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 Hans Bethe by Charles Weiner and Jagdish Mehra on 1966 October 27,
Niels Bohr Library & Archives, American Institute of Physics,
College Park, MD USA,
For multiple citations, "AIP" is the preferred abbreviation for the location.
Natural radioactivity; ideas of nuclear constitution, size in 1920s; Gamow-Condon-Gurney theory of alpha decay 1928; discovery of neutron 1932; Cambridge as a center of research 1933; early theories of nuclear forces; analysis of short-range nuclear forces 1935-40; reasons for writing Rev. Mod. Phys. review articles 1935-37 and detailed review of articles' contents; beta decay and the neutrino hypothesis; application of group-theoretic methods to nuclear physics 1936-37; compound nucleus model 1936; nuclear models in general (compound nucleus, evaporation, liquid drop, direct interaction, statistical); contemporary knowledge of nuclear physics 1938-39; stellar energy production; energy limit on cyclotron; accelerators and theoreticians; nuclear physics at Los Alamos; post-war conferences; origins and development of the shell model of the nucleus; many-body theory in nuclear physics; current algebras in particle physics; origins and development of the optical model; of the collective model; autobiographical comments on political, social, scientific conditions in Germany and England in early 1930s ; nuclear studies at Cornell after the war; building the H-bomb; the Oppenheimer hearings; work as a consultant 1950-1970; involvement with PSAC 1956; views on disarmament; receipt of 1967 Nobel Prize.
I would first of all like to greet Professor Bethe and congratulate him on the events and celebrations in which the whole physics community greets him as one of the most remarkable leaders in theoretical physics in the century.
I think that's a fitting beginning to the tape. I'd like you to respond.
Well, you are very kind and I am somewhat embarrassed and very pleased.
I'd like to note that today is October 27th, 1966, and we are sitting in Professor Bethe's office with J. Mehra and Charles Weiner listening and occasionally interrupting with clarification questions. Our intention is to discuss the development of nuclear physics as a field of inquiry.
Nuclear physics of course started a long time ago, and I know very little about the start. I know only what you read in the books, and so therefore it is not very useful. I think you should try to get other people who were present in the early days of nuclear physics to tell you about the excitement of those days. Nuclear physics in a way started with the discovery of radioactivity by Curie, then the exploration of radioactivity by Rutherford, which happened in the first two decades of the century. During this time one got an idea how radioactivity worked, that it was a transmutation of the elements, that alpha radioactivity decreases the atomic number by two and beta radioactivity increases it by one. Many people contributed to this, and in the course of time—especially in the '20s—the Rutherford school proceeded to quantitative experiments about the energies of radiations emitted from nuclei and thereby put it on a really scientific basis. They found, as you know, that alpha particles have very definite energies, and it was possible already in the '20s to make energy level schemes of nuclei corresponding to the various groups of alpha particles which are often emitted from the same nucleus. They found also—and that, I think, was discovered a little later—that beta rays are emitted in a continuous spectrum. This gave rise to a great deal of speculation in the early '30s—namely, the question whether the law of conservation of energy was violated in the case of beta radioactivity, and somebody as knowledgeable as Niels Bohr proposed that point of view. Others, particularly Pauli, opposed it; and I will later on tell a little about the way it was resolved. Then it was discovered that there are gamma rays, which were found to be X-rays of particularly high frequency, and these gamma rays, I think at a very early time, were attributed correctly to the transition of a nucleus between two energy levels. Now, the main question in these days was, of course, what the nucleus consisted of; and this question was asked very far back—I think certainly as far back as 1920—and the first idea of the workers in the field was that they consisted of protons and electrons, these being the only known fundamental particles at the time. So the more or less accepted theory in the 1920s was that the nucleus consists of protons, as many protons as its mass indicates, and then electrons to the extent necessary to give the right charge to the nucleus. Well, then came quantum mechanics, and with quantum mechanics came the first triumph of theory in explaining nuclear phenomena. This was the theory of Gamow and of Condon and Gurney of the alpha radioactivity. I forget the date. I could look it up if you want me to.
'28, wasn't it?
I think it was '28—something like that, it can be looked up very easily—in which they explained it in a very brilliant way as a penetration of a potential barrier. This explanation is still as correct as it was to begin with and gave the first confidence to physicists that the behavior of a nucleus could be explained theoretically on the basis of standard theory.
If I may ask you a question here, Professor
Would you say that the first real attempt to explain the nucleus in terms of the proton and the electron was really not the first theory, but the first theory of the nucleus was this alpha decay theory which tried to encompass perhaps saturation and isotopic structure, nucleus sizes, and so on. What would you say was the first theory of the nucleus?
I don't know. I think you would have to dig in the old literature. I have never tried to learn it. I just don't know. It is likely that you'll find something about it in the old book by Rutherford, Chadwick, and Ellis, but I don't know what they believed. Certainly they believed, as you suggest, that the alpha particle was an important constituent of the nucleus. I am pretty sure that Rutherford already believed that the alpha particle itself was composite and was composed originally of protons and electrons.
But that is prehistory.
That all is prehistory.
And in the modern sense, as you said, is the Gamow-Condon-Gurney theory.
The Gamow-Condon-Gurney theory was the first theory that explained something about the nucleus. It did not explain its structure. It did not explain its energy levels. It did not explain its stability. It explained only one thing—namely, the lifetime against alpha decay as a function of the energy of the alpha particles. And I think it has two great points. One is that it was the first theory, as I said, which gave people confidence that something useful could be said by theory about the nucleus; and because these other things we mentioned—protons and electrons and alpha particles, and protons and electrons—certainly were in no way quantitative and nobody could make anything of it which could be dignified by the name "theory." It was the first thing that was really a theory. It explained only a very small feature of the nuclear phenomena but it explained those quantitatively.
This was also, I gather, one of the first applications of quantum theory. When you referred to theory, you meant quantum theory?
No. By theory I mean anything which gives a quantitative and logical explanation of some phenomena. If this can be done by means of classical mechanics and electrodynamics, I don't object to it. I don't think any nuclear phenomenon has this feature, that it can be so explained. But if there were such a phenomenon, I would not object to the lack of quantum mechanics.
I just checked. All the papers—Gamow's paper and the Condon- Gurney paper were in 1928.
The second great importance of this theory is that it was later on found applicable to a lot of other nuclear phenomena—in fact, all cross sections of processes which involved charged particles involved this theory in some form. So that was the first great success. Then on the other side quantum mechanics made even greater trouble than people had had before with the phenomenon of beta decay because you could prove essentially without question that electrons simply could not be in the nucleus, the nucleus being much too small to contain electrons. Essentially one can use a loose form of the uncertainty principle and say that no particle can be contained in a stationary state in a space which is smaller than its Compton wavelength, which is h/mc, which for the electron is about 4 X l0- cm, whereas even the biggest nucleus is less than 10-¹²cm radius. So it was impossible to maintain the old theory, and just to make bad things worse, it was then found that also the statistics of nuclei and the spin of nuclei contradicted the idea of the constitution in terms of electrons and protons. In this respect it didn't make any difference whether you considered the alpha particle as a sub-unit or not. Electrons and protons give the same result as alpha particles, electrons, and protons, and in particular it was found that the nucleus nitrogen 14 has Bose statistics and has an integral value of the spin, whereas it should have—if it were composed of electrons and protons—Fermi statistics and should have a half integral value of the spin. So there were at least three independent proofs that electrons could not be in the nucleus, and if you want, you can add a fourth not quite so convincing, that it was impossible to construct a theory of the beta decay which was in any way similar to the theory of the alpha decay. So at this point I think physicists were at a complete loss, and it would have been entirely impossible to construct any logical theory of nuclear structure. This period ended in 1932 with the discovery of the neutron, and therefore I would like to call everything before 1932 the prehistory of nuclear physics, and from 1932 on the history of nuclear physics. And here perhaps is a good point to pause.
Here perhaps I could ask you a question. When and how did you happen to go into nuclear physics from atomic physics? I would like to know what prompted you to work in nuclear physics, especially since nuclear physics was hardly in a good or attractive state, or did you go into it because it was in a bad state?
I went into it only after 1932, and after that it was in a good state.
Was that a causal effect?
This was a causal effect. Not this way around, but the other way around. After the discovery of the neutron in 1932, it was in a general way clear what had to be done, and so after the discovery of the neutron, nuclear physics came essentially into the state in which quantum theory was after 1926. There was a fundamental theory and you could work with it. I think I am not in any way a genius. I cannot invent something out of nothing. Some people can. I think Schroedinger making wave mechanics, Heisenberg making quantum mechanics, Dirac making relativistic quantum mechanics—invented more or less something out of nothing, or out of very little. This I cannot do. The only thing I can do is to take a subject in which the foundations have already been laid and then try to exploit them.
So given the neutron you thought you could do nuclear physics.
That is correct.
Did you know it at the time? Was this a conscious decision? How did you react to the discovery?
I think I should probably describe the discovery of the neutron and surrounding events, and then I'll put my own decision in that framework. The discovery of the neutron has been described very well in a conference on the discovery of the neutron which was held at Cornell about five years ago.
And I think I cannot add anything to this—perhaps only a little bit of atmosphere. In early 1932 I was in Rome with Fermi, who was extremely puzzled by the experiments which had been done by various people—by Bothe in Germany and by Joliot in France and several other people, which showed that there apparently was a radiation coming from a bombardment of beryllium by alpha particles; there apparently was a radiation which had the most peculiar properties. It fitted in no way the properties of any known radiation. It clearly was not charged particles because it went through everything very easily. But it also clearly was not gamma rays because the absorption in some light material like paraffin was as strong or stronger than the absorption in the same thickness of lead; whereas, if it had been gamma rays, the difference would have been tremendous, with the lead absorbing much more. And also, this radiation seemed to be able to eject protons from nuclei, and that again didn't make any sense with gamma rays. Electrons want to be ejected by gamma rays and not protons. And so it was a most peculiar radiation. The existence of this radiation persuaded Fermi to go into nuclear physics, and in 1932, in the spring while I was there, Fermi determined that he would work experimentally on nuclear physics.
You were visiting him in Rome.
I was visiting him. I was a Fellow of the Rockefeller Foundation. It was my second visit to Rome. I had visited there a year earlier, at which time Fermi was completely a theoretical physicist without any question. But in 1932 while he was still doing theoretical work, he was determined to go into experimental nuclear physics. I knew, and had always known, that I have two left hands, and therefore did not decide to become an experimental physicist. The subject was at that time still too unclear to appeal to me. I think late in 1932 Chadwick found the right solution. I don't know exactly which month, but I'm sure this is known, and the right solution was the neutron. And he very soon did a series of really quite ingenious experiments showing beyond any question that this was a neutral particle of a mass very close to that of the proton. The final paper came out I think only in the spring of '33, and I remember this but not terribly well. I seem to remember that it was only in '33. What I remember is that I was to talk about this in the physics colloquium at Munich but was told a day or two ahead of time that I shouldn't because the Nazi students would demonstrate.
Against you personally.
Against me personally.
Had you had any inkling of this sort of reaction before that time?
Well, the anti-Jewish laws went into effect on the 1st of April and this was later, so I knew that there was something coming and I had been dismissed from my job in Tübingen where I was sort of assistant professor. So I knew something was coming. I knew that I had to leave the country, but I thought that a talk was still in order.
These demonstrations would have been from the members of the colloquium?
Yes, Nazi students.
And you had gone to Munich from Tübingen?
Did you hold a joint appointment that year?
No. When I was dismissed from Tübingen I went to Munich to my old teacher, Sommerfeld, who got some fellowship for me temporarily until he could find a job for me outside Germany.
What month was it that you were dismissed from Tübingen?
How did that come about?
Maybe we should come back to that, during the personal discussion.
Well, since you didn't talk about recapturing the atmosphere of those days, and it could not have been a very pleasant one, I would like to ask very briefly about the circumstances attending the development of nuclear physics in the places you had been—say at Tübingen and at Munich— about these centers which had started prospering. Is that reasonable?
I think that's very reasonable. There was no nuclear physics in Tübingen. There was no nuclear physics in Munich. People were interested only in some news like the neutron as exciting news in an area neighboring to their main interest. There were only very few places in Germany where nuclear physics was seriously done. The most important places were Bothe's lab (I don't remember for sure where he was at that time) and then Meitner (Hahn and Strassmann got into this much later) so it was mainly Meitner. With Bothe was Gentner, who is now still practicing nuclear physics. These two—Bothe and Gentner—really did the only important work on nuclear physics in Germany at the time.
You mean post-1932. Or were they involved before 1932?
They were involved before 1932—already in the prehistory. This didn't essentially change until much later, as far as I know, and it changed essentially only because Heisenberg got interested in the theory, and then he inspired other people to go into experimentation.
It was Heisenberg who first proposed that the neutron was the constituent of a nucleus along with the proton, and I would like to ask you about the discussions that attended this new model of the nucleus in which perhaps you participated yourself, and also about the name "neutron." How did it come about? Was it labeled immediately by Chadwick, or by other people?
As far as I remember, it was labeled immediately by Chadwick; and I think it had been so labeled before it was born by Rutherford.
I think in 1920 in his Bakerian lecture he predicted this.
In 1920 Rutherford said it would be nice to have a neutral particle corresponding to the proton, and that would make the structure of nuclei much more symmetrical and much more understandable. But at that time of course there was no such particle. And I think he called it "neutron" already then.
I think this is covered in the Ithaca meeting. We can look that up.
I don't know, but I'm pretty sure that Chadwick immediately had that name. Now, I did not participate in the discussions leading up to Heisenberg's paper and I don't know about them. Heisenberg's paper, as far as I remember, still came in the same year—'32. As soon as that paper appeared everything became very clear, and it now was the question to get the right theory.
I just want to go back to give you an opportunity to answer the question that you started to answer, and that is about the centers of interest. You established the point that Germany was not such a center with certain exceptions, which you mentioned. This is no contradiction, but in 1932 you felt that this field became defined, at least for you personally. At that time you were between England and Italy. How was the situation in each of those countries?
The undisputed center was England, and the undisputed center in England was Cambridge.
And you were there?
Well, no, I was not there. I had been there in 1930, but at that time I was repelled rather than attracted by nuclear physics because it seemed to me really groping in the dark for energy levels and having not enough evidence and giving a quantum mechanical description of energy levels without really knowing what was going on. So in 1930 I did not work on nuclear physics while in Cambridge at all, and I had no idea that it ever would interest me. In fact, I began to be interested really only after Chadwick's big paper in 1933, plus Heisenberg's paper, plus emigrating to England in the fall of 1933. After that I was very closely associated with Peierls. We were both at Manchester University. And while Manchester itself was not terribly interested in nuclear physics, was very much in the air in England; and during this period of 1933 to '4 Peierls and I wrote our first paper about nuclear physics—namely about the disintegration of the deuteron by gamma rays, the photodisintegration. My chief interest at the time still was not nuclear physics but was divided between two things. One was solid state physics, in which I was influenced by Bragg, who was the professor at Manchester, and during that time I wrote a paper about order and disorder in alloys.
This would be Sir Lawrence Bragg.
That's right. Secondly, I was very much interested in the exploitation of relativistic quantum mechanics for phenomena involving radiation, and I mean now electromagnetic radiation—gamma rays—and the result of this was a paper with Heitler on the absorption of gamma rays in matter, their formation and gamma ray emission. At that time it wasn't clear whether this had anything to do with nuclear physics, although it was pretty clear that it didn't really have to do with nuclear physics, but was just a tool in nuclear physics and no more. So there were three different things which interested me. It is very likely in fact that I don't remember exactly when we wrote the paper about the photo-disintegration of the deuteron ...
That was published in 1934 in connection with the International Conference of Physics.
Yes. That's what I thought it was. And this means that we worked on it during the year '33 to '34 in Manchester.
You had worked with Peierls previously in Munich in '27. You marked papers together, he remarked. And he stated that he followed your example, though he thinks not consciously, by accepting a Rockefeller fellowship, going to Rome, studying under Fermi and then going to Cambridge. This implied that it was a rather close relationship.
Yes. We were quite close friends. In Manchester we lived together. He was married and had a child and therefore had a house and lived with them in that house.
Peierls had been a student of Heisenberg?
And where did you meet?
In Munich. He was first a student of Sommerfeld, but didn't finally take his Ph.D. there but then went to Heisenberg to do his thesis.
Getting back for a minute to your visit to Fermi, you said that he became interested at the time while you were there; and I think it would be interesting to talk about that because you said your own interest really didn't get started in a strong way until about two years later. Do you remember the circumstances?
Well, one or two years later.
In '33 you did this work, and so therefore you must have been interested by then. Do you recall—how do you know this about Fermi getting interested? Was this the result of conversations?
We talked a lot. The lab in Rome was a very nice small place where all of the people working there, half a dozen or so, talked to each other constantly, so there were constant discussions. Fermi just was very free in telling everybody that this was what interested him and what he was proposing to do.
Laura Fermi in her book indicates that conversations, or at least one conversation between you and her husband, was in English because that was the best way to communicate.
Well, actually, when I was at Fermi's home we always spoke English because that was the only language which was common to the three of us. However, in the lab we mostly talked German. Fermi spoke German very well.
He had stayed at Göttingen previously?
May I take up the discussion that you introduced in your opening statement about the development of nuclear physics that you weren't interested up to 1932 essentially? Would you again in broad strokes perhaps continue this development of nuclear physics and how it unfolded, again in the light of Charles Weiner's question about the centers, because that relates to the sociology and historiography of physics, and then also relating it to the various models and parents ... ?
Very good. As I said, in 1933 I think Cambridge was undisputedly the center of nuclear physics. There was a lot going on there, and I should mention at this point the building of accelerators because at the same time as the discovery of the neutron—just a few months later, I think—the first machine was built which was able to disintegrate the nucleus, and this was the Cockcroft-Walton generator which got up to about half a million volts and which was then the tool by which the Cambridge group—Cockcroft himself, but equally Rutherford and Oliphant, and then Chadwick—discovered one thing after the other about light nuclei. And the important thing at the time was to get some information about the simpler nuclei, about the light nuclei. I think, to come back to a point you mentioned earlier, Cockcroft certainly should be asked to give a good account of the invention of the Cockcroft-Walton machine and of the first work that was being done with it. Well, from this work we learned many things. One was that almost any nuclear reaction that you can think of which preserves the total number of neutrons and protons, almost any nuclear reaction will go, provided the two nuclei can come together; and whether the two nuclei come together of course is determined by their ability to penetrate the potential barrier. So we again get Gamow-Condon-Gurney. Secondly, we obtained very accurate data on the energy of the various nuclei. We obtained an accurate value of the binding energy of the deuteron, the simplest of all nuclei, and that came mainly from the work of Chadwick and Goldhaber on the photodisintegration. And this was the work which Peierls and I analyzed theoretically. Then we learned that the binding energy of the nuclei of mass 3, the triton and helium 3, was much bigger than that of the deuteron, and in turn the binding energy of the alpha particle, of mass 4, was much bigger than those two. The nuclei of mass 3 were formed for the first time in Rutherford's experiments. They had never been known before. We learned about the difference in mass between these two nuclei—helium-3 and hydrogen-3. We took a long time to find out which one was heavier, but that was enough for us to know that they were almost equal in mass; and then, since the neutron has greater mass than the proton, it followed that the hydrogen-3 was a little more strongly bound because it contains one more neutron; and this could be interpreted in terms of the Coulomb repulsion between the two protons which exist in the helium-3 nucleus. So from this we learned right away two tremendously important things. One is the size of the small nuclei, and the second was that the nuclear forces, the forces other than the electrostatic force, must be very accurately the same between two neutrons as between two protons, and that is now known as charge symmetry. So this came out of the very first experiments of the Cambridge school in 1935 or '34, I think.
May I ask here that it was at this stage that charge symmetry and charge independence ... ?
Not yet charge independence.
I see. It was only charge symmetry that came to be recognized as applicable to nuclear physics.
That is correct.
When would charge independence be recognized?
That was about '36 and I will come to that.
You mentioned—just before you go on here—that this very important work was based on the very early results coming from Cambridge, and at that time results from other places where accelerators were being built and this type of work was going on had not yet become significant.
Although they were doing work, their work did not take effect as this work did?
Right. Then perhaps the most important work at the time was that concerned with the understanding of nuclear forces, and the two most important names in this are Heisenberg and Wigner. Heisenberg in his first paper suggested a model for nuclear forces, and he suggested in particular that they should be similar to the forces in molecules, that they should be exchange forces—exchange forces in which the proton changes into a neutron and vice versa. He didn't have it quite right. The man who got it completely right as far as exchange forces were concerned was Majorana, a very taciturn Italian physicist, who was extremely able but in contrast to other people, did not come to the lab very much. I think I saw him once or twice in the seven months I was there.
Was he a student in that period?
He had his doctor's degree. He must have talked at least during his examination. And he was, I think, about the same age as Segré and I.
Was he associated with Fermi, too?
In a vague way, yes—to the extent that he was associated with anybody. I think the only person to whom he talked a little bit was Segré. Somehow Segrè could draw him out. While Majorana studied the problem of intereaction closely and found that it was necessary to attribute a somewhat different property to the exchange forces—namely, that the two particles should exchange only their positions but keep their spins; otherwise you didn't get any bound deuteron. So this was one line. It should be mentioned that Heisenberg had the idea that there should be an interaction only between the neutron and the proton but there should be no interaction between two protons or two neutrons because no exchange could take place in this case. It was shown in 1935 and 36—and this is somewhat anticipating the history—by Tuve and Hafstad and others at the Department of Terrestrial Magnetism in Washington that there is a strong attractive force between two protons, and thereby this part of Heisenberg's idea was changed, had to be changed. There were forces between any pair of nucleons. Then these experiments of Tuve and Hafstad were evaluated by Breit, who has kept evaluating similar experiments to the present day. He showed that to quite good accuracy the force between two protons is the same as the force between a neutron and a proton in the same spin state—namely, when their spins are opposed to each other. This fact then gave rise to the idea of charge independence, and that I think is in a paper by Condon and Breit—there may be other people involved— in about 1936 in which they postulate a charge independence plus the idea of how to formulate this cleverly—namely, by means of isotopic And the idea of isotopic spin goes mainly back to that paper, although Heisenberg already had a bit of that formalism in his original paper.
He had introduced isotopic spin.
Yes, he did.
May I ask you a question about exchange forces? In the first section of your three-part article on nuclear physics called the "Bethe Bible," under the heading "Saturation of Nuclear Forces," you argued that one had to look for an analogy with chemical forces. The popular one was the homopolar binding. This would lead to the saturation property. You said there, "It is interesting to note that it will also yield exchange forces." Was it in this article that the idea of exchange forces—at least as an important point—was first introduced? One is struck by the importance of the use of analogy in your article. I would like to know whether such an analogy was being discussed at the time, or did you just take it as an example, or had the idea of exchange forces already been established by Heisenberg and Majorana?
The latter. I think that Heisenberg and Majorana had established the idea of exchange forces, and I think that Heisenberg in his first paper—but I haven't read that for 34 years—already mentioned the analogy to chemical forces. Doesn't he?
Well, explicitly I have seen in yours.
But I'm sure he had it in mind and I think it is actually stated in his paper. Certainly, as I remember it, I got the idea from Heisenberg. Whether it's in his paper or not, I'm not sure. It was a current idea, and it was a current idea that it was analogous to molecules, and the idea of exchange forces certainly was discussed by the entire little group which worked on this—Peierls and Wigner and myself and Breit and so on.
This was a scattered group, though. Breit was here. Wigner was in Berlin at the time?
No, Wigner was already in Princeton. The discussions were partly in England, at that time mostly with Peierls and Goldhaber. Let's see who else was interested. There were some small meetings of a discussion group which was organized by Blackett which met maybe two or three times a year to discuss the importance of physics. I'm sure we discussed it in that discussion group. Then afterwards when I came to this country, which was in February '35, I discussed things with Wigner and later with Teller, with Teller also in England, and then with Breit, Condon and a few other people. But certainly the idea of exchange forces was entirely familiar to the group of theorists and also experimenters working in the field beginning in 1933.
Meson theory was formulated around 1935.
In what practical manner did it affect nuclear theory?
Not at all until much later. I would say not until maybe '38 or so, which was really after the first big effort had been completed. It did not affect nuclear theory.
That was also after the mesotron was discovered in cosmic rays which was believed to have been Yukawa.
Yes, not until after the discovery, was it taken seriously by nuclear physicists.
And then it affected nuclear theory.
To some extent, not terribly much.
Would you say that that situation was perhaps like the relation today between the relativistic field theory and particle physics at the present time?
Somewhat similar, not terribly similar. There was the great difference that in nuclear physics we had a lot to work on and we got a lot of answers, I think many more than particle physics has got until maybe the last few months. It is now different. It is now getting somewhere also, I believe. But until a few months ago I think it was not getting very far. I think we did much better and had much more understanding of nuclear physics in the mid-1930s.
About 1935 you knew that the nucleus consisted of the neutron and the proton. You knew about charge symmetry and charge independence and you knew about exchange forces. And you knew that the neutron-neutron potential and the neutron-proton potential were the same ...
In the same spin state.
In the same spin state and to explain the experiments that fitted the data.
Yes. I should now mention two other points which still belong in the same period and were very essential. One point was Wigner's analysis showing that the nuclear forces must be very short range. He did this on the basis of the facts which I mentioned previously about the binding energies of the nuclei containing two, three, and four particles, which binding energy increases very rapidly with the number of particles. On this he concluded that the deuteron was only just barely bound. That is to say, the potential must be very, very deep compared to the binding energy of the deuteron. The binding energy of the deuteron is about 2 MeV, and he concluded from fairly simple calculations that the depth of the potential must be 20 MeV, maybe 40 MeV or so. This was necessary in order to explain the very large binding energy of the alpha particle, about 8 MeV per particle. When Wigner did his first work, he did not know about exchange forces. Luckily, it didn't make any difference because these four light nuclei have a symmetric wave function, and therefore it doesn't make any difference whether you exchange them or don't. The wave function stays the same and therefore you get just about the same result with exchange forces and with ordinary forces. Because he used ordinary forces in this, these are very often referred to as "Wigner forces;" and I think it was during a visit which he paid to England—and I'm not sure what time this was; it was probably early '34—Peierls and I had a long conversation with him in which we convinced him that there are exchange forces and in which he convinced us of his theory. So this was one important point. The second important point about general nuclear theory, were the measurements of the cross-section, for the scattering of slow neutrons by protons. And for this purpose it was of course necessary first to have slow neutrons. They were produced in the experiments of Fermi and his collaborators in Rome, and once they were produced everybody could imitate them and could experiment with slow neutrons. I'll talk about these experiments later on in a different connection. They are logically not important in the present connection. However, the measurement of the cross section of slow neutrons and scattering from protons gave a result which was a complete surprise—namely, Peierls and I had thought that we had calculated that cross section in our paper which I mentioned before and that we had calculated it to be about 2 barns. It turned out 20 barns after some corrections for molecular binding of hydrogen which were only interesting because they led physicists astray for a year or so. Anyway the cross section turned out about ten times the size that we had calculated, and we felt pretty sure of our calculations. This difficulty was resolved again by Wigner, who pointed out—I remember it almost in detail; the conversation took place in the subway going from Columbia University to Penn Station—and in spite of the not very inspiring surroundings, Wigner pointed out that there could be an influence of the spin because if you had a proton and a neutron, clearly you could make of this an object of total spin 1 or of total spin 0. Now, the deuteron was total spin 1, as it had been known for quite some time. And therefore the theory of Peierls and me referred to this case of total spin 1, and therefore all we had calculated was the scattering for total spin 1. We couldn't know anything about spin 0 and so Wigner suggested in the subway that it was the spin 0 which probably gave the large cross section. This must have been in '35 still, and it solved everything. And much much later experiments were carried out which demonstrated that he was right. At the time it was only a conjecture. So Wigner in this way discovered the spin dependence of nuclear forces, and in those days one described it as a mixture of Majorana and Heisenberg forces. It was mostly Majorana and a little Heisenberg. And then it was this spin 0 state which had to be compared to the proton-proton scattering because two protons, when they have 0 orbital momentum can only have spin 0. So when I said before that proton-proton forces and proton- neutron forces are equal, they are equal when you use the neutron- proton forces in the spin 0 state. In the spin 1 state, the neutron- proton force is somewhat stronger. So I think this completed pretty much the knowledge of nuclear forces which could be obtained in those days and very little additional knowledge, only confirmation, could be obtained from then on until about 1948 when the first high energy accelerator came into being.
Just to pin this date down, you described a subway ride. Was this prior to his publication?
I don't think he ever published this.
How widely diffused was this knowledge then?
It was never published by Wigner. It was published by me in the first article in Reviews of Modern Physics, giving credit to Wigner. That's the only publication it ever had.
And so in '35 there was this discussion. And the first publication of the three-part article was in '36. You're saying that there was a period of 12 years—'36 to '48—where the situation in that aspect of it remained pretty much the same. And this takes into account everything that happened in '35—Yukawa's statements ...
Yes. I have to modify it in a few points. Go on.
I just wanted to pinpoint some events. The other point that I was going to make is that the work in cosmic rays didn't seem to materially affect anything in this period.
No. The cosmic ray work never was accurate enough or gave enough data to tell us anything. However, I want to modify my statement in three ways. One is about the detailed description of the nuclear forces, and particularly the forces between two protons. There were lots of good experiments in the late '30s and early '40s on the scattering of protons by protons at energies ranging up to about 14 MeV. All of them were analyzed by Breit and collaborators. The collaborators changed. The subject didn't. And he got a very good knowledge of the potential energy as a function of distance that was acceptable. He found that many forms of the distance dependence were acceptable. He used a square well and a Gaussian and an exponential and a Yukawa function; and they all represented the data about equally well. This result was understood later on after the war by work of Breit himself, of Landau and Smorodinsky, of Blatt and Jackson, and of myself, which is now summarized in the theory of the effective range, which made it possible to put all of these results under one heading and to explain why the results were so insensitive to the shape of the potential assumed. Each potential is described by two parameters—roughly speaking, its range and its depth—and two parameters are sufficient to describe the whole scattering. Well, this was one important additional piece of information. The second, which has already been mentioned, is the use of the Yukawa potential. The Yukawa potential was one of the potentials which Breit could use to fit his experimental data. And of course the Yukawa potential is based on meson theory. This was very satisfactory because it showed that something which could be derived from theory could also explain the experimental data. This is all it did, but this was quite a lot. Third, and this is perhaps even more important—the quadrupole moment of the deuteron was discovered—I forget at exactly what time but I think it was after I had published all my articles. I think it was '38 probably—I don't remember for sure—but it may have been '37. This showed us that we had been still too naive about describing the forces between nuclear particles. We had assumed that they depend only on the distance and then on the relative direction of the two spins. But now we were shown that it also depends on direction and we had to introduce what is called "tensor forces." As you know, and as I'll discuss later on, it finally became still more complicated.
Spin orbit interaction.
Spin orbin interaction. But the tensor forces tied in very nicely with the Yukawa idea—namely, if one took a vector meson or a pseudo-scalar meson, then one could explain these tensor forces and they came out of the theory. I wrote a long and learned paper I think in 1939 or thereabouts on the way this could be used to explain the quadrupole moment of the deuteron. It was much too detailed because the information at that time really didn't warrant such detailed studies. So these are the three additional points.
I would like to ask you about your three-part article on nuclear physics in Reviews of Modern Physics. What prompted you to write these articles because that was not just a compilation. There was something new in them. I would also like to isolate what was new.
The second part may be hard to do. The first part is easy. As I said, in February '35, I came to this country. I found in this country a big community of physicists all eager to do nuclear physics and all eager to learn all about the theory.
Had you expected this to be the case?
I had expected some of that. When I was offered my position here, I was told that a small cyclotron was being built here by Livingston, and that the department expected to go into nuclear physics. I had known of course about the big cyclotron which had been built by Lawrence in Berkeley—big on the scale of those days. I think it gave something close to 10 million volts. Livingston had been associated with Lawrence in this construction. I knew, therefore, that lots of nuclear physics was going on in Berkeley and hopefully at Cornell. But there was much more. Almost everybody seemed interested.
In this country?
In this country. And this was one of the most stimulating experiences really, to come here ...
Was one of the excitements to go to America?
Weisskopf was very excited and interested to come to work in the United States.
Were you prompted for approximately the same reasons?
No. My reasons were as simple as could be. It was the only position which was offered to me which promised any permanence. In England I had obtained , and probably could continue to obtain, yearly appointments with no security and with very doubtful prospects for promotion. Here I was offered an acting assistant professorship, which wasn't very wonderful perhaps on an absolute scale, but on a relative scale it seemed splendid. I had every confidence that once I was here I would get promotions, and even the beginning salary was extremely high on the standards of a refugee in those days, namely, it was $3000 a year. So I came purely for the purpose of finding a secure position.
Now to go back to your article.
I will come to the article very quickly. Now when I came here I found great interest in nuclear physics, and I found, on the other hand, that many of the things which had been common knowledge in the very small circle in England in which I had moved were unknown here, and I also knew that a lot of the things we had done could be much extended by a little bit of theoretical work, and this would be worth doing.
What were the sorts of things you had known about in your circle in England that were not known here? Theoretical? Experimental?
Mostly theoretical and very largely the things which I described to you in the last hour. And I was asked to give many talks. I was asked many questions when people came to my office, and I thought it would simplify my life greatly if I wrote it down once and for all and thereby answered all the possible questions in one big article. So therefore I proposed to Mr. Tate, the editor of the Physical Review and Reviews of Modern Physics that I would write one article about nuclear physics. He said, "That's fine. I'll take it." And then I began writing. After a short time I wrote him, "I have to write two articles, one on the stationary states and one on nuclear reactions." And after a while I wrote him again and said, "No, I have to write three articles." And that's how it came about.
This is 35 that you started?
I think I must have started in the fall of 35. So I began writing this. I had two important advantages. One was that Livingston had a card file of every paper which had been written on nuclear physics— something that I could never manage to get together—but he made this available to me and so this was most useful. The other was that I don't remember exactly from what time on, I had two very able collaborators— Konopinski and M. E. Rose, now at Indiana and Virginia, respectively, who were very eager to do calculations about nuclear physics. Among other things, they did some of the necessary calculations for the articles. They had also quite a lot of out of publications of their own. And so in this manner we got it often. Now you ask what was new. It's a little bit difficult to tell, but I tried at every point to fill in the gaps. That is, I took what was in the literature and then I wondered just how much more was necessary for a coherent presentation of the subject. So in this manner, when something new was necessary, I sat down and did it. I don't think you want me to go through it in detail, do you?
I want the highlights, I'm looking for the crystallization of concepts. You mentioned a while ago about the discussion with Wigner in the subway. Perhaps it formed part of the scheme.
It formed part of the scheme.
We have the table of contents and the bibliography. I Xeroxed them from a bound volume.
I would rather get the articles. [Glancing through articles.] The first chapter of article A is just putting together known things, but they had never been put together in this manner, never so briefly and I think never so completely. The second chapter on the qualitative arguments is a simple statement of the things I have been talking about and a couple more. No, actually the second on qualitative arguments which goes mainly into the evidence from heavier nuclei except for Section 9. Now, in the third chapter the Section 13 on excited states of the deuteron had never been done before; 14, the scattering of neutrons by protons cross-section contains the Wigner idea; 15, the angular distribution of the scattering, was common knowledge, again, but had never been stated and lots of people had said stupid things about it; then 16, photoelectric disintegration is essentially the paper with Peierls somewhat elaborated; 17, the capture of neutrons by protons think is mostly, again, Peierls and myself, plus some ideas which I think were due to Fermi on capture due to magnetic transmission; 18, scattering of protons by protons is more or less what Breit said.
That was a new "deuteronomy."
Now, in IV, the theory of the three-body and four-body objects, Section 20 was, I think, essentially new, and, in 21, I think, I greatly simplified the theory of Feenberg. Section 22, I think, was again more or less common knowledge but put in a quantitative form— namely, getting the radius of the three-body systems from the difference in binding energy.
So three-body potentials have been discussed already at that time?
Three-body potentials were not discussed, no—only three-body systems with two-body forces.
I see. In the formulation of Feenberg's equivalent two-body problem, you were trying to discuss the three-body systems essentially in terms of two-body forces?
In terms not only of two-body forces but also two-body wave functions.
Yes. Was Feenberg working with you?
No, I don't remember where he was but he was not here. #23, excited states of the alpha particle, I think was new. The next chapter V on statistical theory is nothing terribly original but some of the things are new. For instance, the section 30 on "Weizsäcker's Semi-empirical Formulae"—that section is much simpler than his theory and I corrected a mistake that he had made. It's more straightforward than what he had done. Section 27, the "Disproof of Ordinary Forces," is somewhat interesting. I showed in that section—I'm sure I was not the first one to think of it, but perhaps it's the first time that it is put down on paper in a coherent way—that if you had ordinary forces between the nucleons, then the nucleus would collapse or the nucleons would go into a clump of the size of the nuclear forces and the binding energy would be proportional to the square of the number of particles rather than proportional to the first power, similar to what I told you at lunch about Dyson's theory of ordinary matter.
This is like collapse in the theory of gases.
Yes, just so. So I think it's put down coherently here for the first time and it is one of the things that people have forgotten— namely, there are now, 30 years later, some people who still think that they can use ordinary forces between nucleons without a repulsive core and can get saturation. That this is impossible was proved back then. I think it's very interesting and somewhat sad how some things that science used to know just get out of the consciousness of people working in the field.
But you're not implying that this was ignored when it was presented. It was learned, learned well perhaps, but ...
It was learned very well at that time and forgotten in the course of 30 years.
But it seems to have been forgotten again and perhaps reselected afterwards in the theory of condensation. There's no theory available yet, but to try to explain condensation on the basis of entirely attractive forces and the collapse of the volume is the same type of thing afterwards.
The main question there again is whether it is possible to explain condensation either by means of entirely repulsive potentials or entirely attractive potentials. It seems that neither works.
Neither works. You have to have a combination. In this article I made a big mistake which Feynman is very fond of pointing out—somewhere (and I don't remember in which section) I say that you need either exchange forces or you need a repulsive core. The repulsive core had been proposed by Feshbach and Morse in 1935. And I said at that time that it is much too complicated. Forces couldn't possibly be like that. Nowadays we all believe in the repulsive core, and Feynman just loves to rub my nose in this. Now, in the "More Detailed Theory of Heavy Nuclei" [Chapter VI], talk in Section 32 about "shells" of nucleons, which of course is long before the shell model. This was not original. It had been previously suggested by others, especially by Elsasser and it is a critical discussion, but the things which I contributed to this are, first of all, that I took seriously the Hartree model and did a couple of calculations using this shell idea, and this was soon afterwards continued by Wigner and Feenberg in a much more extensive way. But more important—and that was one thing which I was especially fond of in those days—I pointed out experimental evidence for shells. Experimental evidence until then had been very loose and very qualitative, and I had worked quite extensively on obtaining actual values of nuclear masses. And in the process of this, I had better numbers, let's say, for the binding energy of oxygen-16 and so on; and on the basis of these numbers I could show that, indeed, oxygen-16 was more strongly bound than its neighbors, carbon-12 and neon-20, all of which are alpha particle nuclei , and that therefore one could very well claim that in oxygen-16 a shell was closed. And I think this is one new point.
Was it a recognition of the Pauli principle somehow?
The Pauli principle is in this very strongly, and of course the Pauli principle comes in very quickly when you go beyond helium-4. Namely, you have to say: Because of the Pauli principle, you have to start a new shell, and I think we already had the right idea that the next shell will be a P shell rather than an S shell. And because of this, the binding of the next nucleon, nucleon number 5, is extremely weak. In fact, it's so weak that there are no stable nuclei with mass 5. In one of the other articles we go into quite a lot of discussion showing that nuclei of mass 5 are in fact unstable. Some people had claimed to have discovered them.
So the shell model was already in vogue at that time.
It was indeed, but we were only able at that time to get up to mass 40. We could show that calcium 40 should be a closed shell nucleus, and one of the sections  here says: "energy of oxygen-16 and cal- cium-40," these being two closed-shell nuclei. Beyond calcium-40 we got into hopeless contradictions, and at that time we simply had to say that we didn't know how to continue the shells beyond 20 nucleons of each kind.
Wasn't the exclusion principle regarded?
The exclusion principle was fine, but we did not know what angular momenta to attribute to the next nucleon states, and this was only solved in '49 by the introduction of the spin orbin coupling.
And what was your feeling about the shell model before it was in vogue and later on—during this long gap between its original vogue and its resurrection?
Well, I'm just trying to remember. I think I was always confident that indeed the shell model would some day be established. I was always confident that we had the right answer up to calcium-40, but I certainly didn't know how to do it correctly beyond that.
And this was long before Jensen and Goeppert-Mayer?
Long before Jensen and Mayer. But I didn't know how to do right. They did. They knew. Then the second feeling I had—and I was going to talk about this— is that we learned a great deal about nuclear structure in the late '30s which was different from the shell model, if I may say—namely, the Wigner model which talks about the symmetry of the wave function more than the shells occupied by the nucleons and then the compound nucleus. And so in the second half of the '30s these features seemed much more important than the shell-model feature. Nowadays I think it really has become different, and I think the shell model I would now put as the most important feature of nuclear structures everything else being obtained as a modification from this.
It has had a very stormy history.
It has indeed.
I would like you to recapitulate some of the controversies that surrounded the shell model from the beginning.
That would co think better when I progress in history to 1949.
I think we're doing very well systematically going down section by section, and we're only on the first installment. But I think this is very productive. As these things come to you, you're helping us answer the question that you didn't believe was possible to answer—to demonstrate what was new and what was changing.
And indeed the shell model has had a place right here.
The prehistory of the shell model is here.
Let's make a note to follow this up later.
Very good. Now, Chapter VII is about the beta disintegration. By that time there was a theory of beta disintegration, and I haven't talked about that at all because this really went quite independently of the general development of nuclear theory, and it is very independent of the general theory because we are dealing with a weak interaction, whereas inside the nucleus we have a strong interaction. The ratio of the strengths of these interactions is something like (I don't know) 10-40 or some nice number like that. So one can certainly discuss nuclear forces without knowing anything about beta decay, and one can discuss beta decay without much knowledge of nuclear forces. I left beta decay in the year 1930 with the controversy between Bohr on one side and Pauli on the other side: whether or not energy was conserved in the beta decay. In subsequent years—and especially after the accelerators came into operation—we got a lot of evidence on the energies of nuclear states, both ground states and excited states, and we got, for instance, very accurate information on the difference in energy between the ground state of nitrogen 13 and carbon 13 from the study of disintegrations leading to these nuclei. And, in fact, some of this book, Part III, is concerned with the obtaining of good values for these energies. So we knew the energy difference between these two nuclei, and, on the other hand, we knew the spectrum of beta rays emerging from nitrogen 13 in its decay; and from this we could conclude that it was the maximum energy of the beta rays which was equal to the energy releases—that is, to the difference between nitrogen-13 and carbon-13. And this answered experimentally the question raised by these two people in favor of Pauli—namely, energy was conserved but it was not just conserved on the average (Bohr had wanted to conserve it on the average Energy was conserved, but then something had to disappear and this gave rise to the neutrino hypothesis of Pauli. I don't know where Pauli's hypothesis was published by Pauli, and I'm not terribly familiar with the theory of beta decay altogether. The man who probably knows most about it in this country is Konopinski, who has written a book about it. The book probably contains it, but also it might be useful to interview him personally. Now, Pauli's theory was of course put into explicit mathematical form by Fermi, and Fermi's theory in turn is valid to the present day.
This was about l934.
In the same year you published a paper on the neutrino.
Yes, that was a very unimportant paper. The question was raised: could you detect a neutrino after it had been emitted? And the question was: what properties would it have? It obviously had no charge, so it wouldn't make ions. But maybe it might have a magnetic moment, and so this unimportant paper deals with the ionization which one might expect if it had a magnetic moment. Of course no ionization was ever found, and one couldn't set a limit on the magnetic moment, which was very low, and already at that time I think everybody, nearly everybody, believed that the neutrino could be found only in one way—namely, by making an inverse beta process. Peierls and I wrote a paper—and that is more important, although very simple—in which we calculated the cross- section which would be expected for such a process, and that estimate of course later on turned out to be correct. The cross-section was indeed that low, 10-42r so square cm. In spite of this, the neutrino was found, as you know, in the 1950s by Reines, Cowan and collaborators. And nobody in the 1930s would have thought it possible that such a small cross-section could ever be discovered.
You referred to a subsequent paper that you wrote of more importance on this subject.
No, that's probably earlier than this.
I see. The cross-section paper is the early one.
Bethe and Peierls in Nature.
The Nature one is called "The Neutrino."
That's the more important one. There's another one in the Proceedings of the Cambridge Philosophical Society which is "The Magnetic Moment of the Neutrino" or something like that, and that's unimportant.
You mentioned about beta decay. Would you recall the circum- stances in which Fermi's theory of beta decay came to be accepted? I would also like to ask: What ruled out Yukawa's hypothesis of mesons being the mediating agency for beta decay.
Well, this is quite a deep question. Point one: Fermi's theory in principle I think was accepted immediately because it was obviously a sensible theory, a sensible explanation of the phenomenon. In detail, it was modified at one time by Konopinski and Uhlenbeck, who introduced the gradient into the theory, and this indeed is discussed in this Chapter VII of my article. At the time when this was written, this was the accepted theory, but in fact this was due merely to the experiments being bad. Nobody could do good experiments on the spectrum of the electrons in beta decay, and with bad experiments, it seemed to fit the Konopinski and Uhlenbeck theory. With good experiments, which came out I think before l940, the Fermi theory was restored to favor. Then the Fermi theory itself was somewhat supplemented by the theory of Gamow and Teller, and with the combination of these, I think we had a satisfactory theory as far as it went. It was only a short while ago in '62 or so that much real progress was made by the theory of Gell-Mann and Feynman. Now, you asked about Yukawa's theory of the beta decay. Yukawa's idea was that the meson—he had only one, which we have to identify with the pi meson—that the meson disintegrates spontaneously into an electron and a neutrino. We know it doesn't. It does so only very, very rarely. It disintegrates into a mu and a neutrino—and a different neutrino from the one which comes in beta decay. And the disintegration into electron and neutrino is exceedingly slow and is just not sufficient to explain the beta decay of nuclei. It is too small be a factor like 108 or something like that. So the meson, the pi meson, is not a sufficient mediator to explain the beta decay. Everybody nowadays believes that pi mesons are the main term in nuclear forces, and this is well established, but the pi meson doesn't give enough beta decay and therefore one needs some other source of beta decay to explain nuclear beta decay. All right. So then you might look for some other meson to do this. At various times people have suggested that there could be another meson, now called the W meson, which would do this. CERN, in particular, has searched for such W mesons, with their 30 GEV accelerator.
That would be intermediate bosons?
Intermediate bosons, yes.
And this is what Lee and Yang have proposed.
So it seems, in the Yukawa case, that the meson was to be the intermediary agency for nuclear physics, and Lee and Yang have proposed the intermediate boson, but nature doesn't seem to work that way.
It has not so far worked this way. I don't think a negative answer here is conclusive. All that has been proved is that the intermediate boson does not have a mass less than one and a half times the proton mass, but it could very well have a mass three times the proton mass and no existing experiment will contradict this. The trouble with the intermediate boson is, apart from its not having been discovered, that it is very difficult to construct a theory of the electromagnetic interactions of a charged vector particle, a charged particle of spin 1. This gives rise to all sorts of the most horrible divergences. Lee and Yang have tried to evade those, I think not very successfully; and so therefore this vector boson is not a very happy particle.
To bring us back, you got on the subject of beta decay by looking through Chapter VII and then you decided to recapitulate this since in your earlier account you had left out the story of beta decay. And I think you had come to the point where you had brought it up to the time of 1936 when you dealt with it here. But you mentioned, though, that there was another thing that you had left out. I wasn't sure whether you had covered it. Maybe that was the idea of the neutrino.
When I talked about nuclear forces, I mentioned three things. That is not what you mean.
No, I meant here when you said there were two things. The beta decay section here reminded you that there were two things you left out. One of them was the history of the beta decay. I wonder— the positron and pair creation thing is not what you meant?
This is not nuclear physics.
So that's not what you had in mind certainly.
I don't remember what we had in mind, so I guess we better erase that.
May I suggest that we continue through chapters VII and VIII and then perhaps we can take a break and see where we are?
All right. Let me say one more thing about Chapter VII. In those days there was no meson, or at least I didn't take account of the meson. And there were accordingly attempts to explain such things as the magnetic moment of proton and neutron by interaction with electrons. That was stupid, and of course the right answer is that such things are explained by interaction with mesons, and in this respect—that is, in the theory of the nucleons of the elementary particles—meson theory really is important. What we did back in 1936 was just not very sensible. In fact, I think I said here in this article that it isn't very sensible and that you can prove anything about the magnetic moment as long as you would take electrons and neutrinos as the responsible agents for the moment.
Was that one of the areas in this paper that you tried your hand at before coming to a conclusion? Had you worked at that a while?
No. There I just examined the proposals and criticized them and that's all. Now, Chapter VIII was written by Bacher. It's on nuclear moments and the evidence on nuclear moments, their values, quadrupole moments and the evidence from spectroscopy and so on. It is not directly nuclear physics, but it is a tool for the investigation of nuclear physics.
This is the work that Bacher was doing here at the time?
He came after you were here?
He came half a year after me, and I think a year after that, approximately, Livingston left to go to MIT and then Bacher took over from Livingston. He took over the cyclotron and its running. I'm not sure anymore what year this was. It may have been as late as 39.
I talked with him this summer, by the way, on a very short- range interview, and we covered some of the background of that. I think that now we've reached the end of the table of contents or the first installment of your three-part paper, and I would suggest it would be a good time for a break, and we'll find out how much longer you want to continue today.
I'd like to continue till six, if that's all right with you.
Fine. I find this is good to do this. I don't know if you've ever had the opportunity to do it systematically.
I never did, no.
I think it's very nice. This also gives the possibility of examining what were the new elements that were introduced here. [Reel 1, Side 2 of tape; continued after pause of about five minutes.]
Since we have been talking about beta decay, Professor Bethe, may I ask you something about the nature of beta and gamma decay? Was it always assumed that the nucleus acted as a collection of particles, each subject both to beta and gamma transitions? I would like to know who told us that when nucleons are bound in a nucleus, they still exhibit their characteristic interaction?
I guess that was just assumed, and perhaps assumed as the simplest possible assumption. It is of course still one of the important points to verify, that is, to verify that, and to what extent nucleons in the nucleus are the same as they are outside the nucleus. This has lots of ramifications. The most important question perhaps is whether the forces inside the nucleus are the same as you have between two colliding nucleons; and this is a question which we are now only beginning to answer. We are now beginning to come to a point where we can say that the forces, while not exactly the same perhaps, are sufficiently closely the same so that it is sensible to start from the forces of interacting nucleons. In the case of gamma decay, gamma emission, I think the case is the simplest—namely, we have a system of charged particles; and since we have charged particles, you'll have an electric and a magnetic moment associated with any transition. Now, the electric moment one would imagine can be derived from the distribution of charges just the same way as in an atom. So one would expect that one should get from the general theory of the nucleus a description of the wave function of the nucleus, giving the position of the charges, and from that one could calculate the electric moments, dipole moment and higher moments. We have every reason to believe that this is a good description, and all the experiments which have been made in great profusion since the war on electric dipole and quadrupole moments of transition, and also higher moments, seem to bear out the assumption that the distribution of the protons—that is, the wave function of the nucleus as a function of the proton coordinate— will give us properly the electric moments of various orders. For the magnetic moments already exist in the normal state of the nucleus, in the stationary state, not only in transitions; and there are also transitions made by magnetic moments. The moments in the ground state of the nucleus have been measured many times and very accurately, and here the accuracy of the measurements is such that we know that the magnetic moments are not always simply those of the free nucleons, but there is a change of the nucleon when it enters the nucleus. However, the greatest change is in the nuclei of mass 3, the triton and helium 3, where the moments of the odd nucleon are enhanced by some 10% over the value which they have for the free nucleons. This is considered to be a meson exchange effect and an effect which nobody has yet understood in any quantitative detail. When you come to heavier nuclei, it really looks rather better. The magnetic moments of heavier nuclei seem to be closer to what you might expect from the free nucleons. But half of part of this is that we just don't know the theory so well for the heavier nuclei, and so maybe there is more discrepancy than we see. But I think the conclusion is—and this is just a conclusion from shell model and experiments— that magnetic moments are preserved to within a few percent when a nucleon enters the nucleus. We can't say it any better than a few percent. Finally the beta decay. Quite some volumes have been written, and, again, Konopinski is the expert on this, on the probability of beta decay, that is, on the matrix elements of the transitions. In this case, since beta decay is a very weak interaction, the proper thing to do is essentially the same as for electromagnetic transitions; namely, one wants to have the wave function of the system before beta decay and after beta decay and get the matrix element connecting the two. In the case of beta decay, in fact it is better than in the case of electromagnetic interaction just because of the troubles which I mentioned before— namely, just because Yukawa's theory of beta decay didn't work. We know that the mesons which are transmitted between nucleons as part of the nuclear force do not give much contribution to beta decay. The meson decay in fact is weaker than the nucleon decay. So if we get beta decay, we get it from the nucleons in their natural state, and we don't get it when they are dissolved into a meson and a nucleon. Whereas, in the case of charge, this is not so. The mesons have the same charge as a proton, plus or minus, and therefore electromagnetic currents of the mesons are by no means negligible, and these electromagnetic currents are believed to explain the deviations which I mentioned in the magnetic moments of the nuclei of mass 3. But for beta decay we have no reason to expect such deviation. We have every reason to believe that you should just take the matrix elements between the nuclear wave functions before and after the decay. Of course to get these matrix elements is quite difficult and I will not describe this, but I think so far there is every confirmation by experiments that this picture is correct—namely, that one nucleon in the nucleus decays and decays into the other type nucleon with emission of an electron and a neutrino governed by the same beta decay interaction as we have for free nucleons, but one has to take into account the wave functions.
If I could take an analogy of raw cucumbers and pickled cucumbers, it seems that the nucleons behave the same whether they are pickled in the nuclear jar or outside in a dish of their own.
Yes. Very good.
Would you remember something about the discussions, perhaps in the '40s, about the earlier ideas on the universality of the beta interaction?
I don't remember very much about this. I really didn't participate in these discussions. I've heard about that occasionally, but I think you should ask other people to tell you about this. I found after the war that my time was very limited and so I restricted myself to smaller and smaller areas of physics: in the late '40s, mostly to high-energy physics, and then since about '55 to the theory of nuclear matter.
think perhaps this is the point to get back to the project ...
think I would like to take up one more point here.
The point about the history of the universal vector axial-vector interaction. For the record, is it historically correct that Marshak and Sudarshan were the first to analyze the data and show that the only consistent picture would result if the interaction were chosen as vector axial-vector interaction? There were, as you will remember, at that time several experiments, notably the apparent absence of the electron decay mode of the pion. Marshak and Sudarshan suggested that these be redone, and they were redone and found wrong. Feynman and Gell-Mann wrote about it at almost the same time, but did not make the same exhaustive analysis of the data. Sakurai also did that later. Were you involved in these discussions; and if so, I would like to have your comments.
I was not involved. I think the account which you gave is correct historically. I think Marshak and Sudarshan did it first. Not only did they do it first, but they used as a theoretical argument one that is still current and still used and is very useful—namely, the argument that the interaction should be invariant against the insertion of an extra operator gamma 5. They used this. Feynman and Gell-Mann did not use it, to my recollection, but used a different argument, which is obsolete. So on all these counts I think Marshak and Sundarshan should be given credit. I am told—but this I only know from being told by the author—by Marshak that he gave a paper about this at a conference in Padua in the fall before the publication of the Feynman-Gell-Mann paper. Gell-Mann apparently was in the audience. So were many other people. Most of them thought he was crazy, especially in his statement that the interaction should be axial-vector and vector rather than scalar and tensor which had been commonly accepted. The experimenters, in particular, objected to his suggestion that their experiments -night be wrong. Unfortunately, the only manner in which these remarks were published was in the proceedings of that conference. Unfortunately, Marshak and Sudarshan did not write it up as a paper for a normal journal. And I think this has a lot to do with the fact that it is forgotten, that they have been forgotten. Gell-Mann and Feynman wrote it up for the Physical Review, so everybody read it, and the other was just in the conference. Most of the participants in the conference didn't believe it when they heard it. Nobody ever bothers to read the proceedings of a conference afterwards, and I think this was just bad luck.
What year was this?
It was in the '50s.
It was late '50s—I think '57 and '58. I think Marshak and Sudarshan must have been fall of '57, and Feynman and Gell-Mann were early '58. I believe that's right. It could have been a year later.
Professor Bethe, to go back to the Bethe Bible, I would like to ask one question. You had worked on the group theoretic computation of this splitting of energy levels in a crystal. Why were such techniques not applied to the nucleus by you or other people—for example, Hund, Wigner or Racah?
They were applied by Wigner, and this was precisely the next thing I was going to talk about. They're not in the "Bethe Bible" largely because it was done after the first testament was written and it didn't fit into the two other volumes. But Wigner, I think in '36 or `37, did apply the group theory methods to the structure of nuclei, and in fact his papers were exceedingly important and exceedingly good. As a matter of fact, some of this work was very recently studied again in connection with elementary particles. What Wigner observed was that there are two quantum numbers of nucleons, which are the isotopic spin and the spin, which play a very fundamental role in nuclear physics. Each of these two can have two values, plus or minus, so you get four different combinations. Since you have four different combinations, you can put four nucleons into one spatial orbital; whereas, in the case of atoms only two electrons can go into one spatial orbital. Consequently, the first shell is completed at atomic weight 4, the alpha particle, and the same structure can be followed to higher atomic weight. Well, Wigner did this in a really admirable way, using very powerful group theory. You should really get Wigner himself to describe this. But he was able to predict from the symmetry arguments the special stability of nuclei which are multiples of the alpha particle without using an alpha-particle model. It just comes out of the group theory. It is the most symmetrical wave function which you can form of the nucleons. When you can put the nucleons into a wave function which is as symmetrical as possible with regard to interchange of the spatial positions of the nucleons, then you get particularly close correlation between the nucleons and particularly strong attraction by the very short-range forces. Now, on the basis of this, he was able to predict the ground-state energies of nuclei up to about atomic weight 40 or so, and he was able to predict some nuclei which had not been found. For instance, he said, "Sulphur-36 should be a very stable nucleus. It should be stable against beta decay, and it should therefore exist in nature." It had never been found. The mass spectroscopists looked again at sulphur and promptly found it. There is sulphur-36. And I think he predicted two or three other isotopes which had never been found before. Technically his paper (his papers really) are also particularly important because it was, I think, the first time in atomic physics that two groups were put together, so to speak, multiplied—namely, the group of isotopic spin and the group of ordinary spin. And he showed that the nucleus could be characterized by three quantum numbers—the total isospin, the total spin, and then a third quantum number which I think he called Y, which had something to do with the product of ordinary spin and isospin. And he could show that to each state in the nucleus one could assign these three quantum numbers. I think his papers were the first papers in which the importance of the total isotopic spin was stressed. And the total isotopic spin is different from the charge, from the difference between the number of neutrons and protons. 0f course, the basis of Wigner's paper was the observation which I discussed previously of Breit and Condon of the charge independence of nuclear forces. This made it possible to classify states just by total isotopic spin. Now then, partly still in the very late '30s, but mostly in the late '40s and in the '50s, people have found the concept of the total isospin of a nucleus exceedingly fruitful and they have been able to find states in neighboring nuclei which correspond to each other which have the same total isotopic spin. Wigner calls these "analog states. Such pairs or multiplets of states were first found in the lighter nuclei. For instance, carbon-14, nitrogen-14, oxygen-14 are three nuclei having the same total number of particles. Carbon and oxygen each have different numbers of neutrons and protons and need an isotopic spin of at least 1. Nitrogen-14 can exist either in a state of isotopic spin 0 or isotopic spin 1, the ground state is isospin O. There are definite states in nitrogen-14 which correspond exactly to the states of carbon-14 and oxygen-14, and these corresponding states have been traced in great detail, especially by the Caltech group of Tom Lauritsen, and all these predictions have been beautifully borne out. These states are very important. For instance, in beta decay they can go into each other with the greatest of ease with highly allowed transitions. And Wigner took this beta-decay transition problem into account when he wrote his first papers. From the decay of oxygen-14 into the analogous state of nitrogen-14 one derives the best number for one of the beta decay couplings, namely the axial-vector coupling. Now, Wigner in recent years, together with experimenters at Princeton, is following this up to much higher atomic numbers, and he is following it into the continuous spectrum, and this is really what he calls the analog states; namely, you can take a quite heavy nucleus— let's call it barium—and you can change one of its protons into a neutron by a nuclear reaction. Thereby you get an isotope of cesium. And preferentially this transformation goes through an analog state of cesium — namely, to a state which has the same isospin as the barium state from which you started. This state of cesium lies in the continuous spectrum of cesium. Therefore, you get not a bound state but you get a tremendous enhancement of transition probability to an unbound state of the final nucleus. While the example which I gave I think has nothing to do with reality, it has been possible to show that the idea of isospin carries up to very high atomic number in spite of the force. Wigner thought originally that it should stop at about atomic weight 60, but now they are working with weights of 120 and more and it still works. And this classification of states according to the symmetry, according to isospin, ordinary spin and the Y quantum number was extremely powerful and was one of the things which I mentioned previously when we talked about the shell model. I said that in the late '30s and until l949, the shell model was not very much emphasized because it was shown by Wigner that the symmetry considerations based on group theory were really much more important to determine the energy of medium Wight nuclei than the shell structure. So that was the application of group theory, which was eminently successful.
That's quite a clear, coherent explanation. I don't want to press you, but would you now like to continue ...?
I would like to go into Part B.
I would like to ask a question as it relates to that. I notice that there's a year gap between publication of Part A and Part B. The question is this: Did you submit all three pieces of the manuscript simultaneously?
No. I was writing it as it was being published. Part A was published as soon as it was written. Part B and C were written more or less simultaneously because it was a little difficult to separate them, and that's why there is a big gap between A and the other two. And I think B was submitted a little earlier than C, but the writing was done at approximately the same time.
I see. And B was the only one in which you don't share authorship with the others.
Are there any general comments you'd like to make about that before we go on ...?
A general comment is that on Part C, which is the experimental part, Livingston did a good deal of the work. He was an experimenter. He had studied in detail some of the experiments. And many of the experimental chapters were written by him rather than by me. I'll come to that when I come to Part C. Well, specifically, the Chapter XV on "Experimental Methods;" Chapter XVII on "Results of Disintegration Experiments," were Livingston's work. Chapters XVI and XVIII were mostly mine. And that's all there is. So the two longer chapters were his work.
And that also accounts for the reversal of the name order in this one?
Because of his larger role in
Not larger than yours necessarily, but relatively larger role.
No, I think it was larger than mine. I helped a little in the interpretation in Chapter XVII. I didn't help at all in XV. He helped some in XVI and in the Auxiliary Data, and some in Section 108, "Masses from Disintegration Data," 109, "The Excited Energy Levels of Nuclei. So I think it was more his work than mine.
Now, getting back to B, "Nuclear Dynamics, Theoretical."
Here I think the main thing I should say is now the second point which I mentioned in connection with the shell model—namely, the model of the compound nucleus. And now I can resume the story of Fermi. In 1934 Fermi experimented with slow neutrons. Artificial radioactivity had been discovered by Curie and Joliot in '33. This was an excellent tool for investigating the result of a nuclear disintegration. It was used by many people but particularly by the group at Rome, who used neutrons for bombarding nuclei. I think this story is a different story which I shouldn't talk about but which should be told by one of the participants. I think Segrè has written quite a lot about it and knows a great deal and I think is very good in talking about it. Amaldi is another. Rasetti at Johns Hopkins is another of the participants in this early work, which must have been terribly exciting. Well, in this work they discovered at a fairly early time that the effect of neutrons was enhanced when they were transmitted by paraffin, and I think this might never have been discovered if Italy were not rich in marble. Namely, a marble table gave different results from a wooden table. If it had been done here, it would all have been done on a wooden table and people would never have found out.
Sounds like geographic determinism.
Anyway they found out that slow neutrons were much more effective than fast neutrons and they found out very quickly after that that among the slow neutrons there were specific types of neutrons which did specific things. There were some neutrons which activated silver, and others which activated cadmium and others which activated indium and so on and so forth. Well, on the basis of this everybody got the idea. I believe it was not published by the Rome group, but the idea occurred in many places. I think in Copenhagen Bohr got this idea, and here Wigner and Breit, and I got the idea while I was moving from England to this country, that these were resonance levels of the nucleus. And I tried in the first paper I published in this country to give an account of this in terms of the interaction of one neutron with the nucleus. And, sure enough, I found resonances, but I was absolutely wrong in explaining Fermi's resonances in this way. The resonances which I then postulated theoretically, later on came back, in the 1950s, in the theory of Weisskopf, Porter and Feshbach of the one particle resonances of a nucleus. And this was brought back at that time to explain certain experiments by Barschall on the scattering of neutrons in the hundred-kilovolt region, to explain chose experiments in which they found very striking maxima and minima, both as a function of energy and as a function of atomic number. So the theory which I produced then in '35 was entirely premature. There was no experiment which would fit the theory, but these experiments only came I believe around '50, '52 or thereabouts.
The paper that you refer to as the first one in this country was "The Theory of Disintegration of Nuclei by Neutrons." This is a Physical Review paper in 1935.
Prior to that time all of your publications were in European journals.
At a later time, I will ask you why you waited until you came to this country to publish in the Physical Review.
Well, the answer is short and simple. I think I published wherever I was, and it happened that my early publications, were all in German journals. At that time I think the center of gravity of physics was indeed in Germany. Then in 1933 or so with the Rutherford school publishing the nuclear physics papers, and nuclear physics coming into prominence in the whole world, the center of gravity shifted to England. And in 1935 I had the good fortune to come to America just about the time when the center of nuclear physics also shifted to America. The small accelerators which they had in Cambridge were very good for the exploration of the lightest nuclei, but when it came to heavier nuclei you needed the bigger accelerators, the first of which was Lawrence's cyclotron. And I think 1935 was just about the time when significant experiments were begun on the Lawrence cyclotron. So this was just a piece of luck. Now as I said, my theory was inapplicable to the resonances which were found in those days which were resonances at a few electron volts. And the correct theory was given by Bohr: it was the theory of the compound nucleus—namely, the idea that when you add the neutron to the nucleus you form a new nucleus, the compound nucleus, which is highly excited because you have added the 8 MeV binding energy. You are 8 MeV or so above the ground state. And because it is highly excited, you have many levels at very close spacing. Bohr I think formulated this theory most clearly and before everybody else. Breit and Wigner very soon afterwards developed the dispersion formula for cross sections with closely–spaced nuclei. I think they didn't have the full Bohr picture in mind, but a somewhat simpler, if you want, picture where they talked about two or three nucleons rather than all the nucleons acting together to make the energy levels. But they were the first to give that resonance formula. So, accordingly, in Part B, the first few sections are devoted to Bohr's theory of the compound nucleus and to the dispersion formula.
May I interrupt at this point to ask you to clarify something, not that you said but that somebody else said? It seems that Frisch— and I'm not sure where this quotation comes from; I think it's probably in connection with the quantum physics interviews and conversation with Tom Kuhn—talks of 1936 when Bohr gave a lecture to the Copenhagen Academy, and that you were giving a talk and Bohr was thinking during the talk of some of these ideas that led him to the compound nucleus. Here is the quotation that I have: "Frisch feels Bohr was during Bethe's talk realizing that it was absurd to think of the nucleus as bound together by interparticle forces." And not long after, the compound nucleus idea was fully worked out. Do you remember that particular occasion and do you remember if Bohr did ask questions ...?
I remember it vaguely. I remember that I gave a talk about the structure of nuclei and this may have been something like the shell model and things like that. I don't remember what year it was. In a way I would be surprised if it had been '35 or '36.
You were here then.
I was here. I visited Europe during those years. However, I didn't go to Copenhagen for very personal reasons. Namely, I had been engaged to a young lady at the Copenhagen Institute in 1934 and I broke the engagement, and so I didn't think I should show myself for several years. In fact, I didn't until I think '51. I saw Bohr in between. So I'm pretty sure I was not in Copenhagen later than the summer of '34. It is very likely that in '34 I gave a talk. It is very likely that it was about nuclear physics. It is very likely also that it was very much based on the particle idea, on specific wave functions, on individual quantum states for each nucleon.
Independent particle model.
Independent particle model. Whether it was shell model at that time, I don't know. I think I got interested in the shell model only after that, but certainly independent particle model. And I remember vaguely that Bohr asked a number of questions, and the way that he operated was that his questions were very tentative and seemingly very vague. I think it is quite likely that Frisch's story is correct and that indeed he thought of that idea in opposition to my independent particle model talk.
His paper on it came out in '36, and I think the reference was that Bohr himself gave a lecture to the Copenhagen Academy in '36 and perhaps that was the paper that was published in Nature. This followed the seminar paper that you gave. If your seminar paper was 34, it still would fit in.
It must have been August of '34 approximately—maybe September.
And of course, as you said before I interrupted you, the Bohr compound nucleus idea was the first idea in Part B.
Yes. Now, as I remember it, I heard about the Bohr compound nucleus earlier than '36, and it may well be that people in Copenhagen talked about it already in '35, and then I probably heard about it from Placzek possibly in the summer of '35 or some such time.
In this country?
Possibly in Europe. I visited Europe every summer and it's very likely that I saw Placzek during this time, or it is possible that Placzek came to this country on a visit. I'm pretty sure I got it from Placzek. I'm pretty sure I got it as early as 35, and it was then still in development in Bohr's mind.
Did your presentation of it amount to an advancement of the idea rather than just a public restatement?
You mean back in 34?
No, I meant in '37.
You had thought about it and reworked it somewhat?
I had, yes.
In what way?
Particularly in the sections after the first two—well, 51 is just very general, representing Bohr's idea; Section 52 is Wigner and Breit. But now in 53 and following—and this is essentially original work—I tried to be more quantitative; and, in particular, I derived the distribution of nuclear energy levels—how many energy levels per million electron volts or what-have-you. And what I did at that time turned out to be really quite good, with one exception. Unfortunately, I thought that the nucleus was very big, almost twice the radius which later on turned out to be true, so the numbers in here are all wrong, but the ideas are quite good and the formulae are quite right. So I calculated the distribution of energy levels and got the right order of magnitude of the number of levels per hundred volts or so. I calculated the width of the nuclear levels. Bohr had made a general model, the evaporation model, and I now tried to get the width of the energy level from this general theory. And, again, I think this is mostly original work, al- though some of the work was done by Weisskopf at the time.
Excuse me, I don't understand. Done independently by Weisskopf or done in connection with this?
We talked together but I don't remember.
Weisskopf would come and visit with you. He was close by, wasn't he?
Yes. I think it was done independently. I think he did it a little better than I and then I incorporated it in here. I think I had done some of it. Section 55, "Derivation of the Dispersion Formula." A little simpler than Breit and Wigner had done it. Then 56 is essentially original work, partly done with Placzek, where I tried to get numbers useful to the experimenters from the theory and where also I tried to go from the low-energy theory, which had been experimentally established, to a higher energy theory, where the resonance levels overlap and where you want to get cross sections averaged over large energy intervals. So that is what Section 56 is devoted to, and part of this was also published separately with Placzek. The next chapter [X] is about neutrons, which were, after all, the cause of it all and were the instrument by which all these regularities had been discovered. It has a lot of discussion of the Fermi papers. Some of the sections are about interpretation of experiments, especially Section 59, "The Diffusion of Neutrons." There I relied quite a lot on Fermi's previous work. It is essentially a reproduction of Fermi's work in somewhat different language. Section 60, "The Neutron Resonance Energies," getting these energies from observation. I think much of this was developed here by Placzek and myself, especially the "boron absorption method." It had been used previously, but I think we did it a little better. The boron absorption method is a way to determine the energy of the neutrons in a resonance— namely, you observe that, let's say, silver is activated by neutrons, and then you put in a boron absorber between the source and the silver. Boron fortunately has a cross-section which goes exactly as one over the velocity and the constant is known in this formula, and therefore you can determine the velocity of the neutrons which are absorbed by the amount of absorption in boron. This was our method in those days to, determine neutron velocities and that's how we knew the resonance levels. It had been done by a number of other people, and many of the experiments were from places other than Cornell. It was just put together here. So the levels were obtained in this manner, and this is the first compilation of all the resonance levels observed. Then we discussed the width of the neutron levels. It's difficult to tell just what was new and what not. It was again one of those examples where everybody talks about the subject and similar work had been done at Columbia in particular and in England by Moon and several other people and by the Italian group. But I think this was again the first time in which it was put together in a coherent fashion.
Is that P. B. Moon who had himself a stormy history?
Did he have a stormy history?
I mean politically perhaps.
Politically? I didn't think so. It must be a different Moon. This Moon is in Birmingham and I think is quite conservative and had no storms.
Real conservatives are allowed to have storms, too.
But not in England.
So this became a compendium of what to do when one did neutron experiments. Then in Section 62, "Neutron Width [and the Absolute Cross Section as Derived from the Experiments," this is the quantity about which I talked in Section 54, so this was now a connection with the theory and theoretical conclusions were drawn from it. Then 63, "Scattering of Slow Neutrons," again has to do with the question of getting the neutron width, and also in this I point out that there is not only a contribution from the compound nucleus, but also a contribution, so to speak, from the rough shape of the nucleus. That is, the nucleus doesn't let the neutrons in and therefore it has a cross section equal to its geometric cross section, as I then thought. Later on a paper by Weisskopf, Feshbach and Porter, which I mentioned previously, corrected this. It's not as simple as this, but there you now have to take into account those one-particle resonances which I had talked about in early '35. I didn't know that in 37. Then 64, "Disintegration by Slow Neutrons with Emission of Charged Particles" and some processes on fast neutrons—more or less just putting things together so that experimenters could work with it. The next chapter [XL], "Alpha Radioactivity" is just a repetition of what people knew. However, I think I somewhat simplified the theory compared to Gamow and company. (Just today I gave a lecture on this to my senior class, and I found that in the textbook I'm using my theory is used because it is simpler.) But this chapter otherwise didn't contain much new. Now Chapter XII, "The Scattering of Charged Particles by Nuclei," and XI)I, "The Disintegrations Produced by Charged Particles," contained perhaps the most work, particularly the most work of my collaborators, Rose and Konopinski. We tried to put everything in good order. We calculated potential barrier penetrations to discuss the nuclear reaction. We combined the idea of potential barrier with the idea of the compound nucleus, and I think this work in these two chapters is mostly original except where I quote other people. It isn't anywhere a profound theory. It's just putting together these two things which most people knew, but I think it was useful to the experimenters because it gave them very simple easy-to-use formulae for the interpretation of their experiments. I don't think I need to go into detail. Perhaps one point that is interesting is Section 83, "The Selection Rules," saying that not every reaction will take place which is energetically allowed, but many more selection rules have been found subsequently, partly in connection with the point I discussed previously about Wigner's theory of the isospin and other quantum numbers and there are a few more. We tried as much as possible—and this is in Section 84—to make it quantitative, to give absolute cross-sections, not just relative cross-sections. Then the last chapter of this part, Chapter XIV, is about "Gamma Rays." This is related to the subject which you brought up earlier— namely, can you explain gamma ray probabilities on the assumption that nucleons are nucleons in the nucleus and what's the transition probability? So transitions were classified, multiple transitions. An important point which had been discovered, and I think suggested, I don't remember by whom, was the idea of the metastable level. I'll look it up here and find out who suggested it because it had been observed that certain nuclear levels (Weizsäcker was the person) take hours or days or even years to decay by gamma ray emission; whereas on simple theoretical considerations you should expect that it takes about 10 seconds, maybe 10-¹², for lower energy gamma rays, but certainly a very short time. And Weizsäcker suggested that this probably was due to high angular momenta of the excited state. Then you might have a transition from a nucleus of angular momentum 5, say, to the ground state of angular momentum 0, and therefore it cannot be done by dipole or quadrupole but it requires a 32 pole to do this, and these transitions are very unlikely by general laws of quantum mechanics. So this idea of Weizsäcker explained these metastable states, and later on that same idea was very important in the establishment of the shell model. Namely, these meta- stable states of high angular momentum occur mainly in certain regions of the periodic system. And these are just the regions where nucleons of very high angular momentum are being added to the nucleus, and so the shell model got immediate confirmation by pointing out that just where very high angular momenta should be added, according to the shell model, these very high angular momenta do occur in the metastable states. 0f course in 1937 I didn't know about the good shell model, but I did know about metastable states, and these are discussed here. Other things discussed are that transition probabilities are generally much lower than you would expect for dipole transitions. This was explained only in the '50s by the so-called giant dipole resonance. At this time it was just recorded as a posit. Quadrupole transitions are, however, as strong as the theory would predict. So that's one of the sections in this chapter. Mostly these actual estimates of quantitative transitional probabilities I think were original. And that's just about the end of Chapter XIV. We discussed the processes which can follow gamma ray absorption. That's just putting gamma-ray absorption together with the compound-nucleus model. Unfortunately, there is not terribly much to be said about Part C.
Part C, as you explained, was the real experimental review, and this was made possible by the large card catalogue that Livingston had and he made use of it in this one. How did you make use of it, though, in the other sections—in A and B?
For examples. I got my examples from .
Was this a bibliography or were these notes on the articles?
Notes on the articles. For each article he had a little card with a short paragraph of notes.
Did he explain to you how come he happened to have them?
Well, he was working in nuclear physics. He thought there was awfully much literature. He couldn't keep track of it by keeping it in his mind. There were hundreds of papers already in that card catalogue.
I saw him for a few minutes, not for a formal interview, and I think he mentioned a number in the thousands—close to 3000 or something like that. He said they are all destroyed now.
And he said the reason he did it, was because he felt that at Berkeley he didn't have a full picture and one of the things that was a compelling thing to him when he came to Cornell was to get this full picture just as soon as possible. And so he did this and had it on hand when you arrived on the scene.
Yes. Well, the number here used is something like 600.
Already a phenomenal number to absorb.
Yes. We obviously didn't quote them all.
We have a little remaining time, so ...
May I just quickly ask one or two questions that still pertain to this? At what stage did people recognize the parity quantum number? When did people get convinced about it being conserved, and I would like to know whether it was just a matter of educational time?
Parity certainly was known from atomic physics, and it's back in the old atomic physics papers—papers like Wigner and von Neumann and Jordan. Around '28 when people emphasized the parity number, it was emphasized in atomic physics because it explained La Porte's Rule. La Porte's Rule is that the energy levels of a given atom can be classified into two groups which we may call even and odd, and strong spectral transitions can only occur from even to odd or from odd to even but never between two even states or between two odd states. And La Porte's Rule, which had been derived empirically from the spectra before quantum mechanics, was explained in quantum mechanics by parity—namely, that it was the wave function which was odd or even with respect to taking the mirror image at the origin. So naturally when nuclear physics came, we all knew about parity. And when the quadrupole moment of the deuteron was discovered, in particular, one of the main points in setting up the tensor forces was that parity should be conserved, that is to say, that the tensor force ought to be invariant with respect to the parity trans- formation. Everybody assumed this as a matter of course, and it wasn't questioned. I think Wigner was one of the first to write down this tensor interaction, and nobody questioned that you had to take a Hamiltonian of even parity.
Before you get into another technical question, maybe we can defer the technical questions until tomorrow because it's getting close to the hour that you said you wanted to leave. I would like to ask some easy questions, some sociological questions, that do fit in while you're in the mood or even while you're tired of talking of this three-part paper. One is a very simple one, I think, about the time that you put into it, about how long it took you really working on it.
It took me the better part of two years. I started working on it probably in the summer of '35—I can't date it very accurately; it may have been late spring. It was finished in something like April or May of '37. During this time I worked mostly on this. As I worked on it, I got some fall-out. Certain things became clear to me which seemed suitable for a separate paper to be published in the Physical Review, and so I published a lot of papers in the Physical Review during that time.
That's what I was getting at because there seems to have been a great deal of achievement.
All connected to the same subject. Whenever I found something while writing this ...
So when you were getting independent ideas to complete the picture there, you were publishing those ideas elsewhere at the same time.
Yes, at the same time.
The bibliography reflected that, and it shows that you were collaborating, as you indicated, with Rose quite a bit during this period.
And with Konopinski.
Yes, but most of the papers where another author is involved involved Rose rather than Konopinski. As a matter of fact, he's not involved in any of them.
Oh, he isn't?
Not in the 35-'37 period.
When is the first paper with Rose?
Rose comes in on a paper entitled, "The Maximum Energy Obtainable from the Cyclotron," which is another paper that we'll talk about some time. That was 1937, and then the others follow—"Nuclear Spins," and so forth. Rose might have been earlier, but I don't see him.
Nothing earlier? I am a little bit surprised, but maybe isn't entirely complete.
My point in asking this is that you were doing this and you apparently had teaching responsibilities as well. About how heavy a load did you have during this period?
About one course each term. I think occasionally I had two. 0ccasionally I had a lecture course and a reading course. The reading course took very little time—just telling people what to read and then giving them a short oral exam to find out whether they had read it.
Was the lecture course a graduate course?
Always a graduate course with one exception. I taught one section of sophomore engineers who had flunked the course before. This was the most miserable teaching I ever did. It was a most depressing class. Most of them tried to flunk it again.
Then you were spending your summers in Europe.
But you didn't do any writing in Europe.
I did a little writing in Europe. In fact, usually during my summers I would spend two weeks maybe visiting other physicists, mostly in England. I would spend a month at my mother's in Germany and at my father's, and during the time with my mother I worked and wrote.
That was in Frankfurt.
My mother was in Baden-Baden; my father was 1n Frankfurt.
They were divorced?
How was it going back to Germany at such a time?
Let's do that tomorrow. So I did work during the summers. Then we had usually several weeks in Switzerland on vacation. During that time again I did not work. So this was the set-up. I am surprised that I did not have papers with Konopinski. It may well be that this is incomplete. I have the impression that I worked about as much with each of them.
Since we still have a couple of minutes, could I still ask about nuclear physics a little?
Let's give Professor Bethe his choice. The question that I wanted to ask ...
I don't want to interfere with that.
The question is this: the evaluation of this entire series in two respects—that is, your evaluation in terms of the impact this had on the field, including the immediate and long-range response, and the second is what it meant to you, how you felt about it—that is, in terms of your total work, in terms of the time it represented and what it meant in your own development and in your subsequent work.
Well, for myself it was very much worthwhile. I really went through this subject, and I believe I understood it all at that time. I don't any longer. I connected all the parts of it, and I had the feeling that by this work I really had made things clear to myself so that after this I could work in this field without needing to read anything because I had written it all down. And so I had the feeling that it helped me tremendously, and that it stimulated me to work during this period and after this period, and really made it easier to work because I had a basis from which I could then find the interesting problems and go after the interesting problems without much effort. The impact on other physicists, I think, was quite great even if I say it myself. I think it was considered as the standard work. It was for many years I think the most comprehensive treatment of the subject, and I think remained so until Blatt and Weisskopf's book came out. And so in the intervening 13 years this is where people went for information, and certain parts of it were not reproduced in Blatt and Weisskopf and were used even after '52—not many but a few. It came of course at a time when nuclear physics was just bursting out everywhere in this country and in England, and to a lesser extent in other European countries, and so it was, I think, very useful for the practicing nuclear physicists, both theoretical and experimental, to have one book to which they could go to get the first information and one book which would tell them what was already known so that they could then work on from there.
This suggests—and the remarks of other physicists back it up— that the very existence of these articles served to stimulate the field, that it wasn't merely a reflection of interest in the field, but the very existence and pulling together of all that was known into this type of presentation helped to attract people to the field and to stimulate the growth of the field.
That is certainly true.
Was the type of contribution you had made in this generally recognized? You told us here how you had rewritten certain things, how you had reformulated and how much new material was in it, but was this recognized as a review article or as original work?
I think the experts knew what was already learned and what was new, and also we had profuse quotations, so that I think it is clear in every section. In some sections there are dozens of quotations and so this is clearly review. 0ther sections have no quotations or one or two. So it is clear that those are original work.
A final question related to our lunchtime conversation about the function of review articles in absorbing past literature. How useful do you think the bibliography that you published with the article actually was? In other words, do you think people went to that bibliography and used it or were apt to just rely on the article alone?
I think sometimes one and sometimes the other. I think in many cases they just relied on the article alone. I would imagine that most people who consulted it for actual research certainly in the first ten years of its existence probably went and read the appropriate section here and then looked up two or three of the references. So I think the references were useful. I think the article without the references would also have been useful. But I think it's good to be able to dig into the original references when necessary.
That was your reason then for including them.
Well, there is also the reason that you want to give people credit. I think these two reasons were about equally important.
In a sense we're not talking about references but about bibliography. The references occurred in the text and you did give people credit. This was sort of an appendix. It's six o'clock. Just to show that we're honorable men, I going to turn off the tape recorder now to be resumed tomorrow.
This lovely Indian summer day is October 28th and it's in the morning, and we're resuming our discussion with Professor Bethe. I think the point that we left off on yesterday was the period in 1936 and '37 with the three-part article in Reviews of Modern Physics summarizing and consolidating and adding to the general knowledge of nuclear structure until that time. I think we completed that, but there were some general questions of a more technical nature that Dr. Mehra had in mind to follow up on this morning.
Professor Bethe, going back to the history of the development of nuclear physics, I would like to ask you about the role that models have played in physical theory in general and in nuclear physics in particular. I also have in mind that at the crucial moment it seems that Professor Bohr came to rescue a situation with a certain model. As you know, in quantum theory he brought the principles of complementarity and correspondence, and they became the guiding principles. In nuclear physics he introduced the idea of the liquid drop model. He had worked with the surface tension of liquids early in his career, and this seemed to provide an understanding of the properties of the nucleus. Then there was the compound nucleus, the shell model and the optical model. I would like to ask you to comment on models in nuclear physics—your own association with them, controversies with respect to them, and their final establishment.
Well, you mentioned Bohr's models of the liquid drop and the compound nucleus. In my opinion, the liquid drop model has been quite a lot overdone. Many things were blamed on the liquid drop model which shouldn't have been and which really depended on much less than that model. For instance, the question of surface tension: Well, I think it was perfectly clear, and in fact established by Weizsäcker before the liquid drop model was proposed as such, that nuclei had a surface energy. The Weizsäcker empirical formula which I mentioned yesterday contains such a term. I suppose it was conceived by Weizsäcker much in the same spirit as a liquid drop picture—namely, that there is a surface layer which has less interaction. But it didn't depend explicitly on the assumption of the liquid drop. Well, to many people it was a help to have this picture of the liquid drop, but I believe that the most important results which flowed from the liquid drop model were obtained previously in the Weizsäcker semi-empirical formula. Precisely the same kind of formula was used later on by Bohr and Wheeler in discussion of the fission process. They called it explicitly the liquid drop model, but it was no different really from the Weizsäcker formula. It's the same physics only a different name. By contrast, the compound nucleus model I think has had a tremendous influence, and for many years—I think from '36 to about '52 or '53— physics stood completely under the influence of the compound-nucleus model. Every nuclear reaction was explained as a formation of the compound nucleus and thereafter disintegration. Every reaction was explained in terms of a statistical distribution between final states. The compound nucleus model led to still another model, if you want to call it that— namely, the evaporation model; that the products of a nuclear disintegration are distributed statistically; that any final state is equally likely. It was only with the paper of Weisskopf, Porter and Feshbach, emphasizing the single-particle states in nuclear reactions, that this situation changed. In fact, while the compound model held sway in the late 1930s, there were some people—especially Wigner—who complained about it and said this was really a very narrow point of view and one should pay more attention to those features of nuclei which could be described in terms of single particle orbitals—that is, the shell model—and particularly in terms of symmetry of the wave function. talked about this symmetry theory yesterday. The symmetry theory I would say doesn't use a particular model, and I think nowadays everybody would agree that the reliance on the compound- nucleus model was overdone. On the other hand, everybody still believes that this model is extremely powerful and that a large fraction of the reactions with heavy nuclei is indeed described by this model. So this model was tremendously useful and still is tremendously useful—useful, for instance, in predicting the relative probability of elastic versus inelastic scattering of neutrons and the probability of emission of neutrons after you send in some different charged particle and so on. After the single particle paper by Weisskopf and others was published, some people again went overboard in the opposite direction and tried to explain the majority of nuclear reactions in terms of direct interaction. This, again, was called "the direct interaction model," in which it was assumed that the incident particle interacted with only one of the particles in the target nucleus and there was something like a two-particle collision. This direct interaction picture surely is applicable, again, to a fraction of the nuclear reactions. It has been a useful model, and every one of these models has been useful as long as it was not overdone.
Would you say that the compound nucleus model was exploited very considerably in most of its major applications by 1948?
think that's fair to say.
That since its inception, and until the shell model took over the compound nucleus was the major guiding principle?
Until about '48, yes.
Was there a resistance in giving up a particular model on the part of people who had been the proponents?
Yes. I don't know whether we should go right to the shell model now. It is historically not quite in place yet, but I think that I still should answer your question in a general way. There is always reluctance to give up a model if it is well established by experiments, and this certainly was the case with the compound nucleus model. It was well established and it is still well established, I think, for the majority of nuclear reactions involving energies, let us say, from 1 MeV to 50 MeV. And, indeed, it shouldn't be given up because it describes most of these phenomena very well. There was reluctance to accept the shell model—I will discuss that when I come to the shell model—because it didn't seem to fit into the picture of nuclear forces which then existed, but this was not an opposition to the shell model as such, but a disbelief of sufficient foundation.
I think one of the papers you were wanting to recall yesterday was the paper of Konopinski and Bethe of 1938. This was on the many- body model. Was it clear to you at this time that you wanted to discuss nuclear matter in terms of the many-body problem?
No, this was not clear to us. The many-body model was only used as an adjunct, so to speak, of the compound-nucleus model—namely, it was argued that when you have very many particles interacting with each other, there are so many possibilities of sharing the energy (that's the way Bohr put it) that only the state of the entire system made any sense and that every nucleon would give up its identity as soon as it entered the nucleus. So the many-body concept was essentially used to make plausible the compound-nucleus model. It was not something separate. Now, I think the paper that we wrote at the time was concerned with the question of the effect of the use of this kind of many-body model on the radius of the nucleus and on the interpretation of alpha decay, in particular. We had the idea at that time that the formation of an alpha particle is a difficult event, an event which takes place only very rarely in a heavy nucleus, and that this difficulty of formation should be taken into account in interpreting the alpha decay. If you do take this into account, then you are led to believe that the radius of a nucleus is much bigger than was then generally believed. I mentioned that yesterday, and this was the only thing that we had in mind at that time. We did not have in mind any theory even remotely similar to the modern many-body theory of nuclear matter.
During this time the Thomas-Fermi statistical model had also been introduced, and I would like to ask you about its use in the development of the theory of nuclear structure.
This model is, you may say, closely related to the liquid-drop model, and it was used in conjunction with the liquid-drop idea as a means to obtain quantitative results. I mentioned yesterday that one of the new points in the second review article which I wrote was a calculation of the density of energy levels, and this was done with the help of the Thomas-Fermi statistical model. It gave the number of energy levels per million volts or whatever, and it did this by assuming that each nucleon moved independently and that the nucleons formed simply a gas which was moving in a confined space given by the size of the nucleus—the size of the liquid-drop, at it were. It gave quite good results and gave even better results when it was modified for exchange forces, which was done by Bardeen, and it is still a very useful way to calculate the gross features, the statistical features, of a nucleus. Now, the statistical features come into play when you have very many energy levels, so when you have excitation of several MeV and levels get very dense, then a statistical treatment is indicated and the Thomas- Fermi treatment in some form or other is a very appropriate way to do it. It can be combined with all sorts of fundamental ideas about the nucleus—whether it be liquid drop or shell model or what-not. You can then go to a statistical treatment from there.
While we are on this model I would like to ask you whether the statistical model was representative of Fermi's thinking? He was involved with the statistics of gases and by and large he also had a gift for finding a unifying principle in terms of a model which could be very pragmatic and utilitarian, and in this sense is it representative of his thinking?
I don't believe so. At least I don't recall. I don't really know what model he favored. Strangely enough, Fermi never worked very much on the theory of nuclear structure. I don't know why. He did a lot of experiments. He did a lot of theory which was directly related to the interpretation of the experiments, like the slowing down of neutrons and like the effect of molecular bond in determining the cross-section of slow neutrons interacting with hydrogen molecules. But to my recollection he didn't do very much on the main line of nuclear structure theory or nuclear reaction theory. He did more experiments I think than theory. I'm afraid I haven't read Fermi's book about nuclear physics, which he didn't write himself but which was written by his students reporting his lectures. I don't know which model he used there. Do you?
Well, he does talk about the Thomas-Fermi model. He talks about the other models, too.
This would indicate that he didn't really have much of a preference.
Can I ask a question on this point? Was it possible for a person during this period to write an exposition on nuclear structure without adhering to a particular model? In other words, could you still make sense of the field without using a single model to order all of your information? Is that a true representation of the state of it at that time?
You certainly had to use models, and you could do this in different ways. If you were very impartial, you might treat all models; and then if you were clever, as Fermi certainly was, you would put each model in perspective and say, "This works on such and such problems and the other model works on such problems for these reasons." If you were less clever, you would just enumerate them and say, "That's all the models that people have used and you take your pick." Now, Fermi certainly wouldn't have done the latter, but rather the former. I think I may use the Blatt and Weisskopf book as an example. They describe a large number of models. They describe in great detail the compound nucleus model. They describe the detailed foundations of that model in terms of an elaborate dispersion theory by Wigner and his collaborators; and on the other hand, they describe also with great love the symmetry theory of Wigner, which is a very different approach, and, if you wish, a very different model. Unfortunately, Blatt and Weisskopf don't do very much about the shell model yet, but if this had come of age by the time they wrote their book, they surely would have given it the required place. There is a more recent book than Blatt and Weisskopf which is also quite comprehensive. That is the book by Preston, who discusses all the models in perspective and shows where the various models should be used.
All these models have to stand or fall by some essential properties of the nucleus that they had to describe. That is, they had to describe, for example, saturation properties, short range of nuclear forces, charge independence and things like that. They had to be included.
These features have to be included, but I would not call these features part of the model. I would say these are the things about the properties of nuclear forces, and the models which you have previously mentioned are models of the structure which results from these forces— in other words, models of the properties of the wave function under the influence of the accepted forces.
I think the reason we're pursuing this is because it suggests a good area for follow-up work, to find out who was using what model at a particular time and why, and then to try to focus on the process of change between one model and another. This usually yields very interesting results. And so this background would give us a start. One thing we could do is go to published works, other than the major review papers we've talked about or the books, but go and see whether implicitly they are thinking about a model, whether they enunciated it or not.
As I mentioned, Wigner didn't like the compound nucleus at all, and I think never worked on it.
This shows in his publications?
Does he explicitly attack it and say why he didn't like it?
It was from this point of view that I asked the question: What were the major controversies about the questions of these models and nuclear structure at this time? Were there any hot debates at the major centers, perhaps in Copenhagen?
Princeton, yes. In Princeton there were Wigner and Wheeler, Wheeler a complete devotee of the compound-nucleus model, and very much influenced by Bohr. Bohr was frequently in Princeton and worked with Wheeler, and this happened particularly in 1939 when fission had been discovered, and they were using entirely the compound model; whereas Wigner was interested in the symmetry properties. Now, I think it is very easily stated. Wigner was interested in the low-energy states of a nucleus, whereas Wheeler and Bohr were primarily interested in high-energy states and nuclear reactions, and so each was using the proper model for his purposes.
Getting back to the other question, did this ever get to the point of open controversy or a debate?
I don't believe so.
Shortly before the war many other things were coming into focus. In your own work you did some very important calculations of the binding energy of the deuteron. You addressed yourself to the question of large perturbations as a method of handling them. Fission had been discovered, and you personally were concerned with the meson theory of nuclear forces, which was not in connection with high-energy physics at that time. Taking these one by one, I would like to ask you the state of affairs representing nuclear physics, nuclear structure, in the two years before the war— that is '38 and '39.
Well, I think this picture will be rather personal—that is, what I thought at the time. I thought we had a good understanding of nuclear reactions to the extent that we could predict with pretty good certainty the cross-section of any nuclear reaction that you could name and that was not yet investigated. We used this knowledge in a general way in the Los Alamos project where quite a number of nuclear reactions were investigated; and, on the whole, I think our predictions were correct. Weisskopf was very much in this; and, in fact, at Los Alamos he was the person who made most of the predictions. The experimental nuclear physicists used to come to him, to his office, in long lines asking his advice what to do, how to measure, and what would be his estimate of the result. He was known as the "Los Alamos oracle."
And also the mayor.
That also, yes. I think he and I worked very closely together in the time before the war and had much the same outlook. Many experiments were done on the Rochester cyclotron, which was somewhat higher energy than ours, which generally served to establish and confirm our ideas about the theory of nuclear reactions. All this was using the theory of the compound nucleus. I believe that in 1938-'39 we thought that we had some idea about nuclear forces from the study of proton-proton scattering, which I discussed yesterday, as analyzed by Breit from the saturation properties which we knew from the heavier nuclei and a few other pieces of evidence. We were certainly aware that we had very little idea how the nuclear forces came about, and this brings me to another point you mentioned—namely, the meson theory. I think we all were attracted by the meson theory as soon as mesons were discovered and wanted to take it very seriously and use the meson theory for the explanation of nuclear forces. That was really the attempt I made in the paper which you mentioned.
The 1940 paper.
1940. I was not the only one. There was of course Yukawa himself. There was Nordheim, who was very active in this field, and later on during the war there were Pauli and Wentzel who did a great deal on this theory. I think I was terribly optimistic, and so were most other people, in the early days of meson theory. We thought that the theory should explain immediately and without much effort the forces between nucleons and we were of course terribly wrong. It turned out to be a very complicated affair, which was not even approximately solved until about 1955 by Chew and Low.
So these questions about nuclear forces and the various calculations of perturbations, calculations of binding energy, the explanation of reactions— all of these were at the very forefront of discussions.
They certainly were. I think you have put it quite correctly. I would say there were two areas of special activity. One was the question of nuclear forces and the other was the prediction of nuclear reactions. As I said, we felt pretty confident on predicting nuclear reactions and not very confident on nuclear forces. So the endeavors to find nuclear forces were extremely important. In my paper in 1940 I tried to explain on the basis of meson interaction the binding energy of the deuteron and the quadrupole moment of the deuteron. Of course one should really turn it around. One should say, "Using these two data, we tried to determine the constants in the nuclear force that were constants in the interactions between mesons and nucleons.
In these constants you were not bothered by any infinities?
Well, yes and no. If you took the old theory of Yukawa—this involved the exchange of scalar mesons—then you get, at least in first approximation an interaction e -ur/r , the so-called Yukawa potential. And this Yukawa potential in first order does not give rise to any infinities nor any trouble in solving the Schrödinger equation. If you use that, however, you don't get the quadrupole moment of the deuteron. In order to get that, you had to use either a vector or a pseudo-scalar meson. By the way, in the paper which we mentioned, in 1940, I was very stupid—I didn't know about the pseudo-scalar meson or I 1gnored it. And using vector mesons, I had to assume that only neutral mesons were exchanged between neutron and proton in order to get the right sign of the quadrupole moment. When you take pseudo-scalar mesons, as people later did, they got the correct sign, assuming exchange both of charged and neutral mesons. So this is a particular case of stupidity. So you had to use vector or pseudo-scalar meson exchanges. As soon as you do that, you get into trouble. You get a tensor force which goes as the inverse cube of the distance; and in any such force the Schrödinger equation is not soluble. It gives infinite binding energy, so I was bothered by infinities. The device I used and everybody else used and still uses is to cut the interaction off at small distances or high momenta and that's done to the present day. And so we used in this connection the same procedure which had been used by electromagnetic field theory for many years in order to make the self-energy finite.
Well, this period shortly before the war is very exciting from another point of view. Two of the most remarkable processes which seemed to have determined the direction of the evolution of technological society were also discovered. I have in mind fission and fusion. You were yourself responsible for the theory of energy production in stars, with your two pioneering papers of 1939, which was at approximately the same time that Weizsäcker also thought of these questions. I would like to ask you about this—what led you to these questions and also the excitement of working on this very remarkable problem.
Very good. I should go back even a little bit further. The thing which brought me to work on this problem was a meeting in Washington, a very small meeting of some thirty scientists, maybe even less— astrophysicists and theoretical physicists mixed—which was arranged by Gamow and Teller every year and which was sponsored by the Department of Terrestrial Magnetism of the Carnegie Institution. The department was under the direction of Tuve, whom I mentioned earlier, and this department had started in '36 to sponsor yearly conferences in April on theoretical physics, and that's why I wanted to go back to '36. These conferences I think were among the most stimulating meetings that I have ever attended. It was a very small group at that time—something like 20, I think, in the first two meetings—and in these conferences we discussed the problems which most interested us, mostly nuclear physics of course. I don't remember who all was at the conferences. I'm sure there is some record of this.
I have it here.
I don't have the two earlier years. I have the record of 1939. Maybe many of the names will be the same.
Well, no, it was different each year. It would be interesting to find out the '36 and 37 conference.
We've written to Tuve on this and we intend to get all these records if we can.
Very good. For instance, in one of these conferences, I think it was in '36 probably, Teller discussed the modification which he and Gamow were proposing for the Fermi theory of beta decay, which by then I reported in my Reviews article and which is known as the Gamow-Teller coupling or Gamow-Teller selection rules, and which is now known as the axial vector coupling in beta decay. But there were many other interesting papers.
Now, in 38 Gamow and Teller decided that they would have something new, that they would have a new conference on stellar energy. In this they brought together some of the best astrophysicists, for instance, Strömgren, with the theoretical physicists who had normally attended these meetings. It seemed very exciting because apparently the astrophysicists were at a loss as to what to do. 0ne of the participants was Weizsäcker at that time, and Weizsäcker reported on some of the attempts which he was making to explain energy production. I think nobody at the conference had any question but that the energy production must somehow be due to nuclear reactions. This of course was very different from the original ideas of Eddington several years earlier—ten years earlier or so. Eddington thought to use annihilation of matter to produce the energy. But nuclear reactions were well established by this time and gave a good, large amount of energy, and anybody could calculate for himself that nuclear reactions with abundant elements were sufficient to keep the sun shining for the past life of the universe and of course many billions of years thereafter. So this was more or less unwritten common background.
Was this articulated and discussed or was it just felt by everyone?
I think this was probably discussed even. Well, I don't think it was discussed in detail because it was assumed as so obvious. People probably referred to it and said, "Well, of course, there are nuclear reactions but now what? What are these nuclear reactions?" I was otherwise impressed by the total ignorance which pervaded everybody at the meeting. People really were at a loss as to what to do and what reactions to consider. Weizsäcker in particular tried to do two things at the same time—namely, to build up the elements and simultaneously generate energy. And to build up the elements he wanted neutrons on the basis that neutrons easily enter the nucleus. And so he was discussing at the meeting various ways how you could get neutrons. The main step which I took was to get away from this, to get away from the coupling between building up the elements and generating energy. But that of course came later.
Another point which excited everybody at the time was that apparently even without knowing the source of energy, there were internal discrepancies in the calculations of the astrophysicists—namely, in one way they calculated the central temperature of the sun as some 40 million volts and in another way they calculated less than 20 million volts. The mistake which they were making at the time, which was soon afterwards corrected, was that they assumed that most of the material of the sun and of other stars was of the same composition as the earth—namely, mostly heavy elements, heavy starting from carbon and going up through iron. And if you assumed that, then you get this discrepancy. Later on they discovered— and I have forgotten just how and when this came to my knowledge—that the main constituent is really hydrogen; and with hydrogen, the two determinations of central temperature came into agreement. Now, I am very vague how this came to my attention. I remember that at the meeting in April '38 everybody talked about these discrepancies and everybody talked mostly heavy elements, and I remember on the other hand that later on in that year and not very much later, before the summer, I was told that it was all right to assume hydrogen as a major constituent—if not the main constituent.
But you had come independently to the same conclusion?
I needed this conclusion, but I had not come independently to this conclusion, but I somehow got this information and used it. I simply have forgotten how this went, and I think Strömgren is the man to ask just what the situation was. Now, it may also be true that I remember the April meeting incorrectly. I know there was a discussion about this subject between von Neumann and myself and a couple of other people in which von Neumann pointed out this discrepancy, but it is entirely possible that Strömgren already knew the correct answer.
Weizsäcker did not take part in any of these meetings?
With von Neumann?
With von Neumann and Strömgren.
No. Strömgren was not part of it either. This was just a group of three or four people, ignorant people, who hadn't previously worked on it, discussing the deplorable state of affairs. Neither Weizsäcker nor Strömgren were included. Nobody, in fact, in this small group knew anything about the subject. So we didn't have the benefit of first-hand information in that small group. That was a discussion of half an hour or so. The important discussions included Weizsäcker and Strömgren, the actual meetings; and I just happen to remember this half- hour conversation because these discrepancies were mentioned.
Did you mention to us who was present in the smaller informal group?
I don't know, I remember only von Neumann. It may have included Teller.
The story is told that when these problems were mentioned in the meetings that it was said that "Bethe will find the solution before dinner is over."
That is purely apocryphal, and similarly apocryphal is Gamow's story about my solving it on the way home.
Yes, I read that story, too.
That is, in order to get your dinner on time, you had to come up with a solution.
But at least before some dinner in the following summer, you had worked out the proton-proton cycle.
Yes. Now, let me go through the entire sequence. The proton- proton reaction was not my idea and not Critchfield's idea. The two of us wrote the paper and got the credit for it, but I think it was really Gamow's idea to look at that. Critchfield was Gamow's student, a Ph.D. student, and I think this paper was probably Critchfield's thesis. So Gamow suggested to Critchfield this topic. Critchfield did it, but didn't feel certain about his methods and so he sent me the paper for criticism and correction. And I found a few factors of two which I would change one way or the other, and I used perhaps a somewhat more powerful way to calculate the wave function of the deuteron. I made minor additions to the paper. I was very much interested in this paper of course in connection with the conference, but I don't remember whether the paper was sent to me before or after the meeting—after this April meeting. I have a vague feeling that it was before the April meeting. Now, this feeling gets a little stronger now because I have the impression that when I went home to think about the conference, I had clearly in mind the proton-proton reaction, and we had been told enough about the properties of stars so that I could figure out that this reaction was insufficient to give the energy for the hotter stars—in particular Sirius A and Cygni Y or something like that. I don't remember. It's in my paper.
The hot radio stars ...
No, it's not a hot radio star. Those were not known at that tine. It was at that time the hottest of the well-investigated stars in the main sequence. There may be a hotter one now, but at that time it had the greatest luminosity of all stars in the main sequence which people had good data on. All right. I was told that the central temperatures were not terribly different between Cygni and the sun, but the luminosity differed by a factor of 10,000, I believe; and since I knew the proton-proton reaction very well, I knew that it could not explain these very large differences in energy production. On the other hand, I was very fond of the proton-proton reaction. I should have been even fonder because it is now believed that it is the reaction in the sun—the dominant reaction in the sun. And so I set out to prove that it couldn't be anything else. Now, I had at my disposal—and this has never been sufficiently credited—a paper by Gamow and Teller on the rate of thermonuclear reactions. This came out I think early in '38. And using this and my knowledge of nuclear reactions from the three volumes I had written, I set out to discuss the various nuclear reactions that could occur between protons and other nuclei, and I would have gone on—and did go on, in fact—to discussing reactions between alpha particles. So doing this, very soon was able to rule out reactions between protons and helium because they don't give any product; the reactions between protons and lithium, beryllium, boron, because they immediately consume these elements, and there isn't very much of any of these elements and you could very easily show that the energy supply would last only for a few hundred thousand years if these were the elements responsible. And having come to that point, I was almost sure that nothing would work but the proton– proton reaction. But then I looked at the next, which was carbon, and that did give the right order of energy production; and looking at it for about a day, I recognized that there was a cycle which returned to carbon. This must have been sometime during May, 1938.
There you used the compound nucleus again.
I certainly did, and I used the theory of the probability of nuclear reactions, which I had worked on. And then after a few more days I found that this reaction in contrast to proton-proton gave a tremendous dependence on the temperature, about the 17th power of the temperature, which was sufficient to explain the difference in energy production between the sun and Cygni, with a very modest ratio of temperatures, which the astrophysicists would allow me. Moreover, somehow I must have known already then, which was in May, that hydrogen was a major constituent, perhaps the major constituent; and this process used up only hydrogen, and that was fine and would mean that the fuel supply would last billions of years. The whole work took me about a month from the conference.
Thirty dinners. It may have been forty. And it was a wonderful time because I knew that I really was getting somewhere.
During this time you had no association with Weizsäcker.
And Weizsäcker was also independently doing some of these things.
He did. Now, I must say in general I don't like to engage in priority fights, but I must say Weizsäcker had very very little, and I think in this particular case it is quite unfair that he is mentioned often on a par with me. He maybe had one per cent of what I had. I read his paper recently just because I happened to be in the library, and I was really amazed by how little he had. He had not investigated the things which I just mentioned. He had not investigated the rate of the reaction, nor the dependence on temperature. He had not obtained this explanation that the sun and the hot stars could go on the same reaction. He had not looked at the fuel supply question, and at the same time he mentioned this cycle in one section of a long paper which discussed absolutely everything in the sun—I shall not say under the sun. Absolutely every reaction was discussed, and it was not stated that this was it. So I think it is not right that he is mentioned as a co-discoverer.
During this period you started to mention the sense of excitement that you had. This was because you knew you were getting somewhere— is that right?
Can you tell us just a little bit more about the circumstances?
Well, it's the kind of satisfaction I have gotten about five or six times in my life when I really got to some important result. I can't tell you much more. I work very feverishly. I guess I probably worked 15 hours a day or something like that, and I just couldn't leave off working on this. I probably gave very bad lectures—I would imagine. I probably didn't spend much time preparing them. I don't remember that part. I just remember the feeling of tremendous curiosity how it would go on, what would be the next step. I can't tell you specifically about this occasion because I don't remember it to that extent.
You have told me what I really wanted to know—the amount of involvement of time and how you felt about it. This was the first major continuous piece of work after the completion of the three review articles. Is that right?
Certainly the first major, yes, that's right.
In your work on the energy generation in the stars you had got both the essential thermonuclear reactions—the proton-proton cycle and the carbon-hydrogen cycle. You mentioned about those occasions in which you felt the sense of excitement and discovery. These great occasions, I think even in the life of a Hans Bethe, are numbered. I would like to ask which were the other occasions.
Well, I think the first was the theory of stopping power, which I did in 1930, and which I still consider perhaps my best paper. Maybe not. But it certainly is one of my best papers.
It was your first love.
It was my first love and I stayed with it, and I think it was in a way quite perfect. It is called something else, not theory of stopping power in this list, but passage of swift particles through matter or something like this.
1930, yes. [Zur Theorie des Duhgangs schneller Korpuskularstrahlen durch Materie, Ann. d. Phys. 5, 325-400 (1930]
This was the paper which submitted to become a privat-dozent in Germany.
Your "Habilitationsschrift"? ("Inaugural dissertation"?]
Yes. And this has remained good all the time. I have made improvements on it in detail, and several of my Ph.D. students have worked on it and have added to it, and a few other people have. But fundamentally it still is the same as it was in 1930. So that was the first. I did not feel the same exaltation on another paper which I wrote a little earlier, "The Term Splitting in Crystals," which was about a year earlier, and which also has been very much used. But that was much more an application of a given technique, namely group theory, which I had just learned, to a problem. It was solving an interesting but fairly straightforward problem. I did not then see the applications that this would have. I would have been excited if I had seen this. So 1930 was the first really satisfactory paper. The next was the theory with Heitler in 1934 of bremsstrahlung and pair production. In the case of these big nuclear physics articles I had the feeling of a job well done but not this feeling of having made a great discovery. So I think the third was in fact the energy production in the stars, and the fourth was the theory of the Lamb shift in 1947. I think these probably are the four most important ...
On these four occasions you felt that you had a special contact with nature.
Yes. That's right.
I would like to pursue this, and it is almost a psychological question I am asking. It requires an answer based on your reception of what you were thinking. The satisfaction was derived perhaps from the understanding that you were beginning to get closer to an understanding of nature or of certain processes occurring in nature. At the same time did you feel that it would be regarded highly within the profession by your peers? Was that part of the excitement? Knowing that it was good work and that others would regard it as such?
I don't remember. I think I was certainly not conscious of this on the first occasion—1930. I think I must have been conscious of it in 1938 on the stars because it was a problem which everybody was interested in. I don't believe it was the principal motive or even a very strong motive. The strong motive was to find out.
Well, perhaps I may cite an example of fallibility which I suppose later on must have given you some joy that you were wrong? And that is your 1937 paper on the maximum energy obtainable from the cyclotron.
Yes. Well, that's an interesting paper. We were wrong and yet we were right, and I discussed it during the last year or so with other people who are actually in the business of designing high-energy machines. We were right in saying that the design then existing, the design principle, had an energy limit. We were wrong in thinking that this was the only possibility. Well, maybe we didn't even think so. We don't say that you couldn't design some other machine. We didn't have the imagination to think of other machines. But I think on the whole our paper has perhaps done more good than harm. It got us into a big scrap with Ernest Lawrence, who didn't like it at all, but I think the existence of this paper was a great stimulus for other people to think of new ways. If this paper had not existed, I think very likely people would have tried to just increase the size, increase the electric field of the radio frequency accelerator, and would have tried to do it by brute force. But since our paper existed, McMillan and Veksler in Russia thought very hard, and for themselves, and found a very ingenious solution totally different from anything contemplated in 1937. I suppose they couldn't have thought of this solution if it hadn't been for the tremendous development of radio frequency techniques during the Second World War, which made frequency modulation a very easy technique; whereas in 1937 it would have been quite hard.
Did many people address themselves to this type of problem? It occurs to me that there weren't too many papers of this type, that the machine builders were writing about their work; other people were analyzing the results of experiments, but there were very few theoretical papers devoted to the inherent limitations of the instrument.
That is correct. There was, however, one person at Berkeley who was working on it. That was Robert Wilson, who is now a director of this laboratory [Cornell]. He was then a graduate student and a very clever one, and he was thinking about the problems of magnetic focusing, electric focusing, and doing very interesting work on these problems. So the Berkeley group was aware of some problems, and I think Wilson was given the task to understand why the cyclotron worked in the first place, and he did that very well. It was sort of a miracle that it did work and that the beam stays together in the rather narrow space between the dees rather than spread all over the place. And this is done by focusing, which Wilson investigated.
Were there others who played this sort of role?
I don't believe so. However, stimulated by our negative paper, Thomas, who is now at the IBM Institute, the Watson Institute, invented one way of getting around our limitation—namely, to shape the magnetic field in such a way that the magnetic field would vary, as in Wilson's directions, around the cyclotron. And such cyclotrons have been built, only in the '50s, not immediately, and are quite workable.
In the later period, were there any people who were looked to for this type of theoretical guidance in inherent limitations and possibilities of uses of machines? I think in one case of Fermi as a possibility. Someone mentioned that he played some role in this, and I wasn't clear about the context of that.
I don't remember about Fermi, but it's possible. He did almost everything.
But you started to give me another answer before I mentioned his name.
Yes. Well, since the war, this has become quite customary. I think every major machine design group includes theoretical physicists, and one such group really made a brilliant discovery. That was the Brook- haven group discovering strong focusing. This was discovered principally by three people—namely, Courant, Snyder and Livingston. (Livingston is the same Livingston who was associated with me in writing Part C of the big book.) Livingston was the only experimenter. The other two were theorists. They knew in particular the theory of solids in which there are electron bands, and they applied this to the strong focusing of beams in the synchrotron. This was extremely important. It was mainly theoretical, and they did a brilliant piece of work both in inventing it and then discussing it. It is true that the same principle was discovered by Christofilos in Greece, who was an elevator engineer. Christofilos has invented many other things since then. He is now at Livermore. I think not more than ten per cent of his discoveries work, and this is perhaps not for publication in the present form—he is almost impossible to understand because he puts things in a form which is very unconventional. He does not take the trouble of putting it in a form which is understandable to other physicists, and it is usually very very sketchy, and by contrast the Courant-Livingston-Snyder paper was very well thought out in all its consequences; and after the theory was published, it was perfectly clear that it ought to work. I haven't Christofilos's first paper, but judging by his later efforts I would say that it is a somewhat similar relation as between myself and Weizsäcker.
The other great event just before the war was the information that was received in this country about Hahn and Strassmann's discovery of fission. Were you present when Niels Bohr brought this information back and during the discussions that ensued?
Not immediately. The first discussions were held at Princeton with Wheeler and a few other people, and I suppose that Fermi must have been involved and Szilard. I got involved in the discussions at the 1939 Washington meeting, the small Washington meeting of the Department of Terrestrial Magnetism, which we mentioned before in connection with the stars. And at that meeting it was a very live subject of discussion. There was a closed session. You certainly have heard about that before, and I don't suppose I need to go into this.
The members of the press were also asked to leave when the discussion started. That was the first example of self-imposed secrecy in this effort.
Was everyone aware when the session was closed what the reason was?
I think not everyone was aware when the session started, but everyone was when the session ended. Certainly Fermi, Bohr, Wheeler, Szilard were aware, and I'm not sure whether all of them were there. sure that Szilard and Fermi were there, and I'm pretty sure about Bohr.
Bohr made the announcements, and Fermi was there, and, strange, our list doesn't show you. [Chart with Conference attendors].
That's interesting. Whether Wheeler was there, I wouldn't be sure.
And then Bohr went back to Princeton, and Bohr and Wheeler worked out the theory of the fission reaction. And that was I suppose a great triumph of the compound nucleus again.
And this time even more for the liquid drop model.
Both together, yes. In this case they used the liquid drop model more than just getting the surface energy, which they could have got on the Weizsäcker formula. By the way, I should say that Weizsäcker has many merits and has made many contributions to nuclear physics. It's only on this one that I think he has been overrated. But they worked out the theory. In particular, they found that it should be the odd atomic weight isotopes which undergo fission when bombarded by neutrons. This was a very important result and they were very confident of the theory when they stated this result.
And it was the final glory of the liquid drop model because the idea of the instability of the heavy nucleus was explained by the deformation of the nucleus.
By the deformation. The liquid drop model now was used much more than before. That is, deformations were considered and in particular a deformation, which would lead to fission—to a constriction in the middle and fission.
I would like to ask you about the war years at Los Alamos. Now perhaps the stage was set with the great discoveries of the concept of thermonuclear reactions and then the fission reactions; and perhaps historically it is very interesting that science was in a sense ready to tackle the problems you were faced with at 'Los Alamos. At least in the background they did exist, and it was your burden to cope with the problems and to use these principles for some purposes.
Before I answer this, I would like to add one more point to the energy production in the stars.
I have a question on that, too.
Namely about the time involved. I did the work in May '38, here and at Harvard in '38. I was rather slow in writing it up because I wanted to have it perfect and work everything out, including, for instance, the point that no other nuclei can participate to any appreciable extent, including a theory of how the stars will change their temperature and luminosity in the course of their lifetime, including a discussion of the reactions between alpha particles and a few other things. So it took—I don't remember exactly—till September or so to write it up. I don't know when the paper was sent into the Physical Review.
It wasn't published until 39.
It was not published until
There were two papers.
One is apparently a one-page paper ...
Yes. When was that published?
Let me find it. You give me the number of the volume.
Volume 55 in 1939, page 100. [Looking at Physical Review]
There is 103 and 434. Well, the story is that when I had written the paper, my good friend, Robert Marshak, who was then a graduate student, drew my attention ...
The first one is a letter.
December 15. The other may be dated earlier. Robert Marshak drew my attention to the fact that there was a (this is September 7th, '38)..
Well, that's not too much of a gap since you did the work in May.
Yes. Well, he told me that the New York Academy of Science was offering a prize of $500 for a paper on the energy production in stars with the condition that the paper must not have been published previously. And $500 was a good deal of money for me at the time, so I asked the Physical Review to delay publication and sent the paper in for the prize, which I got, and of which I gave a finder's fee to Marshak. Then after I got the prize, I asked the Physical Review now to publish it; and in order to accelerate publication I sent in the letter to the editor, which is really an abstract of the paper after I had sent in the paper, and that explains why there are two papers and why they were so much delayed. The Physical Review in those days usually printed things much faster, but, you see, it took six months for publication, which was not the fault of the Physical Review,
There were fewer papers at that time.
Much fewer papers.
Now, regarding the work at Los Alamos, the stage was set. Did you have much time or attention during this period for work in nuclear physics? How did things go—at least in the period you were the chief of the theoretical division at Los Alamos, and the number of participants who came—there was not a very strong compartmentalization of things and the ideas were very fluid?
Well think we should consider not only Los Alamos but also the work of a small group consisting of Oppenheimer, Teller, myself and three other people in Berkeley in the summer of '42. In the summer of '42 work we mostly worked on the super, on the hydrogen bomb, which was very foolish of us since the atomic bomb did not exist and nobody knew how to make t. The only thing we knew pretty well was that there would be material of which to make an atomic bomb, but even that was far in the future. It was before the chain reaction had been proved by Fermi. Well, foolish or not, part of our attention then was on the thermonuclear reactions which could be established in deuterium, and we recognized that hydrogen-3 and helium-3 would be formed, and hydrogen-3 would be important for a subsequent reaction. And we didn't know the cross-section for the collisions of hydrogen-3 and deuterium, so we sponsored an experimental project which was then carried out at Purdue University to investigate the cross sections of hydrogen-3. It turned out that this cross section was very large, which made thermonuclear reactions much more possible than they had been otherwise. But during our work we had discovered a lot of other reasons why it might not be so easy to establish a thermonuclear reaction. All this is quite well summarized in the so-called "Super Handbook" of Los Alamos, which is declassified and is accessible. So the principal results of our work in 1942 and some of the work during the war at Los Alamos are published. I don't think it was ever published in the literature, but you can get it from Los Alamos.
As one of the reports.
One of the reports.
What brought you to Berkeley that summer? Was it an invitation to work precisely on this problem?
Precisely on this problem. I was asked to work on the assembly of a possible atomic bomb. We got diverted, mostly by Teller's ideas, to work on the super.
You started to mention about the group. It was Oppenheimer and yourself and Teller and three others.
Konopinski, Serber, Van Vleck. I think there must have been somebody else—a few graduate students. Now, in spring of '43 Los Alamos started, and we now began thinking about the assembly of an atomic bomb in earnest. 0nly a relatively small part of the problems were nuclear physics. The lab started out with considerable emphasis on experimental nuclear physics and on theoretical physicists who knew nuclear physics. But in the theoretical division only one group of five or six major groups was concerned with nuclear physics. That was the group of Weisskopf. And only half of its work was nuclear physics. The other half was to calculate the efficiency of an atomic bomb if it was made. So this would mean that about ten per cent of the work of the theoretical division was nuclear physics, and I think less than ten per cent of the work of the whole laboratory. The nuclear physics came in only in one respect really—namely, to estimate the size of an atomic bomb, the critical mass, before we had enough material to construct the critical mass. In fact, the amount of material we had was a few milligrams of uranium 235 to start with, any amount of normal uranium we wanted; and I remember t was a great event when Oppenheimer got a wire of plutonium, which was a few micrograms—maybe a hundred micrograms. So we had very small amounts of material—however, enough to do some experiments on cross-sections; in fact, quite good experiments on cross-sections. And as time went on, the amounts increased and the experiments got better. It was the task of Weisskopf and his group to sort of interpret and direct these experiments in the direction which would be theoretically most useful. It was then the task of other groups, especially Serber's, to calculate the critical mass on the basis of this information. That wasn't nuclear physics anymore. That was the problem of diffusion of neutrons in the given substance, in uranium or the surrounding reflector which we would use to keep the neutrons in. And then other groups were concerned with the assembly of the material, which was a hydrodynamic problem. That was especially Peierls' group; and still another group with numerical calculations calculated on ordinary (hand) desk computers, another group with calculations—that group was established later on—on IBM machines, and there were a couple of groups on special problems. But there was relatively little nuclear physics. Now, Weisskopf's group—and he himself in particular—used all the knowledge he had on nuclear reactions. So this was extremely important.
And this period was characterized by study of nuclear reactions, calculation of cross-sections. There was no essential advance in the theory of nuclear structure.
Can you speculate what was the state of the work in the theory of nuclear structure such that this was a retarding influence? In other words, did it interrupt something that appeared to be developing very rapidly?
Yes. It did interrupt the study of nuclear structure. All the experimenters went to war work of some sort or other. There was, for instance, very interesting work going on here at Cornell on the velocity selector. This was Bacher, who mostly started this, where you measured the velocity of neutrons by the time-of-flight, which was done by radio techniques. This work was stopped in '42 I think after it had given some information pertinent to the uranium project. Similar work, however, was continued at Columbia under Havens and Rainwater, and I would say that this work of Havens and Rainwater got more support because of the uranium project. There were the cross-sections of the uranium isotopes which were of great interest to the project. The cross-section is a function of energy. And the Columbia project was specially supported by the Manhattan District. So this went fairly well. On the other hand, the fast neutron work, and generally the work on fast nuclear reactions was essentially stopped, partly because the people went to war work, partly because one of the best machines—namely, the Harvard machine—was bodily transported to Los Alamos to do work directly pertinent to our problems.
This was also in a certain sense a period of gestation because there had been previous to this period and during this period a surfeit of ideas, and the energies of the theoreticians and the experimentalists were employed toward very specific goals. So in the subconscious there were growing problems and in the conscious there was growing a certain technology.
Why don't you put a question mark on the end of that?
Now, this was a period of gestation because immediately after this period things started exploding in all sorts of directions.
That's absolutely correct. However, I don't think it was a period of gestation of ideas because none of us had the time to think about it. There was a lot of unfinished business, and particularly of course there had been a great desire to have higher energies, and at least Ed McMillan must have had a period of gestation—namely, of thinking about new ways to accelerate particles, and he came up with this ingenious idea of frequency modulation and phased focusing, which led to both the synchrotron and the frequency modulated cyclotron, the synchro-cyclotron. I think there were very few people—I think he was very much the exception—who had time, even spare time, to think what they would do next. Certainly I didn't give any thought to these matters, and I doubt very much that anybody in the theoretical division did. 0ther projects may have been not so busy. On the whole, the war was an interruption of the development and an interruption even of thinking. There were only very few things that were going on. One was the effort to make sense of meson theory, and I mentioned already before the names of Pauli and Wentzel, who did work hard on meson theory and found all sorts of trouble. But it was not until after the war that we got any important clues on it.
There was immediately after the war the first international conference on elementary particles at Cambridge in l946. Do you recall that?
I'm afraid I don't. probably was there, but I don't.
There was one in London, the international conference on fundamental particles.
This took place in Cambridge in England. It was sponsored by the Physical Society.
Yes, and I don't know about it.
0ne question about the effect of the war: At Los Alamos you had a very concentrated high-powered group of physicists from all over the world.
And Bohr was in and out and others. I know from what others have told us that the social relationships were rather close whether you wanted it or not, because these were the conditions. And yet you say you didn't have time to think about these problems. I gather then that the discussions were not on the problems of nuclear structure; that there was interaction among the physicists, but these didn't lead to informal seminars, walks up the mountains to discuss these questions.
We had walks up the mountains in great profusion, but we discussed the problems of the project, not pure nuclear physics. If we discussed anything outside of nuclear physics, it was the political impact of the atomic bomb to be expected. I personally was not involved in this until very late, but Bohr was deeply in this and this was his real interest, and Bohr had long conversations with Oppenheimer which brought 0ppenheimer into this at a very early stage. Oppenheimer was very much indoctrinated by Bohr's ideas of international control. But of course on our walks we might have discussed the progress of the war quite a lot, but certainly on the walks on which I went, there was very little general nuclear physics.
The other thing about it is the fact that a whole group of young men were brought into contact with the leaders of the field almost instantaneously. Do you think that this had an effect on their own careers?
This certainly had a great effect. Very many of these young men became very productive immediately after the war. Certainly there was a feeling at the end of the war of urgency to go back to those interesting problems we had left without knowing very specifically what these problems were and what we would tackle next. We certainly wanted to get back to nuclear physics theory and so on, and this feeling which was very strong in my generation was transmitted to the younger generation. So they brought home the enthusiasm for it.
And the older generation brought home the youngsters with them. You brought home Feynman. Others brought home new people, and there was a lot of swapping going on.
Indeed. And then the other thing which you mentioned should be more emphasized—technology. A tremendous amount of technology was developed during the war, perhaps most importantly the radio technology in the MIT Radiation Lab and similar places, which made t possible for the experimenters to use completely new techniques after the war. In Los Alamos and other nuclear projects of course we thought mainly of studies of neutron physics, neutron diffraction, neutron cross-sections, velocity selectors and so on. These were the things in our minds as the techniques that would flow from the Manhattan Project and that have really been exploited to a very great extent after the war.
Since unfortunately my time with you today is coming to an end, I would like to pose if I may to complete this part of the story a set of questions, the discussion of which perhaps you could start, and my friend Charles could take them up with you. 0ne of the first would be to start the story as it was taken up after the war and in the theory of nuclear structure the greater events were the resurrection of the shell model and then later on the calculations of the optical model—your comments on these. I would also like to ask about the mathematical matters which became common to low-energy and high-energy nuclear physics. There were a number of things that were discussed and used in low-energy nuclear physics. SU 3 was used by Elliott and symmetry considerations, which only much later used in high-energy physics. So the question would be: the matters common to low- and high-energy physics, and then how the developments in low-energy nuclear physics affected particle physics and vice versa.
And then the other question prior to that was models.
Could you leave us this page because there is so much in this.
I shall indeed. Close to the heart of this story of course is ideas on nuclear matter and the nuclear many-body problem, the Bethe Goldstone theory, and then more recently you were doing some very fascinating work with Reed in the use of very realistic nucleon-nucleon potentials, the knowledge of which has been gained by experiment, and then you are using t in theory. I think that it belongs in a very pertinent fashion to the history of this discipline to find out how the use of these realistic things as against more simplified and idealized models is going to affect the development of this discipline. I would also like to ask about your judgment—and this is a matter of opinion, of preferences, of what is the dead wood of nuclear structure, of nuclear physics—things that have been relegated to the wastebasket. And one question which has been the guiding light through all these years: that is, what holds the nucleus together? What is your judgment of that? This is a series of questions which in my view I would like to pursue, and I am sure that you more than anybody else would be receptive to them.
That gives a basis for a lot of answers and it is essentially the whole story of post-war nuclear physics that centers around these questions. I'll try to give enough emphasis to the points you raise. I think I'll continue to proceed on the whole in historical fashion— that is, do it by time rather than in this order.
Perhaps now is a good time to think about your original statement. You said you preferred not to have things introduced by questions, and yet we've tended to do some of that because we've needed a great deal of clarification.
Well, it has made t more lively.
But in this remaining time perhaps you can think in your own mind where you left off this long-range history that you had been giving us and what is the logical point to pick it up ...
These questions have not been to prod you.
0f course. I think I have got essentially to the end of the war, and I think I can just start at the end of the war and start from there. Maybe I can try to answer one or two of these questions.
When you perhaps take up the story, one would be very interested in the Lamb shift because in this context one would like to know about the shell model because that was the next one to come.
Of course the next thing really to come chronologically was new accelerators.
Let's talk about that then.
The great invention was that of McMillan and Veksler of phase stability—namely, that if you sent a beam of particles around an orbit and saw to t that they were accelerated in synchronism with the orbit— that is, whenever they come to the accelerating voltage, that voltage is in the right direction. He proved that if you do this, then the requirement of synchronism is not very strict. We had thought—Rose and I—in 1937 that there had to be strict synchronization, but McMillan proved that was not necessary; there was phase stability. If you had particles which came a little bit too early, they would be accelerated less; and particles which came a little bit too late, they would be accelerated more than the bulk of the particles, and thereby the particles would automatically be synchronized if only you had rough synchronization between the frequency of revolution in the orbit and the frequency of the radio gadget which accelerates them. Now, many devices were built all around the country exploiting the McMillan-Veksler idea. There were many synchrotrons for electrons— one here for 300 million volts, one at Purdue, one at MIT, and one at Berkeley of course and one or two others in addition. Three hundred million volts was chosen because this seemed enough energy to make mesons. And here of course we have the connection with high-energy physics. We knew that nuclear forces presumably were made by mesons. We knew that the meson had a certain mass. We didn't quite know what it was. It turned out to be 140 MeV, but we didn't know that yet. To be on the safe side, everybody agreed that 300 million volts was a good number, and indeed it was a good number. Simultaneously a lot of synchrocyclotrons were being built by Berkeley of course and by Columbia and the Carnegie Institute of Technology in Pittsburgh and Chicago and Rochester and a lower energy one by Harvard. And these machines came into operation gradually around l948, and they really changed the picture completely because with these machines we got high-energy particles and we were then able to explore the nuclear forces, and perhaps the most important problem as everybody saw it at the end of the war, and the problem which I emphasized in the set of lectures I gave at Schenectady in '46, was to find out something about the shape of the nuclear forces. And this was brilliantly done as soon as the big accelerators existed, and this of course is the basis of the theory of the potential between nucleons, which you mentioned, which Reed is now pursuing with me, and for which the basic work was begun in 1948, and the first somewhat complete results were available by 1957. So it took about nine years before people had a fairly complete picture. But quite a lot happened in the meantime, and I'll discuss that after lunch. [pause in recording]
We are resuming after a break for lunch and I think off the tape we just agreed that it would be good to talk about the effect of small meetings, informal and formal conferences, of physicists on the development of ideas in the field.
Very good. I mentioned before the small conferences in Washington in 1936 to '39. These I think had a lot of influence on the thought of the participants, on the problems that were later on treated. They certainly had a lot of influence on me, and I think quite similarly on others. I should mention in the same vein the summer schools at Ann Arbor, Michigan, which were regular events until the war. I participated in 1936 and '38. I know there were quite a number before '36. They were extremely stimulating. There was always a group of very well-informed people who had a lot to say and t was very much worthwhile to listen to each other's lectures, and I think this brought together the thoughts of several people for everybody's knowledge and stimulus. Each of these Ann Arbor meetings involved perhaps a dozen theoretical physicists, the Washington meetings about two dozen, and such conferences are of course very wonderful. The last ones of this type we had immediately after the war, and the first of this series was the Shelter Island Conference of 1947.
There was more than one at Shelter Island.
They were called Shelter Island conferences, but only one was at Shelter Island, as far as I remember.
That was in June of 1947.
Was there not one earlier at Princeton, a different sort of conference but bringing people together?
There was one at Princeton in '46, and that should be mentioned. Unfortunately I wasn't there, but this must have been very good. This was just following a Physical Society conference—I don't know what time of year—but that was somewhat bigger.
This was September 23 and 24, 1946, at Princeton. It was the bicentennial conference. That was bigger. It was about two hundred people, not all physicists.
Yes, Now, Shelter Island was very exciting and especially because just before that conference Lamb had discovered the Lamb shift. At the same time there were talks by several people, but particularly Kramers, on a reinterpretation of the self-energy of the electron in electrodynamics. And I think there was a feeling at that conference that the two things somehow might be connected. Well, I was very much stimulated by this, and in this case Gamow's story was right, only t hasn't been written about this case—namely, I did get the main part of the solution of the problem on the train going from New York to Schenectady, and what remained afterwards was just to put in some numbers. And this theory of the Lamb shift then in turn gave rise to a lot of work by other physicists to put electrodynamics in order—particularly Schwinger and Feynman, who in their separate ways went after the problem and gave two separate solutions which were found to be equivalent in results but entirely different in method. The second conference of this type ... The second conference was Pocono, wasn't it?
In March of l948.
And the third was on the Hudson near Newburgh, in '49, I think. It was in a very pleasant semi-private place on the Hudson close to Bear Mountain bridge.
Is Ram Island the same as Shelter Island?
That's not Ram Island here you're referring to?
Oppenheimer in his talk last January mentioned Ram Island. I can look t up. Was Feynman there? I think he mentions something about that.
I think he mentioned the name of it.
Well, anyway, in the second conference both Feynman and Schwinger presented their theories of quantum electrodynamics, and this second conference was mainly under the impression of quantum electrodynamics. And t seemed to us that now the door was really opened to go ahead and understand everything. It turned out not to be true, but on the other hand the theory went a great deal farther from that point—the theory of renormalization. The third conference, if I remember correctly, had as one of its exciting news the shell model. That was the '49 conference. I am not sure whether Mrs. Mayer was a participant. I am sure that the shell model was discussed together with a lot of meson physics.
Do you remember where this conference was?
Well, I told you: t was on the Hudson near Bear Mountain.
Oh. The second one was the Pocono and the third one was that in 1+9.
That was '49. This was the last of the really small conferences.
How many people—perhaps a couple of dozen?
Yes, maybe up to 40, but certainly not more than that.
Was there a cumulative effect? I know new people were added.
New people were added, and very few people were dropped.
Does that account for the discontinuing of these conferences because they got too big?
Well, they were not discontinued, but they then grew into the Rochester conferences. After this conference on the Hudson. Marshak said that he would like to have one of these conferences at Rochester, and anyway the conferences were growing too big for a small resort hotel where they had been held before, and it was better to hold them in university lecture rooms.
On the earlier conferences, who organized them? Who did the inviting and who paid for it?
I am not sure. I think the first one was paid for by the Rockefeller Institute, and that Institute did the inviting.
Shelter Island—a man by the name of Maclnnes of whom I never knew before or after, who was at the Rockefeller Institute—I think he was a physical chemist ...
Again, he's connected with the Rockefeller Institute, so this would account for two conferences.
No, that's all Shelter Island. Now, the next two were organized by some of the members of the conference, and I am very vague who did it. I think 0ppenheimer had something to do with it. Maybe the Institute sent out the invitation— I don't remember. I think he had something to do with the selection of the places. I am not sure who paid for it, but I consider t somewhat likely that everybody paid from his own contract. By that time government contracts AEC and ONR) were very widespread, and that would be my guess looking at it from now, but I don't remember.
I have a note that the Shelter Island conference was sponsored by the National Academy of Sciences.
It was? Okay.
Perhaps by the National Research Council of the Academy—I don't know. And then you started to tell me that these three conferences evolved into the larger ones.
Yes, in Rochester. And Rochester started with I think between 50 and 75, and for several years Marshak valiantly kept the number down to 75. The Rochester conferences were I think primarily concerned with pion physics—that is, with meson physics. By that time there were some experiments on mesons. The first conference I think was early 1950. Fermi's experiments were beginning to come out in Chicago, and McMillan, I think, and Panofsky found mesons in the synchrotron, and in Berkeley mesons were produced also by the Berkeley cyclotron. So there was great excitement and interest in both the experimental results and the theories. Lots of theories were being started to account for these. I don't know whether you want me to go into meson physics at all. Let me say, however, one thing. In 1947, Powell at Bristol discovered the true meson, the Yukawa meson, which is the pi meson, and thereby ended the period of great puzzlement in physics—namely, the meson which was discovered by Anderson and Neddermeyer in '38 had none of the properties that Yukawa had predicted except the mass. And that is the mu meson of which t was found out in '46 or '47 that it has really no interaction with the nucleus. We know that even better now. It has absolutely no interaction except due to its charge, and so it was futile to try and explain nuclear forces by this meson. Well, Powell's discovery solved this, and after Powell's discovery meson theory and meson experimentation really came into their own, and we can then forget about them.
You would date this as the beginning of high-energy physics as a separately defined field.
As a separately defined field, yes.
Did people in nuclear physics up until that time begin to identify themselves as being in one field or the other? Was it that sharply defined?
Very much so. In fact, I think with the advent of pion physics, the great majority of the former nuclear physicists, both experimental and theoretical, switched over to meson physics and high-energy physics. And that included the majority of the members of the Rochester conference.
The very existence of the Rochester conference—it had as its name in 1950 its first session: "High Energy Physics Conference"— already implied that this was a very sharply defined field. That was not a symptom of a change about to take place. It was an evidence that the change had already taken place and was already institutionalized.
That's a good point. think it's a good point at which to drop that discussion.
Now, I think there was a good deal of high-energy physics discussion at the last preceding conference on the Hudson. I think there was very little of this at Pocono, and practically none at Shelter Island.
At Pocono there was a good deal of quantum electro-dynamics, wasn't there?
This was when Feynman presented some ideas and Bohr criticized them.
So Feynman feels.
So I feel. So I think everybody feels now, and I would think that even Bohr felt that way in later years.
It could have been because of Feynman's presentation.
Very much so.
Which might have appeared to be not mathematically rigorous.
Absolutely. It seemed not rigorous. It's a little strange of course, because if there's any of the older physicists to whom Feynman is a little similar, it's Niels Bohr, because both of them work very intuitively. Schwinger, on the other hand, works very mathematically, very solidly. He always goes by very clearly defined steps from one statement to the next. Feynman doesn't. Feynman has flashes of intuition. So did Bohr. Feynman finds it somewhat difficult to express his newest ideas because they are not yet fully formed. So did Bohr.
That would also explain Bohr's sensitivity to seeing this trait in others and then criticizing it.
Yes. And I don't know any other person among the great physicists who shares this trait to this extent. In many ways they are completely different. Feynman is so exuberant, and Bohr never was—at least not in the times I knew him. Well, that's that.
There must have been a recoil reaction when high-energy becomes defined as a separate field and the majority of people transfer. What happened then to nuclear structure?
That is a very good question. Now, I do remember that in some of the Rochester conferences a lot of the discussion was on the experiments on scattering of nucleons, of protons by protons at very high energy. And that I would still consider nuclear physics inasmuch as it was designed to give information about nuclear forces. Now, this work was done to begin with chiefly by Segrè, Chamberlain and their collaborators on the Berkeley synchro-cyclotron, and this work was reported and discussed in detail at the Rochester conferences. So while this was high-energy work in the nomenclature of 1950, now we would call it medium energy—a few hundred million volts. It was definitely designed for nuclear physics, and I also remember that there was quite a lot of discussion, and there were many papers attempting to derive nuclear forces from the interaction between mesons and nucleons. Brueckner and Watson had a paper which was discussed at one of the Rochester conferences. Marshak worked on it. There was a lot of Japanese work. So the connection with nuclear forces was quite strong, but only with nuclear forces—not with nuclear structure. The connection with nuclear forces still exists, and just in recent years, starting with ideas of Chew and then lots of other people, attempts have been made to derive nuclear forces from the existence of many mesons which have been observed in the meantime, and this has been quite fruitful although it has not given a quantitative description of nuclear forces.
The final question I have on the conferences before we get back to picking up the historical thread: Did anything take the place then of the Shelter Island, Pocono Manor, Hudson River type conferences in the field of nuclear structure?
No. In the field of nuclear structure, I don't believe so— at least not for a long time. There are of course conferences on nuclear structure. There just took place one such conference at Gatlinburg about a month ago in the middle of September. The one preceding that at Paris in 1964. Both of these were very good conferences, but they were big; so of course are our high-energy conferences now. But there are now separate conferences on the two subjects. I don't know when the first conference specifically on nuclear structure took place. For quite a number of years I wasn't really concerned with the subject, so I wouldn't know about it.
Which years would you say were those intervening years?
Oh, essentially immediately after the war until '55.
Well, I think that really has gotten into ...
Let me say one more thing. The conferences, of course, now are completely different from these Washington and Shelter Island conferences. They are very big. They are very tightly organized. There is a very severe restriction on the number of participants. They have become international with all diplomacy involved. There have to be enough Easterners invited and the right kind of Eastern people, and there is some quota for the United States and for western Europe, for Russia and other Communist countries and for the rest of the world. And the conference has now got up I think to 500, the latest one having been at Berkeley.
I heard that figure-4 or 500.
I don't know which. These were the actual delegates and then a number of other people could attend and I think could listen to the speeches by TV in a separate room. This was instituted I think first at CERN in '62 (two conferences back). So the conferences are still extremely good—well-organized; they really bring the newest results in experiment and theory. People have good discussions. But they no longer have the intimacy that we had in the Shelter Island and similar conferences, and that makes a lot of difference. You can't think during the conference any more, and there isn't really time for a full exchange of reactions or for full discussion. I think all of us who were in the early conferences miss very much this intimacy of the early conferences.
There's another effect, and that is that because you are pretty much in public, is there a tendency to be more competitive in a sense? Do you think this is how you could characterize many of the presentations, the desire to come up with a result first and to announce it at this The question was on the effect of of large audiences on the participants.
I don't believe that the field has become more competitive. 0n a personal level at least it seems to me that is not so. There has always been some competition. On the other hand, I think physicists of my generation and the next generation have not often engaged in any fierce and unseemly competition. I think there is relatively little on the personal level. There is somewhat more on the level of the laboratories. For instance, Berkeley, CERN and Brookhaven might compete to some extent.
For example, this theta-meson business. Would this be an example of it?
Yes, that would be an example. It's a very good thing that there are at least two laboratories because sometimes mistakes are made and they are then corrected by the other team, and this has gone both ways. But I don't think that this competition is really very serious. I do believe that it has had its effect on the participants. I think the anonymity rather than the competition is one of the effects. There are so many people now you can hardly remember who they all are. An individual can hardly make his voice heard, and he is just one in a large mass of people all contributing. We have certainly lost something by that.
I think this is very much a part of the history of nuclear physics and the fact that has led into new research techniques which have in fact affected the style of research and the life style of the men who do the work.
And this is particularly true for the high-energy physics where the apparatus is extremely large. Not only the machine which produces the particles, but the observing, detecting equipment is extremely large, cumbersome. You need a big team. You often have in Physical Review Letters more names than text of the communication. The names fill half the page, the communication a quarter page. I think it must give much less satisfaction to be in this kind of work, although I see some of my colleagues who seem to be quite happy and who are engaged in this kind of work.
But it implies that the motivation of a person going into physics today would have to be somewhat different than the motivation of a person going in a decade or two earlier.
I imagine so.
Because if you went in with the same motivations as a couple of decades earlier, you might be disappointed because of the limitation of the opportunities.
Yes, particularly as an experimenter. As a theorist you don't need to work in a big team. Theoretical papers are usually still one, two or three authors. However, their competitiveness comes in a different way. There are so many people working on the same subject that I suppose in nine cases out of ten the theorist will start some work and get three- quarters done and then somebody else publishes t, which was quite a rare occurrence even in the '30s and even rarer before.
The interesting thing about this—this is just personal opinion— is that instead of getting at the roots of this, people seem to be trying to institutionalize it by getting the information to people that much faster—so that he can get seven-eighths done before someone else publishes it.
Well, I don't know. Maybe t will reduce the waste motion. I am rather in favor of getting the information there more quickly in a live field. It may reduce the waste motion somewhat.
But in an informal sense some of the proposals today are for institutionalizing these informal roads of communication, preprint exchange, and so forth. I want to lead from this point, before we get into the post-war narrative, to another point about people coming into the field. We touched on it yesterday, about the quantum theorists, the people who had transformed quantum theory. Many of them then became interested then in the physics of the nucleus. I'm curious to know a bit about that. It appears that some of them entered the field and then withdrew at a rather early stage in the '305. Did we cover that yesterday?
You mean people like Wigner ...?
Yes, and Heisenberg.
And you wanted to know why they did this?
Yes. Was it just because the equations of quantum mechanics were available and this was a logical way to apply them?
It was like Mount Everest. It was there to be climbed. Well, you might say that the most interesting problems in atomic physics were exhausted by 1930 or '31 or so. With the exception of solid state physics I think the principles had been discovered and a lot of details had to be worked out. The subject was mostly taken over by the theoretical chemists, by the physical chemists. There were a few physicists who stayed with the subject fortunately, like Hartree and Slater, Mulliken and a few more. But it mostly went to the chemists. Now, so there were two directions which seemed very interesting and proved very interesting. 0ne was nuclear physics, and the other was the extension of quantum theory to higher energy. And by that I mean in these early days in the 1930s quantum electrodynamics, the divergences which appeared in quantum electrodynamics, how to resolve them and so on. Quite a number of physicists stayed and continued primarily in this subject of trying to make sense of high-energy quantum theory. And these included, for instance, Pauli, Dirac, who looked for the cure of the troubles in classical electrodynamics—I think wrongly, but he did— a number of Pauli's students: Stuckelberg. And in addition a number of other people shared their attention between high-energy quantum theory and nuclear physics. For instance, Weisskopf did a lot about high-energy quantum theory in the late '30s when he also worked a lot on nuclear physics. Bohr himself also shared his attention between these fields. Wigner did a little bit of high-energy quantum physics and a lot of nuclear physics. I did very little high-energy quantum theory, and so there was a gradation going, as I explained. Oppenheimer, for instance, I think was all the time more interested in the high-energy quantum theory than in nuclear physics, although he had some interest in nuclear physics. This was late '30s.
Where does the cosmic ray thing fit in?
The cosmic ray thing fits in with both in a way. The immediate cosmic ray results were on high-energy quantum physics. Cosmic rays gave us showers in the atmosphere, showers of electrons and gamma rays, showers in solid material, which could be explained by quantum electrodynamics without solving the difficulties of divergence. In fact, most of it could be solved by extending the theory of Heitler and myself on the fundamental processes. And this was one of the main interests of Oppenheimer and his school. It didn't touch and didn't solve the question of the high-energy limit, but Oppenheimer's school did quite a lot of investigating this and tried to get evidence from cosmic rays on the high-energy limit and some quite interesting results were obtained; for instance, that the penetrating cosmic rays can have only spin 0 or 1/2 but not spin 1 or greater. So these two subjects went hand in hand, not very closely related to each other. Then I already mentioned that in the conferences immediately after the war both of these subjects were discussed quite intensively and the first great progress after the war was in high-energy quantum physics—Lamb shift, Feynman, Schwinger and all that.
From '47 through '48 and '49.
Yes, and going on in, I think, already late '48, Dyson proved that electrodynamics really could be made finite in every order of perturbation theory and that the renormalization theory was in a way complete, and it could be carried out in every order. He didn't carry it out, but he proved the existence, which was a tremendously important paper.
But in general he acted as the interpreter of what others in the field were doing, didn't he?
I don't think so. This was a very original paper. He used the methods of Feynman. But Feynman had only proved that these methods would work in third order of perturbation theory, and Dyson now proved that it worked in any order—91st if you wished.
So he was more than interpreting ...
Then afterwards he did a lot of interpreting, which was very good, too. But this really was probably his best paper, and t was a magnificent paper.
You've helped trace some long-range trends, and I think that is good because it summarizes a lot of the work from 1930 to 1950.
Yes. And I would like to go just a little bit further on. This high-energy field theory, of course, was not finished with Dyson's paper, as some people thought it might be; but people began to worry even if it converges in every order of perturbation theory, do these various orders converge? That is, when you take all the terms of third order, fourth order, fifth order, do they form a convergent theory? And they came to the conclusion that this is probably not so. As a consequence, field theory has stayed very much alive. It has been at times applied to strong interactions, the interaction of mesons with nucleons; at times to electromagnetic; very rarely to weak interactions, the beta decay interaction. But there is a large and increasing group of people who are working, and have made very important discoveries, in this field theory. The field theory has become more and more complicated and in the hands of many people more and more abstract. But in the last two or three years it has found its way back to applications to strong interactions in high energy. And, for instance, one of my colleagues here, Kinoshita, has used field theory to derive many very interesting results on high-energy phenomena. So this is one branch which has pretty much separated from the rest of physics, has loose connections with high- energy physics but only very loose ones, and has gone off on its own. I don't know anything about it since about 1949.
Is that a consequence not only of the fact that you had other interests but that the way the field has developed one has to be totally immersed in it in order to know something about it?
Yes, exactly. It is characteristic of this field in particular but also to some extent of high-energy theory, that you have to stay with it very completely full time, and it has become extremely difficult for graduate students to learn enough during their graduate study so that they can write a thesis on a subject in these two fields. This is one of the reasons for the very long times of graduate study that we are now seeing. You just have to learn an awful lot, and very frequently when they do learn an awful lot, it's very narrow. It is limited to that field. You can't get there without going through the embryonic stages of the kind of physics I know, but the people who then work in these fields are very specialized and very often don't know some of the very elementary things which seem obvious to us of the older generation some years ago.
The complexity and the amounts to be learned suggests, at least to an optimist, that a simplifying principle may be around the corner somewhere.
There is some hope for that just at this moment. One of my colleagues who keeps track of this tells me that there seem to be really very promising calculations which seem to put nearly all the field of weak interactions and a large part of strong interactions into a form that is believable and may even be correct and that gives correct results on a large number of problems. And I think it is quite likely that within the next couple of years this simplifying principle will be fully developed and we will have something here similar to the quantum mechanics of 1925- '26. The simplifying principle is known as current algebra and was chiefly developed by Gell-Mann, although a dozen other people contributed a great deal.
With that glimpse into the future, we'll try to get you now back to the past. Let me ask you at what point did you feel that we had interrupted the coherent narrative or account of the over-all development of the field? Did you feel that we took it successfully to the end of the war?
I think so.
And then I think you even pursued it a bit further with the beginnings of high energy. So now it’s a question perhaps of dealing with nuclear physics and with some specific points. Do you think a good point would be to start on the shell model and then the optical model?
Yes. The shell model was invented, as far as I know, in '49— that is, in its modern form. And it was found simultaneously, as far as the records go, by Mrs. Mayer in Chicago and by Jensen and Suess in Hamburg. I think it was Hamburg. It may have been Heidelberg, but I'm not sure. Now, these people followed a lot of other people who made unsuccessful attempts, and these people for the first time put in the right forces acting on individual particles which explained the shells in agreement with those observed. And the key point was that they postulated strong spin orbit forces, LS forces as they are also called, and they did this entirely empirically. That is, they said, "If we assume such forces of the correct sign, then we get shells in agreement with the observed ones." After the shell containing 20 nucleons, which is closed with 20 nucleons of each kind, which is closed at calcium, there is one shell at 28, one at 50 and at 82 and at 126—very strange numbers. And they were able to explain all these strange numbers on the basis of spin orbit interaction. They had to assume that spin and orbital momentum liked to set themselves parallel, which is contrary to the observations for electrons in atoms, where they like to set themselves anti-parallel. Now, when this was announced, the successes were immediately obvious. They were immediately obvious in spite of the fact that at that time experimental material was not very good. These people had to argue from such things as the abundance of elements. They found that elements with closed shells—so-called magic numbers—are more abundant than their neighbors. The occurrence of certain isotopes—they found that tin, which has 50 protons, has more isotopes than any of its neighbors; that there are lots of nuclei which have 82 neutrons, which nobody had noticed before, and such things. Nowadays we have much better evidence than these people had because we now know the exact masses, and hence binding energies of all nuclei, to an accuracy of better than a million volts, and so we can observe that these binding energies are very much greater for the magic nuclei than for their neighbors. The inventors also found other evidence—namely, they found evidence for high angular momentum of nucleons in certain cases. In some cases this was directly deduced from the spin of nuclei. They looked at the spins of various nuclei and they found that large spins occurred just in those regions where the shell model predicted them. These large spins occur only in nuclei with odd mass number. All nuclei with even mass number have zero total spin, so they are not useful for the purpose. So they had a lot of very good evidence which is very well summarized in the book by Mayer and Jensen on the Elementary Theory of the Shell Model, and t is really elementary. Many books call themselves elementary when they are not, but this one really is elementary and gives a very good survey of the evidence which was available at an early time for the shell model. So the evidence was quite overwhelming and became more and more so as time went on. For instance, at one time—I don't remember when; maybe '53 or so—I proposed a certain test for the shell model, hoping that it would come out wrong. I was hoping that the strict idea of nucleon shells was not quite right, but that one had to allow some leeway. Well, when the experiment was done, it came out completely in favor of the shell model. So the evidence for it was overwhelming, but there was no force known in nuclear physics which would explain the shell model, which would explain the spin-orbit interaction. The forces which we had previously discovered were the central forces and the tensor forces and no spin- orbit force. And if you assumed that the spin-orbit force had a similar cause as t has in atoms, you got a result which was hopelessly too small by a factor of 30 or 50 or 60—I don't know, certainly a result which didn't come near the expected magnitude. There were some attempts to explain the observed spin orbit forces in terms of the already proven tensor forces, and many of us wanted to do this very much. The person who did it successfully was Wigner as usual. That is, he didn't do it himself, but his student, Feingold, did this calculation. It didn't really work. He could explain some spin- orbit force, and in a few cases this indirect spin-orbit force may be the main contributor, but it obviously didn't work in general. Well, this situation remained for quite a long time until people were able to polarize protons, to direct the spin of the proton. I forget exactly at what time this was done; I think it was first done successfully at Rochester, much inspired by Marshak. I forget who did it experimentally—perhaps 0xley. Well, in any case, he found at 150 MeV or whatever energy Rochester's cyclotron gave, that strong polarization effects were obtainable. People had looked for such effects at lower energy and found nothing, and we know today that indeed there isn't anything. No appreciable polarization is obtained up to quite high energy, up to almost 100 MeV. Furthermore, we believed all sorts of stupid things. We thought that the best way to get polarization is to scatter protons from protons; and it turned out afterwards that the best way is to scatter protons from heavier nuclei like carbon, which have zero resultant spin. This is a complicated matter. It has a somewhat complicated explanation, and when it was first reported at one of the Rochester conferences, it puzzled everybody. But such was the fact. I forget the exact dates on this, but it was in the early '50s that this happened. After the Rochester work, the Berkeley group got busy on this, and they had the advantage of higher energy, higher intensity, many more people; and so they soon got a large number of results—many more results than Rochester had been able to get. In order to detect polarization, you have to scatter your protons twice—first to polarize them and then afterwards to analyze the polarization. The problem was theoretically investigated by Wolfenstein, who is now at Carnegie Tech, and a couple of other people. And they predicted that in order to get full information one had to do triple scattering experiments. One had to scatter the protons three times. Each time you lose about a factor of 10,000 in intensity. So you can imagine where you end up. Well, you can be clever about it and lose only a factor of a few hundred, but still these experiments are exceedingly difficult; and the first successful ones were done by the Segrè, Chamberlain and collaborators at Berkeley. I think they were published only in '56, but I think they started about '54. When these experiments were done, they were immediately evaluated by some very competent people, and the evaluations showed that there was a very strong spin orbit force—precisely the thing that was needed for the shell model. And only after this was found—and I think this was really established only in '56—only then did I fully believe in the shell model. Once you have a spin-orbit force between the elementary particles, then of course it's easy to get the spin-orbit force between a big nucleus and one nucleon.
This conversion, that is, your end of resistance to the shell model—does this imply that you had to give up something else? Or during this period were you content to be on neutral ground?
I was on neutral ground. I just didn't do any calculations involving the shell model. That is, if I had been working in nuclear physics, I would have considered t very foolhardy to use the shell model, whereas other people used t to their hearts' content. As a matter of fact, I didn't work in nuclear physics. But this would have been my attitude. So this was a very important result in nuclear forces which came from medium energy accelerators, a hundred to four hundred MeV. Now I suppose we better trace the shell model from here on. I think there is nobody who nowadays still doubts the shell model, at least no serious nuclear physicist. To establish the shell model, you needed still another thing, and this will finally lead me into my present work. You needed to understand how t is possible to have orbitals of individual nucleons, how they can have any reality, how can you get away from the complete statistical model that the compound nucleus of Bohr represented. The main step in this direction was taken by Brueckner in 1954—it may have been '55—who applied to nuclear physics the methods which had been developed by Watson to determine multiple scattering. Brueckner had as collaborators some quite able people, including Levinson and Eden—Eden who is now the chief theoretical physicist in Cambridge, England. And between them they developed a very nice theory, which is known as the theory of the many-body problem. They had some trouble proving their theory. It seemed to work all right. It seemed to be plausible, but it had certain very definite flaws. The real proof of the theory was given in '56 by Goldstone, who worked as my research student in Cambridge when I was on a year's sabbatical leave in Cambridge.
What year was this?
That was '55 to '6. I had been intrigued by the Brueckner theory; it seemed to me right; but it seemed not sufficiently proved. First I tried my own hand at it. I think I clarified a few points, but I couldn't really prove it. But then I had a very clever graduate student— Mr. Goldstone—who has in the meantime become quite well known as a high energy physicist, and he could prove t. In fact, it took him about a month or two to do so. And Goldstone's proof is now generally accepted. Now, this work is the basis of two things. First of all, it explains how each nucleon can have an individual existence, an individual orbital, in spite of the fact that there are tremendously strong forces at short distance, then tremendous attraction, amounting to probably several hundred millions of electron volts, many many times the binding energy of the nucleon in the nucleus. In spite of these simply fantastic forces, it is logical and justified to ascribe to each nucleon a separate orbital, a separate set of quantum numbers. And this is, I think, the most important direct consequences of the Brueckner-Goldstone theory. Brueckner himself was very much aware of this and really set out to solve this problem. Secondly, and somewhat independently of the first, this theory makes t possible to calculate the energy of nucleons in the nucleus, to calculate the binding energy of the nucleus. There were many disappointments in this calculation, and there were many further extensions of the theory before we got the right answer. The latest extension I made in '64, in which I included more accurate treatment of the interaction of three nucleons at the same time. They interact by pairs but they can be closely correlated. This was necessary before we could get the right binding energy. Well, many attempts were made to get the right binding energy. Brueckner and Gammel made lots of them in the early days. We made some here. The final calculation, which we believe to be right, was done by Sprung and a student of his at McMaster University and by Dalman here at Cornell. We believe that now we have the right numbers and we get the right binding energy and we get the right density of the nucleus. From here we are going on. Well, what we have done so far mainly concerns so-called nuclear matter. That means the behavior of an infinite nucleus, and we have now gone on to discuss the surface of the nucleus. And other people, particularly Jerry Brown and his collaborators, have treated finite nuclei by the same method, and there is every reason to hope at this point that we are on the right track and that this theory, which was started by Brueckner, when properly applied, can be used to calculate the properties of the low-lying energy levels of any nucleus. So in this sense perhaps, coming back to something you asked during lunch, we are coming to a resolution of the problems of nuclear structure. I wouldn't say that nuclear structure theory is at an end. It's in a way at the beginning. But when it has a beginning with a good theory, then an end is in sight; and t is likely this now corresponds to 1926 in atomic theory and in four or five years most of the interesting problems in nuclear structure will have been solved, at least for the low- lying levels. It's not certain, but there's at least some expectation that this may be so.
That's interesting because it coincides with your feeling about the impending resolution of certain problems in high-energy physics, and it raises a question about whether in their final form the resolution of the problems in nuclear structure on the one hand and in particle physics on the other will have something in common, will have an interaction.
So far as I can judge at the present, no.
Has there been such an interaction in the past? In other words, with the development of the ideas of particle theory since, say, the late '40s—has this affected the work in nuclear structure?
It has affected it in giving us some leads on the nuclear forces. When nowadays we analyze the medium-energy experiments on the scattering of neutrons and protons by protons, we put into the analysis what knowledge we have about meson-nucleon interaction, and this has proved very fruitful; and, in fact, without this we would probably be in considerable trouble. This was introduced by Moravscik I think in '58 or thereabouts and has since been used by Breit very extensively and by everybody who works on the subject, including my student, Reed. So we are helped in the analysis of experiments by knowing one part of the theory of nuclear forces, and that's the part which is given by the pi meson. There has been a second interaction—namely, high energy physics has discovered a lot of other mesons. These other mesons have been used to tell more about the nuclear forces at shorter range. And if you are not insisting on very high accuracy, then you can analyze the nuclear force as t comes from the experiments in terms of interaction with three or four different mesons. This is useful; it is satisfying in that it connects the two fields; it gives a basis for nuclear forces; t is useful in giving us a little more insight into the shape of nuclear forces. It is, however, not anything like Coulomb's Law where you can just write down a formula. Furthermore, it has only limited precision, and the most important force cannot be described in this manner. The most important force is that between particles which have zero relative angular momentum, and this is the most important when you want to talk about binding energy and nuclear structure, and this force cannot at all be derived from meson exchange up to the present day. But the others can be, and so that has been useful. Then in the opposite direction, knowledge of nuclear physics has been very important for analyzing certain of the high-energy experiments, which involve nuclei; and methods have been taken over. And here I am reminded particularly of Mr. Mehra's question about the SU 3 theory. The SU 3 theory was found useful by Elliott in discussing the consequences of the shell model for the binding energy of successive nuclei and for the structure of the low-lying energy levels, the angular momenta and so on. And much later this SU 3 theory was taken over into high-energy physics. The reasons for the applicability is totally different. They have nothing whatever to do with each other. But it just happens that the same kind of mathematics fits the two fields. Well, it’s similar to many previous cases in physics—wave equations happen in acoustics and in electromagnetic theory, in quantum theory. The phenomena are totally unrelated and yet the same mathematics applies.
I think that's a good answer to his question. And it brings you up to the present. And it gets you to another question that we had in mind. That is, what holds the nucleus together?
Yes, precisely. That of course is very closely related to two things I mentioned. First of all, everybody by now is convinced that exchanges of mesons of various types are the foces which hold the nucleus together. Pi mesons, single pi meson exchange, is certainly important. Simultaneous exchange of two pi mesons is certainly very important. And these two are probably the most important ingredients. The other mesons—rho and omega in particular—seem to be important for holding the nucleus apart. They seem to cause very strong repulsive forces at short distance, and these repulsive forces are absolutely necessary to prevent the nucleus from collapsing. The question of collapse of the nucleus was discussed yesterday. I mentioned the exchange force of those days, and the exchange force of those days has to some extent been placed on the pile of dead wood. People still use the expression. They don't visualize t at all in the way in which it was originally conceived. There is exchange all right, namely, exchange of mesons; and indeed the exchange of one pi meson makes something like an exchange force but not quite. It makes a force which also leads to saturation of nuclear forces that would prevent collapse of nuclei. But it is not simply an exchange force. It's more complicated. It's a force which depends on the scalar product of the two spins and of the two isospins of the interacting nucleons. The exchange of two pi mesons simultaneously, which is extremely important for the binding of nuclei, is not an exchange force but an ordinary force, a Wigner force. It is strongly attractive in all states, and if t and the pion exchange were the only forces acting, then nuclei would collapse. So we needed additional exchanges, particularly the rho and the omega mesons, in order to prevent collapse of the nuclei. And so in this way I think we know in a qualitative way what holds the nucleus together—namely, meson exchange; what hold the nucleus apart— namely, exchange of heavier mesons of mass about 800 million volts instead of 140 million volts. And we know, from the calculations of nuclear matter which I mentioned before—the Brueckner-Goldstone calculations followed by my own—in which way this happens; that is, in which way these forces manage to give the correct amount of glue to hold the nucleus together.
This compared to the 35-year-old prediction of Yukawa of mesons being involved in holding the nucleus together is an end in the sense of a cycle or the beginning of a new one, if you choose to put t that way. It appears not to be simplifying things, though, but to be making them more complicated.
They got more complicated because meson physics turned out to be enormously more complicated than Yukawa ever dreamed.
Was this a gradual recognition that t was this? Or did it only come after 1947 when, in fact, the different type of meson was discovered?
It only came through experiment. And I think a typical and very apt comment was a song by Arthur Roberts which was I guess '53 or thereabouts.
He was originally at MIT. He was then at Rochester. He is now at 0regon. When he did this he was at Rochester. This was where he said, "What shall we do with 22 mesons? Some people don't know when to stop." He was lamenting, as everybody was, the proliferation of meson discoveries. "Some mesons might do us a good turn by turning to glue, but what would you do with 22?" And that was the story at that time. Most of the mesons which were known at that time other than the pi and mu are now known as strange particles, and many of them turned out to be the same. But others have taken their place. The rho and omega were discovered quite late—I think in the '60s—and there are now ordering schemes to put them in some kind of order, and this has to do with the SU 3, which I mentioned previously. There just are an awful lot of particles which strongly interact with nucleons. Each of them has an interaction. Hence, the force which they make between nucleons is just exceedingly complicated. But you may then say: fundamentally what we have to look at are the relations between these various kinds of mesons, the SU 3 group which relates all of them to each other, and now perhaps current algebras which may be again the basis of SU 3 or certainly are closely related to that; so that we may be on the verge of seeing the simplicity behind the complication—the simplicity being current algebra, which is a very abstract and somewhat complicated subject, but which is fundamentally simple, perhaps of the same degree of simplicity as quantum mechanics when we really learn to understand it. Then from the current algebras follows the structure of this whole zoo of different mesons and different states of the nucleon. And from this again follow the interactions between nucleons, and from those again follows the general structure of nuclei and the specific structure of specific nuclei and specific energy levels. So there is a long chain, a chain very similar to the chain from quantum mechanics to orbitals in an atom to molecular binding to the color of a given dye organic molecule.
In describing this chain, you have not summarized the history (not the quantum physics, but the nuclear physics) you have enunciated a program for the future—a step-by-step program.
Yes, right. And the only step which we have I think accomplished fairly well is the step from nuclear forces to the structure of very heavy nuclei.
You left out the optical model, if I'm not mistaken, and I'd like to get back to that ...
And also the collective model.
Good. And there may be others. But I'd like, before we do that, just to take this to another point which seems logical to me— and that is, now that you've summed up the field and you've talked about the program for the future, what about the dead wood that we started to mention; about ideas that essentially haven't led anywhere or have been discarded or temporarily been put on the shelf or in the waste- basket, going all the way back, if you can?
I think there isn't terribly much. Everything more or less found its place. Well, going far back, you may talk about the prehistoric ideas: Bohr's idea of violation of energy conservation in beta decay—that was simply wrong; Konopinski and Uhlenbeck's theory modifying the Fermi theory of beta decay—that was simply wrong; some experimental results on beta decay, which led to the assumption of an interaction by scalar and tensor—they were simply wrong. The experiments were wrong. And a lot of theoretical papers were written in that connection. They are still useful because the theorists mostly considered all five different kinds of interactions, which also include the presently accepted vector and axial vector, and in addition include pseudo scalar. Well, many of the early attempts to get quantitative understanding of the energies of live nuclei—some of those which are in my Part A— are pretty dead wood because the interactions then used were wrong, and the interactions are much more complicated and so one cannot use these approaches. Some of the methods were quite primitive. I think most of the work that has been done has led in a more or less logical way to the next approach rather than being just dead wood. I would say, for instance, the early calculations on the shell model which I mentioned—Feenberg's —were very useful because they gave general insight into how to do such calculations. And even specifically I think they are useful for certain problems. Most of the work that has been done with the shell model is still correct and not dead. Some work has been done with insufficient mathematics, but in those cases it was usually clear from the beginning that this was not terribly good and that it had to be done better. Only at the time, nobody knew how. Quite a lot of the work on so-called direct interactions falls into this category, and I'll discuss that in connection with the optical model. I think on the whole there isn't terribly much dead wood. If you go through Physical Review of the late 1930s, of course the experiments have been superseded by better experiments very largely, but they fulfilled their function at the time. The theories were maybe primitive. For instance, the theory of penetration through the potential barrier; the theory of the compound nucleus are still all right. They have to be slightly modified, but I certainly would not call them dead.
They're useful in a certain context, with certain kinds of problems.
Yes. think there has been remarkably little dead wood.
Maybe "dead wood" is the wrong way to put it but rather to think of transformations in thought, changes. Would you agree that this is characterized by the choice of a particular model in a particular period— that this is the best symptom we have of a transformation of a way of looking at things?
It serves to organize all of the others.
Yes. Certainly we have become much more sophisticated. As I said repeatedly, the compound model, which was the guiding principle of the late '30s, still is useful in describing nuclear reactions at high energy for the most part. It has to be supplemented. The shell model was already talked about; the liquid drop model still is useful for many purposes; and I'll come to that when I talk about the collective model, which is sort of its successor.
Do you think now would be the proper time to get into the optical model and the collective model?
All right. They are of very different type. I'll take the optical model first because Mr. Mehra asked about it. Logically, I could do one or the other. The optical model either way is an outgrowth of the shell model. In the early 1950s, as I mentioned before very briefly, Barschall and his collaborators at Wisconsin did a lot of careful experiments on the scattering of neutrons of a few hundred kilo- volts as a function of atomic number, and he went through the entire periodic table. He found very striking phenomena. Some nucleus might have a large cross section at zero energy and then a small one at 100 kilovolts. Another nucleus would have the reverse phenomenon. So there were clearly some properties of neutron scattering by nuclei which had a scale—let me call it this way—of a few hundred kilovolts or a million volts in contrast to the resonances which had been discovered earlier by Fermi in 1935 or thereabouts, which had a scale of single electron volts; so here was a phenomenon which had a scale a hundred thousand times bigger. Obviously one explanation which was good for one thing couldn't be good for both, and for a long time people were very puzzled by these experiments. Then Weisskopf and his collaborators had the courage to take the shell model seriously for a neutron which is not in the nucleus but comes into the nucleus from the outside. And they said that such a neutron could have an individual existence in the nucleus, an individual orbital, just like the bound nucleons. This was a very revolutionary thought. It went back, as it were, to my paper of 1935. And when t was properly used by these people, they could explain all these fluctuations—fluctuations of the cross- section with energy and with atomic number. These are gross fluctuations of large energy scale which are caused by resonances in the one particle wave function. If the neutron wave function just fits into the nuclear potential, if you have an integral number of half waves in the nuclear potential, then you get a resonance and a large cross-section. Well, Weisskopf and collaborators also connected these large- scale resonances with the small-scale resonances, which had previously been obtained and which are resonances of the compound nucleus as a whole, not of one particle but of a hundred particles together. And they could show that the small resonances could be superposed upon the large resonances; that for some nucleus—let's say nickel at low energy—there is generally a high cross section because of the one particle resonance. And that means that neutron scattering by nickel at low energy just is very high and larger than for some other nuclei. 0n the other hand, there can in addition be resonances—and indeed there are—and these are then wider and stronger than they would be for a nucleus which does not have in addition a one particle resonance. This general picture was justified pretty well by the Brueckner school theory of nuclear matter.
Excuse me for a moment. When did Weisskopf and his collaborators advance this?
I can look it up.
So can I at another time. Just an approximate date. I don't want to nail it down.
Maybe '52. The model was justified by the Brueckner work. And, on the other hand, to apply the model, one had to calculate the wave functions of individual nucleons in a nuclear potential. This is what is meant by the optical model—that you say that an individual nucleon, when in the nucleus, is subject to a potential energy which is strongly attractive inside the nucleus and then gradually goes to zero. It was shown that it's important to make t go to zero gradually rather than suddenly, and this optical model was then used by lots of people to describe the scattering and reactions of neutrons and protons and also somewhat more complicated nuclei. In this optical model also, the spin orbit coupling was introduced of course. It had to be introduced by logic, t was introduced; and it explained the results of experiments on the polarization of neutrons when scattered—neutrons of some 10 MeV, 5 MeV, 20 MeV, when scattered by nuclei. This is different from the subject I have discussed before— namely, polarization at 100 MeV and more. I said before there is no polarization at low energy. That was correct—when you talk about the scattering of neutrons by protons or protons by protons. But in the scattering by more complex nuclei, there is polarization which can be explained by the optical model.
How widely used was it and for how long a period of time?
It is still used and people are still busy connecting the optical model to more rigorous theories of nuclear reactions, and only in the last couple of years have people been able to derive really good parameters for the optical model, like the depth of the potential and spin orbit coupling and so on. This work has been done particularly well at the Los Alamos Laboratory, at the Oak Ridge National Laboratory and at Berkeley.
Do you think that covers that and allows us to get on to the collective model?
Yes, except for one more sentence. I think I want to make clear that the optical model is an extension of the shell model into the continuous spectrum, into the region of free particles. Now the collective model. I think the first inkling of the collective model was conceived by Rainwater from considerations about the shell model. His work I think has not received enough publicity, and it was somewhat tentative, but I think t contains the main part of the idea. The main person to develop the collective model was young Bohr, Aage Bohr, together with Mottelson, an American, and somewhat independent and parallel early developments were by Wheeler, Ford and Hill. For the idea of the collective model, let me start from A. Bohr's ideas. He pointed out that the surface energy of the nucleus is not very great, so you don't need much additional energy to deform the surface. It's only the surface energy which keeps the nucleus of spherical shape. In a way the collective model starts from the question, "What gives a nucleus its shape?" In the case of an atom it's obvious. There's a nucleus in the atom, and this nucleus will hold the electrons and more or less automatically they will tend to form spherical shells. In the nucleus there is no nucleus and instead there are only short-range forces between the nucleons. How then does the nucleus manage to have a shape? Well, the same way—and here it really matters—as a liquid drop. In a liquid drop you also have only short-range forces between the molecules. Still the liquid drop has normally a nearly spherical shape. On the other hand, it's easily deformed. When it’s a falling water drop, it can easily be deformed into a pear-like shape. So relatively weak forces will deform the shape of the liquid drop, and so Bohr told himself the same thing should happen to a nucleus. On the other hand, t is well known that it is very hard to compress water. Its volume doesn't change even under quite strong forces. And so he said, "The same thing happens in the nucleus. Its volume is pretty definite, but its shape is not so definite." So, using this picture, he then examined the consequences of deformations, and the simplest deformation of the sphere is into an ellipsoid of rotational symmetry—make one axis bigger than the other two or, one axis smaller than the other two or you could make all three different. Well, this proved to be a very fruitful concept, and he could guess or calculate the energy levels of a spherical nucleus and the energy levels of an ellipsoidal nucleus. In particular for the ellipsoidal nucleus it is clear that you change matters by rotating that nucleus. So he said, "There ought to be rotation of this nucleus much the same as for a diatomic molecule, and this rotation will have energy levels much the same as in a diatomic molecule where you can show that the first three excited energy levels have excitation energies in the ratio 1 to 3 to 6." I think that's right. So A. Bohr investigated the low-lying energy levels of nuclei and found that indeed about half or at least more than a third of all nuclei had energy levels which followed precisely that sequence. Moreover, these energy levels lay very low, and that's what you should expect because a big nucleus has a large moment of inertia, and the energy of rotational levels is inversely proportional to the moment of inertia, and so you should expect very low levels. Now, these levels then are obtained by the rotation of the nucleus as a whole, and therefore are a collective state and not just the state of one nucleon. This seems to be entirely separate from the shell model. He went further, and he said that even nuclei which are normally spherical ought to be able to have a deformation to an ellipsoid, and in this case this is a vibration, and vibrational energy levels are uniformly spaced so that the three lowest levels should be in the ratio 1 to 2 to 3. And he could show that most of the nuclei have either the ratio 1 to 2 to 3 or the ratio 1 to 3 to 6 for the low-energy levels.
You talked of A. Bohr, but you also talked of Mottelson, but in your description everything was related to A. Bohr.
Yes. Well, it was all joint work of Bohr and Mottelson. I think I got the ratio of the energy levels wrong. It's not 1 to 3 to 6, but it's something like that, with the spacing much larger for the higher level. This seemed a very different approach from the shell model, but very quickly after the first papers, A. Bohr and Mottelson themselves and lots of other people—Weisskopf was of course involved in this, Peierls, Brown, many other people—tried to establish the connection between the shell model and the collective model, and at the same time to give a reason why some nuclei are deformed and ellipsoidal and other nuclei are spherical. And what turned out was approximately this— and here Elliott's work comes in very importantly: When you put lots of nucleons into one shell, then because of their mutual attraction, they try to arrange themselves so as to be close together. And the best way of arranging them close together is by putting them all into two poles of spheres, so to speak, so that then at least the nucleons outside the last closed shell are not spherically arranged but are asymmetric and make something which looks like an ellipsoid. And then you can go on and take the next step and say, "Now that those nucleons are there at the poles, the other nucleons feel more attraction to the poles and so this is a self-aggravating effect and you get the ellipsoid very nicely established." To start this off, you have to have enough nucleons in an incomplete shell to give enough mutual attraction, and therefore you will get this ellipsoidal shape whenever you are not near a magic number, not near a closed shell. And that is exactly what is found. The model was made more quantitative by Nilsson working under A. Bohr's direction, who said, "Let's start right away from an ellipsoidal arrangement. Let's assume that the nucleons are subjected to an ellipsoidal potential and calculate their energy levels," and by doing this he gave very nice quantitative results—just how the deformation should take place, under what conditions, what should be the quantum numbers of the nucleons in this case and how much should the deformation be. Many people have worked on this, and there has been a complete reconciliation and amalgamation of the shell model and the collective model in this manner.
The collective model before Mottelson's idea was advanced from about 1953 on—there was a joint paper of A. Bohr, Mottelson and Nilsson— and then this reconciliation between that and the shell model, about how long did that take?
I think as soon as Nilsson's paper appeared on the individual nucleon quantum states in an ellipsoidal potential, it was clear that reconciliation was possible. And while people are still working on it— it's still a very active field—there are quite a number of papers and a book on the unified theory of nuclear models, a book which is particularly illuminating by Brown whom I mentioned several times, a very small little book. He, I think, perhaps more than anybody else has been devoted to this unification and has shown how everything flows from the same fundamental principles. In fact, coming back to this long chain of going from current algebras to real nuclei, what Brown has been interested in is the end parts of the chain, and he has been very successful in this— from the theory of nuclear matter to the theory of finite nuclei, generally, to the shell model, to the collective model and the combination of all of these.
Your last statement about him sort of covers all of the models that we have covered. Is there anything else that you think is pertinent to add on the models and their relationships or the change of interest from one to the other over a period of time?
I think they all were developed simultaneously. There were certain people primarily interested in the collective model. I mentioned them already. There were other people—and I will particularly mention the Israeli physicists, Talmi and Veshalik (?) and Racah, who exploited the shell model and obtained energy levels from the shell model using the principles of group theory which had been used successfully on atomic physics, and it's much more complicated in nuclei. What they did is essentially an extension of the Wigner work of the late 1930s. So there was that group working on extensions of the shell model. There are of course hundreds of experimenters. There are lots of people interpreting experiments, sometimes using one, sometimes the other model, sometimes both. Elliott whom we mentioned frequently, was one of the people successfully unifying the two models. Flowers should be mentioned— he's at Manchester—as one who has done a great deal about shell model calculations, detailed calculations, fixing parameters connecting with the nuclear force and so on. So in this postwar period I think it was not a succession of different models but different people were each adopting their favorite model until they finally came together. But even now the Copenhagen school always starts from the collective approach, and the Israeli school always starts from the shell model approach. They recognize that the other model has a lot to say, but their main methods are always based on one of the models.
That brings me to what might be a final question on this interview, if you feel we've really covered everything, and that is ... We started off by talking about the various centers of research or of interest really in this field. We talked about the early '30s. We didn't say anything really directly to that point later on in the later period—the '30s, the '40s, the '50s and the present. Is it possible for you to think about where these major centers of interest in nuclear structure have been from the '30s on?
Well, at present I should say that in theory the United States is not necessarily the leading country. There is quite a lot of nuclear theory in this country, and there are quite a number of competent people; but I would say that perhaps many of the most original ideas have come from elsewhere—especially England, Denmark and Israel. There have been very good contributions from France. There's a very good group at Saclay. There are contributions from all over, including many countries I haven't mentioned; but these I think would be the centers of ideas.
Today. 0n the other hand, experimental work is going on everywhere and particularly in the United States. The equipment for nuclear structure physics is cheap on modern standards, very expensive on the standards of the '30s. You get a good machine, a good Van de Graaff generator, which is the best equipment for most investigations, for $2- or $3 million, and a very fancy one for twice that amount, which is of course nothing compared to high-energy physics. 0n the other hand, when we built our cyclotron here in l934 to '6, the total amount of money allotted to it was $2- or $3000, and a lot of time and sweat by Stanley Livingston and a number of graduate assistants. So, on present-day scale, it is relatively cheap and therefore can be had in many different places. It is a favorite subject for experimental research in the universities. We don't happen to have one, but many universities have a Van de Graaff, a good one, and have good experimental research going on. I think there must be 20 or 30 in the country.
Would you say that today the study of experimental nuclear structure or the study of nuclear structure on an experimental basis is comparable to the amount of interest and the wide diffusion of studies on nuclear spectroscopy in the '30s?
Well, it is still nuclear spectroscopy to a large extent, only that people have become much more sophisticated. They investigate not only what levels there are but the exact energy—fantastic accuracies are obtainable now which were not in the '30s—and how the levels behave, how they disintegrate, their angular momenta, all the decay schemes and so on. So there's a lot of spectroscopy nowadays, too, and it is much more widely diffused now than it was in the '30s. In the '30s I think there may have been half a dozen places participating. Now I think it's about 30. And of course it is only through such work and a lot of such work that one gets the interesting facts, and only on the basis of such work that for instance A. Bohr could say, "Now, look, half the nuclei have rotational energy levels; half the nuclei have vibrational energy levels," and this correlated enormous numbers of experimental papers.
And this is a general indication of the close tie between the theoretical aspects of nuclear structure with the experimental continually, whereas in other fields there may be that the theoretical and abstract mathematics may tend to develop into a separate field very remote from experimental.
I think that is quite correct. I think nuclear physics is one of the fields where they go at approximately the same pace, and lots of experiments are suggested by the theoretical developments and vice versa. In high energy physics the experiments have run far ahead of the theory; and on the other hand, take gravitation: You cannot do any experiments but you can do a lot of theory.
Getting back to the question on the centers, now we have both ends. We have the beginning and we have the end. What happened then in the '30s and in the '40s? What were the centers then? You haven't mentioned the Soviet Union, for example. Did they figure in this at all?
I know very little about this and other people know much more. I have talked to one and maybe to a couple of Soviet theorists. They were disappointed by the work of their own experimental people. Whether that's justified, I don't know. I think nuclear experimentation is much less widespread in Russia than here. It is not a glamour subject, as high-energy physics is, and they have relatively restricted manpower, and so they put more emphasis I think on high-energy physics relatively speaking than on nuclear physics. 0n the other hand, other countries— the western European countries with the exception of CERN—have gone on the whole the other way. They have aid, "Nuclear physics we can afford. So there are quite a number of good nuclear physics machines in England, France, Germany, Denmark, Sweden—you name it: Holland, Switzerland, Israel, whereas high-energy machines are very few. Italy, I think, has stuck pretty much to high-energy physics and has one of the first high- energy machines in Europe designed by Robert Wilson of Cornell. [Pause in recording]
Could you suggest certain basic review articles or books that in their own way cover important periods in the development of nuclear structure? One would be the three review articles you did in 1936 and '37. Then what would you think would be the next thing? Or was there something earlier than that that would be a key document?
Well, for the prehistory, Rutherford, Chadwick and Ellis, the book on radioactivity. The next I mentioned was Blatt and Weisskopf, and I think there isn't terribly much in between that's comprehensive. And the next comprehensive book is that by Preston, which is I think four years old or something like that. It is not nearly as deep as the Blatt and Weisskopf book. Blatt and Weisskopf tried, as I did, to give you a complete theory with a complete proof for everything they stated. They wouldn't leave anything unsaid, whereas Preston very frequently says, "Well, now this is kind of complicated," and the reader can find the proof somewhere else.
But if you just as an experiment took the bibliographies of the Rutherford, Chadwick, Ellis and the Bethe Bible and Blatt and Weisskopf as a total bibliography up until 1952, you might have at least a representative selection of some of the most important literature in the field. This is what I'm getting at there.
And then from those, we can select certain ...
Preston is extremely good also as a reference book.
For that he would be good?
He would be good. Let me say it's not as complete as my review articles or Blatt and Weisskopf. It's much more a textbook for graduate students, so he gives maybe 15, 20 articles on each chapter, of something like 15 or 20 chapters. So the references are not as complete as they were before, but still they give you a very general idea of the most important contributions.
You gave a talk in the spring sometime before the New York State section of the Physical Society, and someone told me that you had something to say about the history of nuclear physics. Is that so?
Well, no. I was telling a simple story about nuclear physics, not about the history.
I see. Well, maybe concepts of history is related to anecdotes in the story, so maybe that was the impression. Maybe it was a misimpression I got. I would like in the remaining time—and I know we won't to able to finish in the remaining time—to at least open up some of the autobiographical questions that we deliberately deferred today. And I thought that since we don't have time to do it completely, we could take ones that related to your moving about—the fact of leaving Germany, at least for a visit, to study for a few years, your reactions to the different communities you found and the style of this community, and then to continue on with the return to Germany and the final leaving and coming to England again, and then your expectations, your problems in relocating from there.
All right. In 1930 I got a fellowship from I think it was called the International Education Board, which was Rockefeller money.
How did you get that? Did you seek it?
Sommerfeld, who was my teacher, sought t for me. think that was the approved method. He recommended me and I got it. I wasn't actually terribly anxious to go abroad. I did t more out of a sense of duty. But I soon found that England was a very wonderful place. Well, you must know that the Germany of the '20s was already a somewhat difficult country. There was a great deal of dissatisfaction which finally erupted in the Nazi government. There was already in 1930 a very strong Nazi movement. And quite apart from the Nazis, of whom I knew very little at that time except that they were generally objectionable, the general atmosphere in academic circles was very bad. I found that nearly all my colleagues, nearly all my professors, were terrible chauvinists talking of nothing else than restoring the glory of Germany and of the unfair treatment that Germany had received in Versailles. I didn't feel that way—I think largely influenced by my father, who was an old-style liberal from way back. In fact, in the early years after the revolution he was one of the few professors at the University of Frankfurt who did not deplore the change of fortune but was ready to make the best of it. He gave quite a number of election speeches supporting the Democratic candidates. The Democratic party was the party of the left and nonsocialist. He was almost alone among his colleagues in this. So I found life in Germany in the 1920s much less than satisfactory in every respect except in work. I found that I couldn't talk politics to anybody, and this colored the outlook generally. It was generally a very narrow outlook. People were grumbling all the time. They were dissatisfied. I had a very small circle of friends—two or three—with whom I got along very well and with whom I went on walks in the mountains from Munich, where I mostly stayed. But apart from that, I just felt that I didn't fit into the surroundings. Now, coming to England, I found that people were not all that way, and this was a great revelation to me. I found that people in England, although they talked English instead of German, spoke my language and that it was easy to get along with them; that they felt much more open; they had much more liberal ideas; you had sensible discussions about politics and about life in general. So it really was an awakening in England. I found that some of the old people sitting around the fire in the Cambridge colleges after dinner were just delightful and were telling stories and were doing new things. One told us that he was learning to fly at the age of 60. This was so completely different from the German scene. And so I felt much more at home in England than I did in Germany. In England I spent a little over four months. I didn't do terribly much work. I did some work. But I mainly learned the language and enjoyed the people and definitely did not enjoy the food. Then I went to Italy where I didn't learn the language. As we mentioned yesterday, I spoke German to Fermi and to most of the people in the lab and English to Mrs. Fermi in their home and also occasionally to the other people; and instead I did a lot of physics.
How long did you stay there?
Again about five months in '31 and then four more months in 32. It was so interesting ... In Italy I mainly concentrated on work. I found delightful people to discuss physics with. I found that Fermi was a man from whom I could learn a tremendous amount. He gave me a lot of my style, and I think he was a second scientific father to me, although he wasn't much older than I. But he had a lightness in his approach and a directness which had been missing in Sommerfeld, and so I asked that I be permitted to return there the next year and the Rockefeller Foundation granted me an extension of my fellowship.
And you stayed right on. You didn't return to Germany.
I returned to Germany in between. I returned in June of '31 when also it got very hot in Italy and people went on vacation and I stayed then in Munich for the fall of that year and returned I think in February for another four months. So during this time I learned a great deal. During the second visit I also started writing my article on one and two electron problems, which was published in the Handbuch der Physik in 1933 and which was then in '53 republished with Salpeter, and also published as a separate book translated into English. Then I returned to Germany in June or so of '32, spent a couple of months in Munich and then got an appointment. It was called beauftragte Lehrkraft. That means somebody charged with giving the instruction in the field of theoretical physics. I was not a professor. I was more than privat-dozent.
Was this a customary appointment?
This is an appointment that is normally made in remote fields— that is, let's say, the Burmese language, where you don't want to appoint a professor.
An adjunct professorship type of thing?
You might call it that, but it doesn't have the privileges of a professor. In this case it was made because the previous professor had resigned and had gone to the United States, and they couldn't make up their mind whom to appoint as a new professor, and so they picked me up from the street.
Who was the previous professor and why did he go to the United States?
Landé. Maybe he saw things coming—I don't know. Landé is a man who contributed a great deal to quantum theory before 1926 and not terribly much after. He went to the University of Ohio, and I don't know what made him go. So I taught there during the winter semester, which lasts from November to early March, and as far as the work went t was very satisfactory. Personally, it was most unsatisfactory because Tübingen was one of the most Nazi-infested towns of southern Germany and I was quite lonely and the only relief I had was that t was very close to Stuttgart where I had very good friends—Ewald, whose daughter I later married. Then at the end of January '33, the Nazis came to power. On the 1st of April they had the first boycott of Jewish stores, and about the same time they published a law according to which anybody who had one Jewish grandmother could not hold any official appointment. Well, since I had a Jewish mother and not only a grandmother, it was clear that this meant me and that sooner or later I would have to leave. I didn't expect t would be quite so soon. I had the somewhat interesting experience that one day in April I got a letter from one of the two people who were taking their doctor's degree with me saying that he had read in a small-town paper in Württemberg that I had been dismissed—what should he do? This was the first news I had. Then I wrote a letter to the professor of experimental physics who had been very friendly to me and really had indicated that he liked my work and liked me to be there and so on, and I got back a very stiff letter that presumably the lectures in theoretical physics would have to be arranged differently the next term.
No other hint in there?
No other hint. And about a week after the first letter from that student, I finally got a letter from the Minister of Education of the state of Württemberg saying that I was dismissed according to the law effective the first of May and the salary for May could still be paid to me and that was that.
How much before May 1st was this?
0h, maybe the 15th of April. I should say that there are the university vacations in March and April in the German universities. So thereupon I went back to Munich to my old teacher, Sommerfeld, who got without much difficulty a small fellowship for me for the next three or four months. Back in Munich I did mostly research. I think I did not give a course of lectures but I'm not sure. Maybe I did. Otherwise it was quite pleasant. It was quite a good time for work. I think it was the time when I directed my first Ph.D. thesis success- fully on electron scattering, which was I think quite good. And in the meantime. Sommerfeld, who was a very kind person, was looking for positions for several of his former students, including me. After a relatively short time— I think it was maybe sometime in June—a position was found for me in Manchester, where Sir Lawrence Bragg was the professor. So I went back to England.
Was this period in Germany—the several months when they informed you they no longer wanted your services—a depressing time for you? You said otherwise t was quite pleasant. How did you manage to separate this feeling of knowing you were being forced to leave the country ...?
Well, I took this an impersonal act of mad people and didn't consider this as directed against my person. The people in the Institute were generally very friendly. There was one other person in the Institute, who was an American actually, with rather leftist leanings... He and I used to tell each other the latest jokes. We were then at one time warned by the Diener, who was a man of all trades who did everything in the lab except for the professorial duties—he did the machine work and he was the janitor at the same time; he kept everything straight; he knew everybody and he made models of crystals which he sold for his own account; he was a very versatile man—warned us not to do this because there were ears in all the walls, and I think this was a very good warning. Sommerfeld was very good and unchanged and so were nearly all of the people who worked there.
Your friend Peierls has indicated in an interview with Tom Kuhn that at Cambridge he saw an advertisement concerning a job opening and he sent the information to you suggesting that you should apply. Do you remember that?
I remember that. I don't remember where it was.
Well, I assumed that it was ...
It was not Manchester.
I know you were both offered jobs by Bragg in Manchester. He indicated that in England at the time there was local unemployment and so there was a problem of absorbing people, but nevertheless you did get a job there.
Yes. The English really were extremely good and really went out of their way when they themselves didn't have too prosperous a time. They looked for jobs for as many as they could absorb. They created new jobs. They created new fellowships. They had the Emergency Committee for the refugees. Szilard played a great role in working there and helping them and by the creation of additional jobs. They, for instance, got one job for Fritz London, of superconductivity fame, and later on for Schrödinger and for Heitler and at least a dozen of the refugee scientists.
Not all of these scientists were leaving because of the so- called racial laws but some of them because they were political.
Mostly. I found very hospitable conditions in England. They received me as one of their own. I could work very well. I did a lot of work. I discussed some yesterday on solid state inspired by Bragg, some on nuclear physics with Peierls, and the work with Heitler and a few occasional things in addition. I stayed for one year in Manchester. In Manchester I had the good fortune to have a regular position as a lecturer. The normal lecturer was away on leave. Then this expired at the end of the year and I got a fellowship in Bristol, which was very nice, too. The professor there was an old friend of mine of the 1930 days—namely, Nevill Mott. Heitler was there and we continued to work somewhat on similar problems as before. But even before I went to Bristol I got the offer from Cornell, which had come through Lloyd Smith, who was then a young assistant professor here at Cornell and who had worked with me in Munich in '32 or so during the time between my Fermi visits.
And had you thought of coming to the United States prior to that?
Not consciously. If somebody would have asked me, I probably would have given an open-minded answer, but I had not definitely considered this as the place I was likely to go. In fact, prior to that I thought I would stay in England.
I think you explained yesterday that they couldn't offer you a permanent position.
Were you aware of physics in the United States, of what was going on? Did you have any real idea of that?
Not really. I knew of the cyclotron being built and that seemed very interesting—in Berkeley. I had met Arthur Compton and had some quite good talk with him on physics.
No, during a conference in London. But on the whole I thought that the center of physics was in England and not in the United States.
A final question. Were you at the 1934 International Conference on Theoretical Physics in London?
Yes, that's the one.
I'd like to talk about that next time, too.
I don't know whether I remember about t. I just saw in my bibliography that I gave three papers to that conference, which surprises me very much in retrospect.
There are many people who are listed, whose papers are included in the collection of the conference but who never attended.
Yes, but I was there.
Well, next time I'd like to pick it up from this point and about the circumstances of arriving in the United States—how you made contact with a new physics community, how you were received by the administration of the institution as well as by your colleagues, and then how you made contact. You mentioned Weisskopf, you mentioned coming down to New York. You seemed to be in the swim of things very early, and I would like to trace that.