Notice: We are in the process of migrating Oral History Interview metadata to this new version of our website.
During this migration, the following fields associated with interviews may be incomplete: Institutions, Additional Persons, and Subjects. Our Browse Subjects feature is also affected by this migration.
Please contact [email protected] with any feedback.
This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.
This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.
Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.
In footnotes or endnotes please cite AIP interviews like this:
Interview of William V. Houston by Gerald Phillips and W. J. King on 1964 March 3, Niels Bohr Library & Archives, American Institute of Physics, College Park, MD USA, www.aip.org/history-programs/niels-bohr-library/oral-histories/4682
For multiple citations, "AIP" is the preferred abbreviation for the location.
Postgraduate work at University of Chicago; early work in spectroscopy using the Fabry-Perot interferometer; studies of e/m and hydrogen fine structure. Study at Universität München with Arnold Sommerfeld and the development in electron spin research in the 1930s; work with and impressions of Werner Heisenberg and others. Later work in solid state; interest in quantum statistics and its relation to statistics of ensemble. Discussion of major problems in modern physics; teaching methods and responsibilities, administration and research, solid state developments.
Well, let’s see, Professor Houston, to start off the interview, I was most struck with the work on your double Fabry-Perot interferometer. I wonder if you would tell me something about the history of that.
Well, that was just an attempt to try to get more spectral range out of the interferometer. I got interested in interference devices as a student at the University of Chicago because I did my first work with one of Michelson’s interferometers and with Michelson’s first big echelon grating. It wasn’t, by modern standards, very good but it was the first one that we had. And eventually we went to the Fabry-Perot interferometer for studying the fine structure of hydrogen. And the difficulty, of course, with the Fabry-Perot is, as one increases the separation in order to get greater resolving power, the spectral range between one fringe and the next becomes smaller. And so it pretty soon got so small that you couldn’t get the hydrogen line in it. And just thinking about it, it seemed to me that if you put another interferometer on with smaller separation to get a large spectral range, then you could put the second interferometer on and get the resolving powers. Well, it turned out to be moderately so, and we built one; we built one at Caltech. This kind of instrument was always hampered by the fact that you had back surfaces on the plates. And these produced a lot of spurious reflections although we tried to keep them down — and didn’t have at that time the good non- reflecting coatings that are available now, so the reflections were there. So the thing was only moderately successful, but I did use it a little bit. And I was especially interested shortly after that in going to Germany to study. I went around by way of the Volta Conference in Pavia and happened to sit across the table at one dinner from Professor Paschen who told me about a German who had invented a similar kind of device. But Paschen was inclined at the moment to give me credit for it because I had published both the account of the construction and what he said was an adequate theory of it. And I was always pleased to find that he had read about it. But it was never used very much, as far as I know.
It seemed to me a very, very clever idea. In all of the work that you did, starting back at that time and then continuing on for a number of years thereafter, in the fine structure of hydrogen and helium, it would be interesting, it seems to me, if you would review how you recall the history of the theoretical development of this subject that, of course, finally culminated, I suppose, in the Lamb and Rutherford experimental work.
Yes, that was the final conclusion to the hydrogen fine structure study. Well, it started in the summer of 1921 when I was a student at the University of Chicago and Professor Millikan asked me to report the work of Sommerfeld on the application of special relativity to the theory of hydrogen which gave the theory then of the hydrogen fine structure. I got interested in it. That was the last summer that Millikan was at Chicago. He then went to Caltech permanently at the end of that summer. And I did my research the following winter under the supervision of Dr. Henry Gales who was interested in optics. And so I took it up because of the interest I got in the theory from this paper of Sommerfeld’s and from Sommerfeld’s book which everybody read in those days. I did a master’s thesis on it, which didn’t do much more than identify the doublet really well. Then I continued it at Ohio State with the Fabry-Perot interferometer and I guess I used a prism for the other dispersion, and pursued this to the point where it was quite clear that there was more than a doublet. The other components couldn’t be resolved; I always thought of them as — thought of it — as merely a third component. And I was trying to associate that with the selection rules that were normally used at that time to describe transitions in spectra. The results didn’t fit very well. Then at the last minute — that was after I had finished at Ohio State and had gone to Caltech and had redone the measurements with a new Fabry-Perot interferometer that they built for me — news came of Uhlenbeck and Goudsmit’s suggestion of an electron’s spin, and the necessary revision of the selection rules that accompanied this.
That must have been a very exciting thing.
And so it was interesting to me that just the micro symmetry that I had observed in the shorter wave length lower intensity component of the hydrogen doublet fitted better with the selection rules that you would get from Uhlenbeck and Goudsmit than from the original Sommerfeld selection rules. So I kept up the interest in the hydrogen fine structure. I suppose I was just interested in optical measurements, because I had a new interferometer and I liked it; it did do various things. So I first showed that the Helium spectrum was a triplet spectrum instead of a doublet as had been supposed. And I went on to measure e over m, not only from the hydrogen lines which I did then about 25 or 26 or some such times. But later, I made a more precise measurement of the Zeeman Effect to give a result which was a half percent off the accepted value of e over m at that time, and I thought I made a serious observation about the way people do experiments. Apparently — at least in those days and I shouldn’t wonder if it’s still true — everyone adjusted his e over m apparatus until he got the result he thought was right. And so, up until the time when there were some rough observations at the Mount Wilson Observatory and I should know the man who did them —
Babcock. That’s it — it was Babcock — who made quite an analysis of Zeeman spectra and concluded that the value of e over m was not the accepted value. Well, I went to a lot of trouble to build a large solenoid in which we could get seven thousand gauss over a space about 2 inches in diameter and maybe six inches long with some precision. I spent a long time measuring it, and finally I got, accurately I thought, to a tenth of a percent and measured the Zeeman Effect on a number of elements. At that time — this was later, 1930 or so — there was an adequate theory of the Zeeman Effect so that, except for not knowing quite the magnetic moment of an electron, which was off by a few parts in a thousand, we were able to get the value. As I say it was half a percent different from the accepted value, but after that, everybody continued to get the value that we had. It’s been adjusted a little bit since then.
Was this because the new value as obtained from the study of calcite —
I’m not quite sure what you mean — from the study of X-rays, if I understand —
— in calcite. Well, that was all involved, you see, with the determination of a lot of constants, and Dumond in particular took up a study of the fundamental constants. And calcite was used to determine the wave length of X-rays, and that was involved then in the value of h. And then one had the Rydberg constant which included e and m and h. And these things were all put together, and this was the direct determination of e over m. I can’t quite remember — I think they’re probably about the same time that this work to which you refer was done. And just getting interested in that was the reason that I went ahead then and did an experiment on the viscosity of air, because Millikan’s value of the electronic charge had depended on the viscosity of air, and, apparently, it was off. Although he had students working on it in Chicago for quite a number of years, apparently they never got quite the right done of it, so I found it a rather interesting experiment, because it involved hydrodynamics, which I hadn’t studied much before. We built a rotating cylinder device and showed if you put in the right value of air, the right value of the viscosity of air, that Millikan’s determination of e as quite close to the more recent better determinations.
But your motivation for this research came, more or less, from your desire to follow up this experimental research.
Well, the study of the fine structure was following up the theory that I think is quite clear — the developing theory of hydrogen fine structure, because it was the only spectrum, hydrogen and ionized helium that the theory could predict precisely.
It provided a very important test for the new ideas of the quantum mechanics.
Yes. So it was very interesting to follow that up. Now, these other things sort of came along, because I had the apparatus. In the first place, I guess it was this way. The measurements on hydrogen gave them the e/m because it was the so-called Mitbewegung of the nucleus that displaced the lines and gave me a measurement of e over m that way, but that was the difference between two very large numbers and never looked to me like too good a method, although the results were not too bad, and with the spectroscopic equipment available, the rest of these things sort of seemed to follow along. I think, as a matter of fact, my general philosophy on what you do is something like this. You do something whenever you get an idea. No matter what it’s about. If you get an idea of something you can do and you have the equipment to do it, why that’s what you do.
I noticed in all of your work theirs been a very large amount that was in one way or another associated with the development of the idea of the spinning electron. You just mentioned the importance of this in the interpretation of the fine structure of hydrogen and helium and what not; and, of course, in your work on solids. All of this, of course, has been very important. And your recent work on superconductivity, of course. So I guess that, in a sense, the spinning electron has been very important to your scientific career.
Well, it has, because it was involved in the hydrogen fine structure. I remember on one occasion there was some European woman physicist visiting at Caltech — it may have been Hertha Spooner, although I’m not absolutely sure about that — and she was having a little discussion with Epstein, as to what was the most significant development in that year, which was 1925 I believe, or perhaps ‘26. The two points that were mentioned were first the development of the quantum mechanics, both the Heisenberg and the Schrodinger form and the electron spin. She was inclined to think that the development of quantum mechanics was the more important. Epstein was inclined to suggest that the electron spin was the more important, because, as he said, this was a real physical discovery. Which is the more important I suppose is always a matter of opinion, but the two things were very closely combined together, of course, in the hydrogen fine structure. And when I went to Germany in 1928, I went with the idea of studying the theory of the electron spin, but Sommerfeld rather discouraged me, I think very wisely, on the grounds that Polanyi had tried it. He said Polanyi hadn’t gotten very far, although as a matter of fact, he had got to the two by two matrix theory of electron spin, which we still use quite extensively, and Dirac, I guess at that time, had already published, or was about to publish his equation, which included it, so I didn’t pursue the theory of it any further. I’m sure I wouldn’t have got very far, as I was not prepared with the kind of mathematics that seemed to be involved in it.
I wonder if we could go back just a minute to your decision to go to Caltech. How did you happen to choose Caltech?
Well, I’m not entirely sure that I know, but I think it is probably because of the influence of Dr. Millikan. I went to Chicago in the summer of 1921 and he was there for his last quarter, and I attended his class on quantum theory and saw him in some other connections and was tremendously stimulated by him. He had the knack — at least as it appeared to me and I think to many others — of making the subject of Physics sound like the most important thing in the world, although he wasn’t always so very precise in his teaching. You had to go look up the things afterwards to understand them clearly, but he conveyed this idea of urgency and importance in a way which very few other people have –- at least to me. And I think that was probably the reason why I went to Caltech when I got a National Research Fellowship. I had, I guess, some choice as to where I could go, but I never really considered any place else. Of course, it was a new and developing place. It had the excitement of a new and growing institution. And in those days — in 1925 and certainly through to 1935 — it was a place of tremendous excitement, because everybody from all over the world came through at some time or other. And we had such people as Heisenberg, Schrodinger, and Born, and Einstein, and almost everybody else you can think of come for a month to three months. We had three colloquia every week, and everybody attended them, and it was a very satisfying experience, of course, I didn’t know that before I went there, but I suppose I followed Millikan — that’s probably the best explanation. I’ve never really sat down to try to figure out why I did, and perhaps I couldn’t remember very well at this distance.
I noticed in your autobiography you mentioned Professor Michelson at Chicago. You felt that he made some contributions to your own intention to follow up in Physics.
Well, possibly that’s right. He was my first introduction to theoretical physics, and as I had previously studied Physics, one didn’t make any distinction between theoretical and experimental Physics. I just studied Physics. At Ohio State nobody emphasized the difference. At the University of Chicago nobody emphasized the difference, and I learned my classical theoretical Physics largely from Michelson, whose lectures were magnificent in the sense that they were in great detail on relatively few subjects — a procedure which I liked. So I think maybe that it is true that he gave me some desire to have a precise mathematical formulation of whatever I was doing. It’s true that I got into the real theories of quantum mechanics only after I went to Germany.
That must have been a very exciting time in your life when you went to Munich and studied with Sommerfeld.
Well, it was. It was. Of course, it was a period when one could take this new theory and apply it to almost anything and without too much trouble. You could get out a nice piece of work. They were all wide open. It was just like getting a new technique to apply to all the old problems. I went to Munich because I had started studying Sommerfeld’s book my first summer in Chicago, and in my mind he was the authority on all of these things.
During that time that you were there in Munich you did this first work on the conduction of electrons in metals, and I believe it’s true, isn’t it, that your work treating the electrons as quantum-mechanical waves is one of the first treatments of electrons in metals in that way.
Yes, I believe that was true, and it came about somewhat in this way. When Sommerfeld somewhat discouraged me from studying electron spin, he gave me the proof of his first long paper on the application of quantum statistics to free electrons. He gave me that proof to check through, and I checked it through, and checked all the equations, and learned in that way what he had done and learned about the statistics, which, of course, he approached from a very classical point of view. He did the statistics as I guess it was originally formulated, by the old-fashioned methods of statistics of counting up complexions and so forth and so I learned that kind of statistics a little bit there. And, then, after having read that paper through thoroughly, everybody recognized that the thing that was still missing was something about the mean-free path. The mean-free path had been introduced as an arbitrary function in Sommerfeld’s work, and he suggested —
It was simply a parameter.
It was just a parameter. He suggested that I look up the various kinds of mean-free path, because if you made the right assumption about the mean-free path, you could get any necessary variation of resistance with temperature. And a lot of people had been making rather out and out assumptions of that kind. One had been made by W. Wien who was Professor of Experimental Physics there at Munich. So I looked over these a while, and I found out that if you expressed the mean-free path as an arbitrary power series in the velocity of the electrons, it didn’t change any of Sommerfeld’s results, and, eventually, it developed that the reason that it didn’t was because it was only the mean-free path for the Fermi velocity that counted anyhow.
Well, did you and Sommerfeld realize that at that time?
Well, pretty shortly. I can’t just remember who it was — I think it was Wentzel. Wentzel came to visit — I think this is mentioned in Sommerfeld’s paper — but my recollection isn’t entirely sharp. I think it was Wentzel who came to visit and he pointed out this point, which then was included in the final draft of Sommerfeld’s paper.
Well, you got the T to the fifth dependence of resistance, and —
Then I just thought that one might take seriously the idea of the electron as a wave, and so I looked up Debye’s work on the diffraction of X-rays by crystals as a function of the temperature, and it was interesting that if you put that in for the determination of the mean-free path, that the temperature dependence at high temperatures came out right. It was a function of the temperature, and I was greatly pleased and stimulated when I took that in to Sommerfeld. He was greatly pleased. He got up and paced back and forth across the room, and he said a sentence which I can still remember in German. He said, “Die erste anstandige Bearbeitung des Widerstands gestoes.” Well, for a young fellow looking for some approval that was…
But the trouble was, of course, that this was good and fitted the observations quite well, only if one ignored the zero-point vibration of the nuclei. Sommerfeld then reported this work in his colloquium, and he said that he would report it himself. And the real reason he reported it himself was because my German was probably not very understandable. But he was polite about it and he said that he would report it, because he could perhaps speak German better than I could, and also because he could express his pleasure a little better that way. Well, he did and then it was after I came home in the fall of 1928 that I realized that the resistance would go to zero because of the restrictions of the Fermi statistics.
No matter whether one ignored the zero point, even if one didn’t ignore the zero point vibrations, and so I got out the paper in the Physical Review in 1929, which gave the key to the fifth power law, which I believe is still the current law that is thought of as the limit for very low temperatures.
Had Professor Sommerfeld been speculating about the use of a wave in the metal of the electron in the metal before that time?
Well, I hadn’t known about it, at least I don’t think he had really, because as I remember his reaction to this, it didn’t suggest that he had been thinking of that idea at all. No, his use of the statistics was still confined to the classical picture of an electron as a point, because to a person in those days, and even to me, so much younger, one didn’t think easily of an electron as a wave. It came — it just came awfully hard, and it came only slowly. So all of these early treatments were treatments in which the electron was still regarded as a point, but you had to describe it in some funny way by this particular kind of statistics.
Did Professor Sommerfeld use the model of the electron as a wave in his work with graduate students, or only occasionally? Did he use it systematically?
Well, he didn’t use it in his theory, in his metal theory, there in that first paper. There was no mention of the wave at all. About that time, he started to prepare an appendix to his famous book on atomic structure and spectral lines, and he got out a supplementary little volume, of which I have a copy inscribed by him to me, in which he dealt with the wave mechanics, so he must have been giving it in his lectures. Yes, he was giving the theory as it appears in this supplementary book. He was giving it in his lectures at that time, because I do remember now, it was in those lectures that I first saw some of the elegant ways in which he treated spherical harmonic functions, hermite polynomial functions, and things of that kind. He treated them in his lectures in a very nice and elegant way. He had a tendency, as I think probably all good lecturers do, to gloss over the difficulties. He treated the nice and successful parts. I have always thought that was one of the great merits of his early work. He made the early form of the quantum theory popular in his book, because he made it sound so successful. It wasn’t as successful as it sounds. There are a lot of difficulties with it, and in his lectures on wave mechanics, he wouldn’t be troubled by the difficulties that mathematicians would see in some of his statements. He wouldn’t be troubled to emphasize all of the limitations that had to be put on his statements, but he presented what he did in such an elegant fashion that you remembered it and were impressed by his power. I think he was in that way a wonderful teacher.
Well this is something of a paradox, because Sommerfeld himself was really a great mathematician.
Well, he was much more of a mathematician certainly than Bohr, who never claimed to be a mathematician at all. Well, I think, though, that when he was doing Physics, he probably thought that mathematical difficulties were not of great importance. He was a mathematician. I remember the first time I was in Munich, which was for a few weeks in the summer of 1924, he had his students working on problems of propagation of radio waves. It was at that time or about that time; I think that he was working out much of the mathematics of wave propagation. Yes. That perhaps is somewhat of a paradox, but he certainly had a means of picking out those parts of his mathematics that looked nice — or maybe he made it all good — I was not in a position to judge that, and I, perhaps, am not now.
Returning to this business of the propagation of electrons as waves through a solid, well — shortly after this work of yours, Loch did his work, I suppose, on electrons and solids and treated the periodic potential that — how much of an overlap of ideas and intercourse was there between Heisenberg’s group and Sommerfeld’s group?
Well, I was in Munich from about October until March or April, I guess, 1927-28. Then I went to Leipzig with Heisenberg, and during that late spring and summer of 1928, Bloch was working out these things and Heisenberg was working out his theory of ferromagnetism. I suppose there was some overlap. I couldn’t say that Bloch got his ideas from my work at all. I don’t know that that’s at all true. I do know that — I think it was Heisenberg — came down to Munich sometime during the winter of 1927-28, with the thought that the association between electron diffraction and temperature was the explanation for superconductivity. Now he soon gave up that idea, because this was currently only the ordinary fall-off of resistance with temperature. When I went to Leipzig, I sort of quit for the time being — I guess following my usual policy of doing whatever seems to be handy — I quit the study of metals, although I kept in close contact with Bloch’s work and went into the study of essentially the spin orbit interaction two electron spectra. It was suggested by Heisenberg, as a matter of fact — he suggested it as a follow-up to Wigner’s work on group theory. I didn’t know much group theory. I didn’t learn a great deal of group theory, so I did it by hammer and tongs method, just applying the zero order perturbation procedure, but it came out rather nicely.
That was one of the early examples of what I guess we call intermediate coupling now — the calculation of —
— between the Russell Saunders and the J.J. coupling schemes.
It was possible in that approximation, which was good at both limits, but only moderately good in between, to work out the way the Zeeman Effect, changed from one end to the other. And it really was the beginning of a whole lot of work of that kind, which then led up to Condon and Shortly’s book in which they developed these things at some considerable length. But this again was a case of a problem, which had a relatively straightforward solution by these new methods. Just how much interaction there was between Sommerfeld’s and Bloch’s and Heisenberg’s work, I, perhaps, am not in a good position to say. There was certainly some, because Heisenberg and Sommerfeld were in touch all the time, and I told about my work in Leipzig, and Bloch’s came out very shortly afterwards, at least in the next six months or so. I don’t remember the dates.
How did you happen to go to work with Heisenberg rather than, say, with Born or Bohr? Was this on the advice of Sommerfeld or your personal choice?
Oh, I don’t remember. See, Carl Eckart and I were there in Munich together. Carl Eckart went on to Berlin to study with Schrodinger when I went to Heisenberg. I can’t remember now why —
Just seemed a good thing to do.
Seemed like a good thing to do. I don’t recall that I had any specific reason. I suppose I didn’t go to Bohr because the newer formulations of quantum mechanics were associated with the names of Heisenberg on the one hand and Schrodinger on the other. Of course Heisenberg, Born, and Jordan all went together in a sense, but Heisenberg was the one that was usually put at the front of that triplet for some reason which I don’t well know.
How was Heisenberg as a teacher compared with Sommerfeld?
Well, he was, of course, a much younger man — he was the same age as many of his students — younger than some of them, such as myself, so he operated on a much more informal basis. Sommerfeld had a certain amount of the formality of the old German aristocracy. He came from East Prussia, and was a “Herr Geheimrat” — never to me, at least, in the forbidding sense in which that term is sometimes used — but, nevertheless, he had a certain amount of dignity and formality which he always maintained. Heisenberg, being so much younger, didn’t have that — didn’t give that same impression. Now, I don’t have much of a recollection of Heisenberg’s lectures. I may even not have attended them. We had a seminar every week, after which we all played ping pong. I particularly remember that because at that time I could play ping pong along with the rest of them, but when I went back, for short visits, some three or four years later, they’d been playing ping pong every Tuesday since, and I hadn’t, so I was completely outclassed. I remember that very clearly.
Interesting anecdote I had not heard at all.
I thought I could play ping pong pretty well, and in fact, on one of those trips I was coming back on the boat with Helen Wills and they had a ping pong tournament, and I did all right. I got to the finals with Helen Wills and halfway through I was ahead of her, but then she decided that her bat was bent and called time out for a new bat. By that time she had recovered her poise or something and went on to beat me.
You were batting in good company.
I thought I was batting in a pretty good class. But you know, Heisenberg was just a whiz. After a few years of this he could play a wonderful game of ping pong. So my associations with Heisenberg were much more personal and concerned specifically with what I was doing, and I don’t recall his lectures at all.
Was there much informal guidance from Heisenberg? Did he more or less leave you alone? And discuss other things?
Well, in this particular problem that I was working on, he suggested it to start with. After that, I don’t recall very much guidance. I just steamed ahead and worked it out. The process was fairly obvious. It took me a long time to do it, just because I was unfamiliar with it, and sort of slow at it, but I finally got it done, and he felt it had departed a little bit from his first suggestion, which was to follow up the group theory. I didn’t follow up the group theory. I worked instead on a specific problem, but he thought that the evaluation of the Zeeman Effect for all stages of coupling was a rather nice result. I went back for a visit a few years later and he suggested another problem. He suggested that I try to extend Issing’s model of ferromagnetism. Well I tinkered on that for a summer. This was in ‘32 or so and I made absolutely no headway. I had been pleased to learn — pleased in one sense that nobody else could make much headway on it either — that it turned out to be a very difficult problem in combinatorial analysis and people who had approached it had done it another ways — and it had become quite complicated. So I guess my stimulation at Leipzig was as much from the other students there, which included Bloch, of course, principally. Bethe was a student in Munich when I was there at that time. He didn’t take very much part in this work on solids. If I remember rightly, he went up in the mountains that winter once and froze his feet so that he was laid up for a time, and was out. There was another man at Munich named Haisch, who, I guess, became a mathematician. He seemed to be working on this problem about how many colors it takes to color a map. I don’t know what he did about that. I haven’t paid any attention to that problem.
As a former student of yours and present colleague, I have appreciated for years your interest in the powerful nature of the idea of normal coordinates. Where did you first become interested in normal coordinates?
That’s an interesting point. As I say, I got my first idea of mathematical physics from Michelson. Then I went to Caltech, and I got there at the beginning of the semester, and Ernest Watson said: “Would you like to teach a course?” Well, yes, I’d be glad to teach a course. That was regarded as appropriate activity for a National Research Fellow in those days. I think it’s now regarded as an appropriate activity for most Fellows. “Well, what would you like to teach?” There are two courses that I taught. One of them was called Introduction to Mathematical Physics. I forget what the other course was, but I jumped on this one, and the reason I jumped on this one was because a little while before I had run across a copy of a little book by R.A. Houston, entitled Introduction to Mathematical Physics, and it had a section on harmonic oscillators that I found most delightful. I was so impressed that I wanted to read the book more fully. I never did, as a matter of fact, but the idea of this as mathematical physics rather impressed me. This course at Caltech was being based on a book by Haas of Vienna, and so I took that book by Haas and we went through that, and I made up an elaborate series of problems to go along with it. I made the problems up as I went along, and in the course of that, I came across normal vibrations, and then I read about normal vibrations in A. G. Webster’s book mathematical physics, and it was there that I got so impressed with the power of this idea that I still feel it.
Well, you’ve influenced me profoundly on that and I think you influence many of your students in that direction.
Certainly, always in my course in Mathematical Physics, I have given a reasonable amount of emphasis to that idea and to its great generality.
Of course, then, that certainly bore fruit in your treatment of the normal vibrations in solids.
I had a number of students that worked on that problem at Caltech and there were a number here who worked on it. And I think that particular problem has come to a very nice culmination in the work of Van Hove on the topological properties of the frequency distribution. Yes, I have always had the hope that somebody would invent a similarly good transformation — a good idea for some other cases.
Maybe this is out of the question. At least nobody has. Yes, I suppose it was sort of going back to this work for Sommerfeld at the end of the 1930’s that got me again back onto the idea of solids. I tried to understand the effect of an electric field on Bloch-wave functions, and tried to understand the conservation of the so-called crystal Bloch momentum, and it has always seemed to me that there are conservation laws in there whose extent has perhaps not been fully exploited, although maybe it has been in more recent days.
The pedagogical remark you made a while ago about making up a lot of problems for the students in a Mathematical Physics course I believe you’ve always rather held the notion that students should work quite a lot of problems.
Well, that was quite in keeping with the philosophy at Caltech at that time and I think maybe I went further than anybody else in making the course in Mathematical Physics essentially a set of problems with some comments. And I wrote my text with that idea in mind. And I still think it’s a good idea, because I don’t believe one learns Physics by memorizing a series of statements, but by getting a sort of a feeling inside himself as to what you do when you meet a problem.
How you go about it and how much precision is required in your reasoning.
Well, one of the things I’ve wanted to ask you is: “What do you find is the most exciting present topic in physics that you would like to work on, and would like to see others work on?”
Well, it seems to me that Physics is pretty definitely split into what I would have called the objective of the study of atomic physics during the first forty years, say; and the nuclear and high energy Physics. It seems to me that the basic reason why one studied the structure of atoms was in order to understand the structure of tangible material, especially solids. And so that is to my mind the fascinating field, but the reason I find that fascinating is because I think I know something about it. I haven’t kept in close touch with the nuclear field. I can understand the nuclear field to the extent that it uses the same rules that atomic physics uses. The rules of angular momentum and the quantum mechanics as applied to systems of a small number of particles, but the further developments of high energy Physics, I just haven’t kept up with. Now, there are those who say that’s the only field in Physics. I know Rabi, who tends to make extreme statements on many occasions, was saying the other day that was the only field of Physics. Nothing else amounted to anything — perhaps in his mind that is true — but I want to work on solid state Physics because I think I have some insight into it, and because it is a fascinating problem. There’s this fundamental philosophical problem in it as to whether a large assemblage of bodies, by reason of being an assemblage, can have fundamentally new properties as opposed to those of its constituent parts. I think that’s an old question that’s often been raised — and just what its answer is I wouldn’t have any idea — but I find that interesting, I also find interesting the things that come up in teaching — in particular, the question of the appropriate formulation of quantum mechanics, just ordinary, simple, and unrelativistic quantum mechanics. I think that, so often, in teaching one does what is proper. He glosses over the difficulties and concentrates on the successes and effectiveness of it, but I have found myself in recent years very uncertain as to statements that are made in nonrelativistic quantum mechanics — uncertain as to their generality. How much can one talk about the equivalence of the operators for momentum, for example, with the idea of Fourier transforms and wave — mechanics? Apparently that equivalence is subject to a certain number of restrictions which, at least, I don’t know. Maybe somebody knows, but these things aren’t apparently getting so much attention. And then I was interested always by the way in which Bohr tried to get at the fundamental significance of the quantum idea, and how he rather complained when he was here a few years ago over the fact that philosophers didn’t pay much attention to this. He thought it would be useful if some of them did. I find every now and then somebody giving some attention to that.
I know that you and I in the past have talked about the difference between quantum statistics and a —
— statistics —
— the statistics of ensemble.
This seems to me a very interesting subject, and I know you have thoughts on it —
— I’m trying to formulate that right now, and I’m trying to see if you and I can’t get into agreement on this point, which we —
I think perhaps we only have semantic difficulties —
Yes, I think that’s right! But so much of semantics is involved in the way in which you think about things, and I have always felt that there was something more to the statistics of quantum mechanics than there is to the ordinary classical statistics. But I have not come across any clear statement of just how much more and just what the limitations are, and so, I’ve been approaching that by asking mathematicians about the mathematical restrictions on some of the things that we say in wave mechanics. The way I posed the question doesn’t seem to get an answer from any of the mathematicians I’ve talked to thus far, but I perhaps haven’t got the question in the form in which they have an answer. But I do find that a fascinating field. As I’ve said, my philosophy is to do whatever I get an idea about, and so it depends on whether I can understand some of these things.
Well, I suppose your main principal interest in research right now is superconductivity, isn’t that correct?
I think that is the right thing to say. It’s been a subject that’s fascinated me since the late thirties, and I thought that probably the paper that I wrote on the‘conservation of momentum in solids might have some bearing on superconductivity, and I think maybe it has, in this sense: that apparently superconductivity is now described in terms of pairs of electrons, which between them have zero momentum.
This is the Bardeen theory?
Yes. Yes, in one sense it is. Although apparently, it owes a great deal to the work of Blatt in Australia — or New Zealand, wherever he was — and, I sometimes think that he’s not given adequate credit for that aspect of the BCS theory. But it’s true that it is the central core of the arguments that Bohr, Bloch and Rorschach used in connection with the quantization of flux in superconductors. But the fact that you stick to zero momentum, really, in electrons means that you’re in a ground state — a very very special kind of state and you assume the kind of coupling that keeps it that way.
I’d like to raise a couple of more personal sort of points. One of them is the fact that during your career, you have always taken responsibilities of an administrative sort — you’ve always been a teacher, as well as being a creative person in research. Perhaps you might comment a little bit on this threefold nature of the scientist in the modern world.
I’ve done just what comes along, it seems. I suppose that teaching seemed to be the appropriate thing to do. My first year out of college, I had a job teaching at the University of Dubuque, a little school in Dubuque. I was the sole Physics Department and a lot of other things too. And, then when I went to Chicago, where I really had my first contact with up-to-date physics research, I got the idea that research was being done in universities — and being done as part of the teaching process. Both there and at Ohio State later, and at Caltech later, during the twenties and the early part of the thirties, the general spirit was of research done as part of the University activity, though for some reasons which weren’t perhaps particularly good — research in an industrial organization was regarded as second-rate, and research in the government laboratories was regarded as second or even third-rate. So it just seemed to me that in order to do research one did teaching. And furthermore, it’s been my experience that most of the problems that I’ve discovered of interest have come up in connection with teaching, because it is when you try to teach something that you find out how inadequate your understanding of it is, and how much there is that you can’t really present in a complete and satisfactory fashion. And so problems for research seem to me to come out of teaching in often a much more sharply-defined form than in any other case. Now as to administration — seems like somebody has to do it. I guess I never really did any until the time of the war, when I was in New York, on a war research job and had a small group working with me. Then when I went back to Caltech for a short time, I was head of the division there of combined physics and mathematics and electrical engineering. But that wasn’t very high power administration. We all worked together, and I suppose I came to Rice because I thought I sensed in Rice, and in the development in Texas much of the same kind of excitement, looking to the future, that I had seen in Caltech in the twenties, and in California in the twenties. It looked like an opportunity for real development, and Rice looked like a place that could very well develop along the lines that I think are largely due, in this country, to R.A. Millikan — the idea of a graduate school being a very significant part of the university rather than a little piece, and the idea that the faculty should participate in the graduate work and research work to a very large extent. Now while I was president here, I undertook to keep up my teaching and research, but it was clearly on a very much reduced level, and so right now I’m glad of the opportunity to devote full time to that, and I spend my time studying those problems that arise in connection with teaching. Part of the teaching, of course, is with graduate students doing research, and so the problems arise there too.
May I ask you a rather general question? What do you think is the proper way to teach?
Well, I think everybody has to teach in his own way. I don’t think there is any one proper way, and I think some ways “take” with some students, so that I doubt if it can ever be expected that any teaching method will be effective with all students. But I think each student, if he is successful, finds several times during his life a teacher and a teaching method with which he is particularly in tune. Now I have known over my school periods several teachers that made a very great impression on me. I remember in high school one teacher of English. It was a History of English Literature. And she made a very great impression on me. My first teacher in Plane Geometry made a very great impression, because of the formal rigor that she insisted on in the proof of geometrical theorems. Now her rigor consisted merely in the verbatim quotation of the authorities to which she referred, that is, the theorems and the postulates and so forth, but the idea of organizing a system here, rigorously connected together, struck me, as a sophomore in high School, with tremendous force, and for that reason I found geometry a most fascinating subject. Then, as I say, R.A. Millikan at Chicago, in one way, by his enthusiasm — A.A. Michelson at Chicago, in another way, because of the meticulous detail with which he worked out derivations in classical physics — impressed me tremendously; and then, of course, Sommerfeld again. So I don’t think there’s any one way, but the way that I have found effective with quite a large number of students is just what Ferry referred to: the idea of making a course an occasion on which the students solve problems, and then we discuss the problems, but I think the student has to participate. That’s why I think a lecture is a relatively futile device, except for certain purposes. A lecture can show some experiments that are not conveniently done by students. A lecture can present certain basic theories that the student then may want to apply. But unless the student takes the lecture material and does something about it, in the form of applying it to situations which amount to research on the part of the student, since it’s a problem which he hasn’t seen before, I don’t think the lecture does him much good. So I’ve never done anything that might properly be called lecturing. The classes that I have consist of discussion of problems. Sometimes I solve the problems; in most cases I like to get students to solve the problems, and I comment on and discuss with the class the significance of the solutions and the difficulties that might arise. Well, this is just the way I do it, and at least many of the students that have studied mathematical physics with me that way have seemed to find that it gives them self-confidence — self-confidence that they themselves can go ahead and do something.
I’d like to return to one of the first remarks that you made. It wasn’t until you, I suppose, went to Chicago, or perhaps really later, when you first went to Germany, that you realized that there was a distinction between experiments and theory in physics. I know that you don’t really believe in separating these things. Your career proves that you don’t believe in separating them. Maybe you can tell me something about that view.
Well, I think you’ve just about stated it. I think that that view has come to this country from Europe. I don’t believe it was widely prevalent in the United States before 1920 — unless of course — I was just thinking about that the other day — unless of course one might consider Gibbs to have been a theoretical chemist. I don’t recall that Gibbs did much in the way of experiments, and yet he invented an air brake, didn’t he or something of the sort once?
He worked on an air brake, and he worked on a form of gear teeth.
Yes, but I guess still that was theoretical — I don’t know whether he ever built any. I suppose he was theoretical. But everyone else that I heard about did his experiments and then put his theory on the basis of those. I think possibly that was more a British tradition too, than it was Continental. And the British tradition was sort of — and I think maybe still is working out special cases. You fix up your theory to deal with the problem that you have. The Continental — at least the German and, I expect, the French was also the idea of being very general in your theory. It sometimes seems to me that many German papers started out with extreme generality, wrote down a few very general statements, and then put on so much restriction that all they had left was a special case. But they recognized the desire for generality. The British didn’t bother about that. They just went right ahead and solved the problem. So I had that kind of background, I think, from the beginning.
Do you feel that this idea is applicable or inapplicable in teaching students physics nowadays?
Well, I suppose that one has to recognize that the theories are getting a great deal more abstruse, and that it takes a great deal more background to understand them at least in any mathematical detail. So perhaps we have to back off from it a little bit. But I do think it is extremely unfortunate that the student studies physics with only theory, if he doesn’t have some kind of a “feel” for a kind of experiment that he’s talking about. Now this is particularly true in classical physics. I think there it’s not so difficult in quantum mechanics, it’s a little more troublesome, because you have to get pretty nearly through your quantum mechanics before you get ready to apply it to any very material or tangible situation — at least when you do it in a deductive fashion, as I tend to do. Now, of course, other people — and I’m sure they think it’s better and it may be better try to teach quantum mechanics in an inductive way, by talking about experiments, and only at the end come to the formulation of the theory. Now that gives the student a good understanding of the physics and a good understanding of what might be called the tentative nature of the theory, but it doesn’t equip him to apply it very well, because then he has to sort of start over again and work back in the mathematical details. Yes, I think this separation —
The inductive way, though, perhaps is a much nearer way to the way one actually does research, don’t you suppose —?
Very much more so, very much more so, yes. Yes, it’s certainly the way one does research, and I suppose my experimental research associated with theory has always been a performance of an experiment for which there was a theory, like this hydrogen fine structure, for example. There was a theory, and then you do the experiment to see if it fits. But the other way, in which you make real advances by doing experiments and then modifying the theory or developing new ones, is a different process. Yes, I suppose I began to realize this distinction first in Munich, where there was printed on the door that this was the Institute of Theoretical Physics, and around the corner they had Mr. Wien’s Institute of Experimental Physics. I would never have used those terms before.
Well, I certainly know that in the training of your own research students, you attempt to train them both in experimental and theoretical research.
I try to suggest that they ought to do some theoretical work in connection with the experiment. That’s quite right. At least they ought to do enough to try to understand what the, experiment is about, and how it fits into the pattern of other experiments. In the early stages, that isn’t always easy, because the experiments don’t always fit very well.
I’ve noticed that quite a number of your experimental measurements have involved a considerable amount of precision — for example, your e/m measurements and your work on the fine structure of hydrogen and such things. It seems to me there’s a tendency for rather crude experiments to be perhaps glamourized these days and hard experiments neglected. Any comments on that?
Well, I always liked precision and precise experiments. It gave me a great deal of satisfaction to get agreements out to three or four figures. I was interested in the Balmer Constant, the Rydberg Constant, because you get it to eight figures. Oh, I guess now they’ve gone to ten figures or more in that, I don’t know. On the other hand, I think it is also true that those are not in general the experiments that need new theories. Now, of course, in some cases they do — such as the Bellmut Relativity Output of small differences. Certainly in recent years most of the new developments in theory have come out of these rather crude experiments. They’ve come out by skillful guesswork on the basis of experiments that are not very precise. And a new experiment is probably never precise at the beginning, so I just like precise experiments. I remember one of the things that Millikan has said somewhere or other with respect to the freshman college text that he wrote. When he went to the University of Chicago, as he describes it in his autobiography, he found that high school students weren’t very well pre pared. We all make that complaint and we have ever since, and he didn’t think there were any very good college texts. So he wrote a high school text, which as everybody knows, was adopted throughout the middle part of the country, almost exclusively. And then he wrote a college text. And I got acquainted with this text, because I was an assistant in the laboratory at Chicago where he used it. And he made the point that he tried to build his text on a series of experiments that were done in the laboratory. And he tried to get experiments that could be done with reasonable precision. Well, they weren’t always done with precision. It depended upon who set up the apparatus, and who fixed it, but they could be done with some reasonable precision. Perhaps I got something out of that.
Professor Houston, did it take a fair amount of experimental confidence to go ahead with your measurements of e/m, when you found that they disagreed with Professor Millikan’s values?
Well, I don’t think he had measured e/m. He’d measured e of course. But e/m was done by a variety of people. Yes, it took some confidence, I presume. But I’d done the experiment.
…Professor Millikan was fairly well accepted, wasn’t he, I mean his results were taken…
Oh, yes, his results were accepted, but they weren’t on e/m, so there was no immediate contradiction there. And then on “e” —
As you say, you did a study later that showed that “e” —
Then I thought, too, in reading his work, that the work he had done on “e” was very carefully done. But he had used a value of the viscosity of air which he got from a variety of places and that seemed to be a source of possible error in his results — systematic error — so I did another experiment on the viscosity of air by rotating-cylinder method, which can be made moderately precise. And that showed then, that his experiments on “e” really led to the same value as was being obtained from the X-ray measurements. The e/m of course, was a half a percent difference, and I could stick my neck out and say this was it. But in the first place I haven’t been too much afraid of being wrong. I’ve made speeches on several occasions in which I said that if a physicist isn’t wrong half the time he isn’t doing very much.
That’s a pretty good batting average.
— that would be all right. So I wasn’t too much disturbed. I remember also I measured the moment of inertia of carbon dioxide, by means of the Raman Effect, with a student, I guess it was. Debye asked me, in one of my visits to Leipzig, how much precision I would put on that. Well, I didn’t put a great deal. I put a limit of — well, I said it was better than five per cent. I think it was better. I think it was about two or three percent, maybe. But it again was a somewhat new value for that constant. Yes, I think all one can do in an experiment is do the best you can. Take care of everything you can, and then say: “That’s it.” I couldn’t always satisfy everybody else. It was in the middle thirties, sometime or other, when I sent in to the Physical Review my last paper which definitely showed the Lamb Shift. And it was returned by the editor of the Physical Review on the grounds that he didn’t believe that I had that much precision. But then, in an issue a few months later, they came out with a paper by Gibbs, showing much the same results, which they evidently published. I think that my paper was accepted later. But I’m not quite sure. I would have to look that up to be sure. But evidently they didn’t trust me as much as I trusted myself.
Is there any research that you have started and that you regret not having followed through on?
Yes — most all of it. That is, I could have continued much more vigorously on the matter of theoretical work, electrons and solids. I could have continued much more vigorously on the question of spectroscopy. I could have continued experimentally on fine structure, which has gone into the by hyper-fine structure, and the matter of the nucleus. I did some work at one time on the surface photoelectric effect, enough to establish a few points, especially concerning the nature of the work function. That I think could have been pursued. I see there’s now some work coming out on the nature of the photoelectric effect and the question of the interaction of the electrons with the solid before they get out, which was involved in what I was doing, because I was concerned there only with the surface effect, and it’s the so-called volume effect that has some further interaction. Yes, I should say I’ve regretted almost everything that I didn’t carry further. Because, well, you get tired of it. You don’t see exactly what to do next. Something else seems to be attractive. So I did that. I think, possibly, I could maybe have followed up the fine structure of hydrogen a little more, and I would have if it hadn’t been for the war. I never had in mind the techniques that were used finally for the precise determination of that, but I did have in mind the idea of using an atomic beam to reduce the Doppler Effect. I thought I’d get an atomic beam through some slits, and then I’d look perpendicular to the direction of motion of this beam, and would hope that the Doppler Effect in the line of sight could be reduced to a point where could really resolve some of these components. Of course, the difficulty there is that intensity gets very low and exposures have to be long and this maintaining of a Fabry-Perot interferometer in an active enough state for a long period of time isn’t always easy, because it tends to get shaken around a little bit. So it’s just possible that I might have gotten a really decisive measurement of this shift, although, as a matter of fact, the value of the Lamb Shift that was worked out by Pasternak from my measurements and those of Gibbs was within a half per cent of the presently-accepted value. Pretty good.
You mentioned your war work. I know that you and I’ve talked about that some in the past — your undersea warfare work.
Well, that was a job of jumping in to do what seemed to be needed. It didn’t involve any of this kind of physics that we’ve been talking about. It did involve normal vibrations because it was concerned largely with vibrations of one kind and another, but it was not so much research as development. It was during that period that I obtained such a very high opinion of the Bell Telephone Laboratories because I was closely associated with many of the people from there. And I thought they did a first-class job, always. I did learn some good classical physics: wave motion, scattering of light from bubbles in the water-scattering of sound, rather, from bubbles in the water— and was able to apply some of the principles of background noise to problems of signal-to-noise level.
Wasn’t it your group that developed the frequency modulated sonar…?
Well, I was connected with that group for a while, but I really didn’t participate in that work at all; now that was started at least at the San Diego laboratory, while I was there for a little while. No, the work I had to do with was principally the matter of torpedoes — air-launched torpedoes and homing torpedoes — and we got some of them built in time to be effective.
Yes, I had some personal experience —
You had some?
Homing torpedoes. They were very successful.
They homed, all right. They homed. They were crude, I guess, compared with standards of today.
Well, they got ‘em out. They called them Cutie-Pies?
Well, they may have been called that name. It was called FIDO for a time, and then FIDO was also applied to a fog-dispersal system that was used out on the West Coast, so they had some other names for it, I guess. And, up to a few years ago, those things were still in use. I shouldn’t say “use”, but at least they were supplied to the Fleet. Yes, I suppose with everybody that interfered a good deal with the development of physics that I might have done. In the first place, it made a hiatus of some years in the people with whom I was acquainted. Since the war I have never felt the same kind of acquaintance, because a whole lot of these people jumped into physics, then, with whom it was not so easy to get acquainted immediately…And all the establishments exploded in size…But it was certainly immediately after the war that this high precision work developed from the work at Columbia very largely. There’s one other thing I had to do with, and this, I think, I’ve never looked up carefully. They also determined a new value for the ratio for the mu — the ratio of magnetic to mechanical moment of an electron — as being a trifle different from 2e/mc. I remember just after that was announced, I got a letter from Mack in Wisconsin, I believe, wasn’t he — pointing out that this just confirmed some of the results that I had published with Kinsler on the Zeeman Effect at the time I was measuring e/m. We measured a number of spectra of the anomalous Zeeman Effect. We had to combine the mu for orbital motion with the mu for electronic motion in order to get the result, and apparently, although I hadn’t looked this up — I never looked it up, I guess — I did remember that we got a result a trifle different from two. And according to Mack, that was the result that was observed. So I’ve always been pretty fortunate in the precision that’s attributed to those interferometer measurements. But the interferometer is a good device, and one gets good resolving power on it. If one is careful about where he measures on the linen — and we often did it with electrometer traces, so as to try to get an analysis line — the results can be quite precise. And for those days, they were very precise.
Well, I expect I’ve told you everything I know and a lot more.