investigator on that. I just couldn't see that. I had worked with Low and Neugebauer in the 1960s when they were continually outvoting me on the rockets, and I just didn't feel that that kind of frustration was going to be worthwhile. At the time it was either doing IRAS which I had set up for 10 years to do, by doing the rocket work and proposing time after time what seemed like better versions of an IRAS, or I was going to work on the COSMIC DISCOVERY book and the Hadamard transform book, and then the infrared astrophysics book which I had been working on for some time, and which I always have about one-third finished. I don't know whether that will ever get finished or not. I hope it will.

DeVorkin:

I hope it will have a good historical chapter.

Harwit:

Well, it does actually. You can see. The manuscript is in here. But it's only one-third of a book, so there's an awful lot still of work to be done. I don't know whether I'11 finish it. The field grows so fast, as you are writing that I don't know if it makes sense. At any rate, it was either doing that or doing IRAS. You certainly couldn't do both. If I had to do IRAS and fight people who were as opinionated as Neugebauer and Low are, and at the same time, their having no experience in this field, and our natural antagonism towards each other that had been built up during these meetings in the 1960s, it just didn't seem to make any sense to me to do that. So I said that, if our team won, then we would have several of the people from our team on there. That was clear. If our team didn't win, it didn't make any sense to me for just me to be taken out when all these other people had worked on the proposal also. So I said that if our team did not win I didn't want to serve; that I had talked with Jim Houck and that he was willing to serve, and he could certainly bring all the expertise that we had built up at Cornell over a decade with him and make it available to the IRAS team. So they said, thank you very much, and in fact they did have Houck become a member of the team. I wasn't too upset. I was sorry. I had really spent 10 years working towards that, and I felt pretty upset about that. But the time I really got depressed was two years later, when after having spent a year or year and a half on committee meetings, which were international now with the Dutch and the British, a fairly weighty team of people, they came out with the final description of what the payload was going to look like. It was exactly what we had proposed.

DeVorkin:

You weren't involved in any of those deliberations?

Harwit:

I wasn't involved in any of them. Jim Houck was, and he used to tell me from time to time that it was looking very much like what we had proposed. I think part of it was just his steady approach; he's very quiet, and he probably was able to persuade people that this was the way to do things.

DeVorkin:

He had the experience, after all. Ball Brothers ended up building it after all, didn't they?

Harwit:

That's right.

DeVorkin:

Perkin-Elmer made the optics? I

Harwit:

That's right, yes.

DeVorkin:

There was nothing very peculiar about the optics. It was really the satellite that was the new.thing.

Harwit:

Yes, I think the cryogenic hold time, and also the extreme sensitivity of the detectors that they had were new. I talked to Rainer Weiss, who had been on the judging committee for these proposals, and who then later on also was for the first year on the IRAS team — he later switched off because he was going to be on the Cosmic Background Explorer, and didn't want to do both, or couldn't do both. But he came up to me one time and told me, you know, it came out just like what you proposed. I said, yes, I saw that, but why did your committee not give us the award then? I guess I will have to say here that he was the only person on there who had much experience in the infrared on the judging team. Everybody in the infrared community was competing. Ray 'e7eiss was also, incidentally, but because everybody was on one of the proposing teams, they needed to have at least one person who was knowledgeable on the infrared. I think Ray Weiss was the only one, if I remember correctly, who was on that deciding committee. At any rate, I then said, well, why did you guys not give us the award? And he said, well, your proposal looked so rigid. Everything was defined in it. We felt that you had prejudged the whole situation, and that your were not going to be sufficiently flexible. I had spent six months on that proposal, almost full time. I was doing other things also, but most of my thinking went in to that. We had looked hard at what the various options were, and we picked what we thought was the optimum. There weren't too many ways of going. We kept finding ourselves driven into this particular mode of doing things. Now, you know, there were one or two differences between our proposal and the final design. We were going to include some polarization capabilities in ours; that was left out of the IRAS. But we had a little spectrometer aboard. They have a little spectrometer aboard, also in the focal plane just where we were going to have it. They have these arrays of detectors that we had. They go out to 100 microns, which we said we were going to do. Low and Neugebauer's original proposals didn't go out that far as I remember. We had the orbit picked out that IRAS eventually used.

DeVorkin:

You didn't have this software package for looking for fast-moving objects, did you?

Harwit:

No, but that was a much later thing. That sort of thing was not asked for in the proposal at all.

DeVorkin:

So basically the definition was very similar. That's an interesting story. I wasn't aware of that.

Harwit:

Yes. Well, that was very painful actually, at the time. But in retrospect, after about four years of the thing going on, seeing Jim Houck going back and forth across the country 17 times in one year to California from Ithaca; and the books I was writing coming out slowly, it seemed to me that I had probably made the right choice.

DeVorkin:

Not to be a part of that team?

Harwit:

Yes. Because I rather like the individual work where you have the chance to have your own ideas, and to hone them and to polish them, and to bring out something which is really new. I have been really very pleased with the response that COSMIC DISCOVERY has had. The Hadamard transform book has also been satisfying. It's had almost zero response. It's a much smaller sort of a thing. But you know, I still enjoyed it. We did something which we were able to finish off. It has a certain finality to it, or polish, and it was a sort of a neat job, I thought. I don't know if one can get quite that same satisfaction out of something where there are so many different hands in the pot. I'm not sure. I've never done it.

DeVorkin:

In that regard then, the last question for this session. As you look back on your career so far, what would you identify as the major contrast in how your career has changed. And what would you look at as being the most satisfying project that you have been involved in or completed?

Harwit:

I don't know about the first question. It is hard to say where changes occur or when. I could tell you the pieces of work that I found most satisfying, perhaps. Some of them we haven't even mentioned, because they are sort of little things that never went anywhere further. The initial things on galaxy formation in the steady state universe gave me satisfaction. Then later on, I did some work on interplantary grains, and what I call boulders, things which you can't see very well, but they would be enormous rock sizes. I did some work on that which people found interesting, because it was something that wasn't directly visible, but you could trace influences of it. I did some work tracing the origins of the zodiacal light and zodiacal dust particles, and I enjoyed that, because it was something I could do by myself. And it seemed to have some repercussions in the community. Some of the instrumental things I did I enjoyed, for example, the fact that we built those first liquid helium-cooled telescopes, and that they worked. We had some wrong results that we reported and had to retract. But we also had some right results which were nice, and showed a capability which hadn't been there before. That was satisfying. The COSMIC DISCOVERY book was satisfying. And then occasionally the modeling of new things, or sources in the infrared that one would not have expected, perhaps; for example, a young student at the time, Kris Davidson, a first year student at Cornell and I sketched out what very young stars ought to look like when they were still surrounded by what we called a cocoon of dust, and also had a small ionized region around them. We described that all before the first one was ever seen.

DeVorkin:

The Becklin Neugebauer objects?

Harwit:

The Becklin Neugebauer objects are just like that.

DeVorkin:

I'll be darned.

Harwit:

These are the sort of things that you quietly feel proud of. None of these things get cited very much; people have a short-term memory. But usually we predicted some of these things, well, 10 or 15 years before they actually came out, and the typical memory is about five years. Then I did something with Franco Pacini on infrared galaxies as being very likely caused by bursts of star formation, and perhaps even the quasars, being such bursts. There is an awful lot of that being heard these days. But again, that was done in 1975. Actually that is something that other people were talking about roughly at the same time, independently, but they then followed it up much more. I think there is something in science where just writing one paper doesn't gain you very much, that people tend to go with the person who has written many many papers on the subject. You might think at first that that is unfair. I used to think that it was unfair; but on the other hand, the person who writes more papers often has put a much bigger investment into it really. I have usually written only one paper on something like infrared galaxies because I feel, well, I could write more, but right now, it's not worth it; there is not enough data to make a really good case. So I will sometimes work on something which throws out an idea that I figure five or 10 years from now will be either substantiated or shown to be wrong. And that's it. Other people will tend to work that idea up; for example, on this star formation bursts, Beatrice Tinsley did an awful lot of work afterward. She wrote a lot of papers. It didn't go much farther, I thought; but she did a lot more quantitative stuff, which I wasn't sure the data necessarily were sufficient to support. But she was a theorist, and this was a good thing for her to be doing. So most of the time, I think, if people think about bursts of star formation now as sources of infrared galaxies, they would quote a Tinsley paper. Occasionally, but only very seldom do I see the paper with Franco Pacini and me cited. Another thing that had some impact — again, one of these odd things where I only did one — was a paper that Ed Salpeter and I did on gamma ray bursts as produced by comets falling onto neutron stars. At one time that was regularly cited as being the most zany of all the different models. But two or three years ago, Stirling Colgate and Albert Petschek at Los Alamos picked it up again and changed it slightly, making it asteroids instead of comets, which I think actually is a mistake, because asteroids, at least in our solar system, have almost circular orbits, whereas the comets have these very eliptical ones; and so the possibility of a comet striking a neutron star is, I think, much more likely than an asteroid.

DeVorkin:

We certainly have seen a lot of comets striking the sun recently.

Harwit:

That's right, yes. There is one event which is regularly cited as being somewhat different from other gamma ray bursts, which had a number of repetitions, sort of spikes a few seconds apart, which was one of the predictions that we had made. That one should see it as the comet orbits and comes in, that different parts of it would get sucked into the neutron star. So it may be that there are some neutron stars that produce gamma ray busts in that way. That was quite interesting. Again, it's one of those things where it happened right at the beginning of the growth of the field. And we just got out of it, both Ed and I. There didn't seem much more that could be said at the time.

DeVorkin:

Yes, I understand. That's fascinating. You've had a lot of fertile ideas. You've looked at them and realized that they are interesting, but they are premature, maybe.

Harwit:

That's the point. I think that's exactly the right word. Most of these ideas are premature. If you followed them up you would have to put in an investment which you yourself don't think is the right thing to do at the time. And then of course you are not recognized for it later, which has its own justice. It doesn't bother me too much. Although most of us furtively look at the list of citations at the back of papers to see whether other people recognized the ingenuity of what we've done. Then it's often a little bit of a letdown if the competition is cited rather than you. But you know, I think that's all sort of dotting the "I's" and sort of trying to get that pat on the back which (chuckles) one always is looking for.

DeVorkin:

Is that a driving element, or are you just looking for it?

Harwit:

I don't think it's a driving element for me. But I am looking for it, sure. I don't think it's a driving element for me as much as it is, perhaps, for some other people, partly because I have almost always worked on oddball things, where recognition, if it came at all, came 10 years later, or 15 years later. I got to the point where I decided that the other people's approval was probably just not that important a factor. You just couldn't do the sort of things I wanted to do, and expect the approval, both. That just isn't compatible. If you want the approval, you have to work on something the community already recognizes as important.

DeVorkin:

That's very interesting. Possibly, you might bring up these same kinds of questions when you talk to Alpher and Herman.

Harwit:

That's a good point, yes. Motivation is a very complex thing. But I do think it is different for different people.

DeVorkin:

That's exactly true. But it's interesting to see what motivates creativity, true creativity. In this case, I think it is something for you to concern yourself with in your study of theoretical discoveries, discoveries in theory.

Harwit:

I hope to, yes.

DeVorkin:

I'll keep that on the tape so it will remind you. Thank you very, very much.

Harwit:

I thank you.