History Home | Book Catalog | International Catalog of Sources | Visual Archives | Contact Us

Oral History Transcript — Dr. Masatoshi Koshiba

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.

Access form   |   Project support   |   How to cite   |   Print this page


See the catalog record for this interview and search for other interviews in our collection


Interview with Dr. Masatoshi Koshiba
By David DeVorkin
At the Kamiokande Facility in Kamioka
August 30, 1997

Oral history interviewee photo
open tab View abstract

Masatoshi Koshiba; August 30, 1997

ABSTRACT: This interview is part of a small program to document the recent history of the American Astronomical Society. These interviews were used as background studies to help authors of chapters for the centennial history volume of the Society research and organize documentary materials. The volume to be published in 1999.

Transcript

DeVorkin:

I am with Professor Masatoshi Koshiba, Professor Emeritus — is that correct? — of physics, the University of Tokyo. And the auspices of the interview are the American Institute of Physics. Please state your name.

Koshiba:

My name is Koshiba, Masatoshi Koshiba. Just call me Toshi.

DeVorkin:

Toshi. Okay. Toshi, could you tell me something about your family, who your father and mother were, their names, something of their background, the kind of family you came from.

Koshiba:

Alright. My father was a professional Imperial army officer who fought in China during the last World War. And my mother was named Hayako. She died when I was 3 years old. My father married my mother's elder sister, who became my second mother.

DeVorkin:

And your father's full name?

Koshiba:

Toshio. Because my father was an army officer, I was told to enter the military school during the war. Luckily or unluckily, one month before the entrance examination I got polio, which made my right arm numb. It's still numb.

DeVorkin:

But you have full control of it.

Koshiba:

Well, I don't have any strength. That made me exempt from military service during the war.

DeVorkin:

Tell me, when were you born and where?

Koshiba:

I was born in 1926 in a city near Nagoya, Toyohashi City.

DeVorkin:

Okay. Now your father was a professional military officer. He was stationed there?

Koshiba:

Mm-hmm [affirmative]. That was the time when he was stationed in a regiment in Toyohashi City.

DeVorkin:

Okay. Do you have brothers and sisters?

Koshiba:

I have one elder sister, and from the second mother I have two brothers, younger brothers.

DeVorkin:

So you are the oldest son.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Could you tell me what your siblings are doing, or what their lives were? Your older sister, did she —?

Koshiba:

She married an accountant. They are living in the southern part of Yokosuka City.

DeVorkin:

Okay. And your two younger brothers?

Koshiba:

My next youngest brother is living in the western suburb of Tokyo, and is rather well-to-do financially. He is the president of a company.

DeVorkin:

President of a company?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Does he have training?

Koshiba:

He graduated university, and majored in finance and economics. He then married the daughter of the company owner, and sort of inherited that company.

DeVorkin:

Okay. And your second brother?

Koshiba:

The youngest one has been a teacher of the Japanese language in high school. He is retired now.

DeVorkin:

Okay. Did your father have university training?

Koshiba:

No, he went to military cadet school.

DeVorkin:

Okay. So tell me a little bit about your childhood and your activities. What were you interested in as a child? Did you have hobbies?

Koshiba:

Well, after I got polio —

DeVorkin:

How old were you when you —?

Koshiba:

I was 13. I was in the first year of middle school. Before that I was rather good at Japanese fencing, Kendo.

DeVorkin:

Kendo, yes.

Koshiba:

But after polio I couldn't do anything athletic, so my hobby at the time was to build model airplanes.

DeVorkin:

Oh really? Okay.

Koshiba:

I was seriously thinking of becoming a model airplane shop owner, selling model airplanes.

DeVorkin:

Did you join a club? Did you have friends who did the same thing?

Koshiba:

Yes.

DeVorkin:

And did you build them from kits, or did you just build them from scratch?

Koshiba:

From scratch.

DeVorkin:

Were they models that were representative of specific aircraft, or did you make them to fly?

Koshiba:

I made them to fly.

DeVorkin:

Of any particular aircraft?

Koshiba:

No. I didn't aim at any specific type of airplane. When I made a flying model, I made it very simple so that it flew a longer time.

DeVorkin:

So you were going for performance.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

What was the mode of power? Was it a gas engine or —?

Koshiba:

A rubber band.

DeVorkin:

Rubber band. Okay.

Koshiba:

Simplest type.

DeVorkin:

Balsa wood? Using balsa wood?

Koshiba:

In those days balsa was not easy to obtain, because it was not produced in this country.

DeVorkin:

So what materials did you use?

Koshiba:

Bamboo.

DeVorkin:

I made some model airplanes too, and we would sometimes make the skeletal structure and then cover it with paper and then dope the paper. Did you do the same thing?

Koshiba:

Yes, I did the same thing.

DeVorkin:

I'd like to know more about your interests then. Was it in aircraft or aviation? What kinds of books did you read, what kinds of things were you interested in?

Koshiba:

The Japanese educational system followed the old German educational system. And after six years of primary school or elementary school, we had five years of what is called middle school. After that, we had three years of what is called higher school.

DeVorkin:

Higher school.

Koshiba:

There were a number of higher schools, although in Japan the one called the first higher school in Tokyo was supposed to be the best school. And so we wanted to enter this first higher school. I liked mathematics, but when I was hospitalized because of polio, I also had diptheria.

DeVorkin:

Diptheria, yes.

Koshiba:

Yes. I was hospitalized for about half a year. And my teacher of this middle school happened to be a mathematics teacher.

DeVorkin:

Was this school in Toyohashi?

Koshiba:

No, in Yokosuka. You see I spent only one year in Toyohashi after birth. And because my father was sent to Manchuria, we were left behind in Yokosuka.

DeVorkin:

So that's where you had your early training.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

So you had a teacher who was a mathematician.

Koshiba:

Mm-hmm [affirmative]. He was very good to me. He is still alive in Yokosuka. While I was in hospital, he gave me a couple of books to read. One of them was a small book which was a dialogue with Albert Einstein.

DeVorkin:

The World As I See It or —?

Koshiba:

I haven't seen the English version of it, and I don't remember the name, but a dialogue takes place with Albert Einstein about how physics was created.

DeVorkin:

Oh, could this have been Einstein and Infeld?

Koshiba:

Yes, that's it! It was translated into Japanese. Even though I didn't understand the simple statements in this book, they did attract my attention. In those days I didn't have any notion of becoming a physicist. Further, because of polio, I didn't have to worry about military drafting. However, I didn't do very well in the entrance exam and I failed.

DeVorkin:

This was an entrance exam for the higher school?

Koshiba:

Yes.

DeVorkin:

Okay.

Koshiba:

The second time I tried I got into the first higher school.

DeVorkin:

Now this would be in Tokyo.

Koshiba:

Mm-hmm [affirmative]. And everybody had to live together in the dormitory on the campus, and that was a very good thing, because science majors or literature majors, everybody had to mix.

DeVorkin:

Let me ask how could your family afford to send you away to school? Was your father relatively well-to-do, or was there money available?

Koshiba:

Well, by that time my father was an army colonel. He sent money for us.

DeVorkin:

So he was quite a high-ranking person.

Koshiba:

Yes.

DeVorkin:

When did you go to the higher school?

Koshiba:

That was in 1945, four months before the end of the World War.

DeVorkin:

Oh. Now Tokyo was certainly —

Koshiba:

Yes, devastated.

DeVorkin:

Devastated. So, were you actually in Tokyo during this time?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

You were not evacuated.

Koshiba:

Our school was on the southwest side of Tokyo, and the majority of the bombing was down in the central and eastern part of the city. Even though the area around our school was pretty much burned down.

DeVorkin:

This must have been a very, very difficult time for all of you.

Koshiba:

Yes. What else can I tell you?

DeVorkin:

Well, what was the instruction like in the higher school? Were there teachers that are memorable to you or important to you in the development of your interests?

Koshiba:

Oh yes.

DeVorkin:

You also mentioned that everybody had to live together in a dormitory, and so people interested in literature were put in with other arts people and that sort of thing. I would like to know if you were concentrating in any particular area at that time yourself.

Koshiba:

Well, in those days I was quite interested in literature. As a matter of fact, I was seriously thinking of entering the German literature department in the university. As you know, in those days, because boys of different fields of interest lived together for three years, we made many good friendships. For instance, among my friends I count not only physicists and scientists, but also a very well known poet, the president of the Bank of Japan, and many others in different fields.

DeVorkin:

How big was your class? How many people?

Koshiba:

Well, altogether this first higher school had about one thousand students.

DeVorkin:

Oh, one thousand students. And you all lived together.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

But this was, as we'd say in the United States, the cream of the crop; this was the highest ranking high school in Japan.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Okay.

Koshiba:

There were three physics teachers in this higher school, and they all gave us lectures and training. One physics professor gave me a very bad mark. The other two gave me good marks. When I was in the second year of this higher school, I was elected to be the vice president of the student body. This is a very responsible appointment. Even during the war we believed in self-government. So the student body elected the president and vice president, and organized a cabinet. This cabinet controled everything — daily life and everything in the dormitory. When I was in that capacity, I couldn't pay attention to the courses, so my grades went down considerably. When I was graduated from this school I was just about in the middle. So, all the students in the science field wanted to enter the physics department, because that was the department that was very attractive to science major students.

DeVorkin:

When did you become interested or decide to become a science student?

Koshiba:

This I am going to tell you.

DeVorkin:

Okay.

Koshiba:

At that time I was more interested in German literature. You know Japanese people like hot baths.

DeVorkin:

Hot baths. Yes.

Koshiba:

Because everything was burned down around the school, our friend in the student government went out and found a transformer which was still usable, got it to the campus, and used this transformer to boil water to make a big hot bath. So everybody around the area came to take a hot bath. One cold night, I was taking a bath; because the window was broken, cold wind came in and because it was very hot water there was a lot of steam. You couldn't see people 2 meters away. And I could hear the voice of this particular professor who gave me the flunking grade.

DeVorkin:

A flunking grade you said?

Koshiba:

Well, the very low grade.

DeVorkin:

The lowest grade, yes, okay.

Koshiba:

He was answering his favorite student. His favorite student, of course, was aiming at the physics department.

DeVorkin:

Who was that? Do you remember who it was?

Koshiba:

I know. I remember the name, but I won't tell you the name. You'll find out the reason why later. The professor asked, "Young man, which department do you think Koshiba will apply for?" This colleague of mine, a student replied, "Oh, I don't know. He is interested in German literature." And the professor said, "Well, whatever he chooses to apply for, it cannot be in the physics department. Either German literature or Indian philosophy." That's what this professor said. That statement made me furious, so I started studying physics. After one full month of concentrated work, I passed the physics department requirement, while the favorite student of the professor failed.

DeVorkin:

Really? Oh, boy.

Koshiba:

Yes. That's why I'm not going to tell you the name.

DeVorkin:

Fair enough. I would be interested to know, though, of those who were among your classmates who did go on in science, in astronomy and physics. Are there any names that immediately come to mind?

Koshiba:

Kozai.

DeVorkin:

Kozai did.

Koshiba:

Yes. There were other well-established scientists from my school. Let me see. One of my colleagues became a professor of I think meteorology at Columbia many years ago, but I don't know where he is now.

DeVorkin:

Okay. No, I meant just the people that you knew directly.

Koshiba:

That was a few years after the war ended, and my father was retained in China for one year, and then was released and came back. However, because of his career he was banned from any public service. So I had to ask a friend who is rather high up in society, actually this man was from the same higher school. Simply because you are a graduate of this school, you get some sort of credit or status.

DeVorkin:

Credit or status?

Koshiba:

Yes. You see, the graduate of this higher school goes to the University of Tokyo, to various ministries, and then becomes a high ranking officer. Those people, when they need somebody, seek out a younger person who is also from the same higher school. They try to be nice to graduates of the same school.

DeVorkin:

We call such a school an alma mater.

Koshiba:

Oh yes, that's it, alma mater.

DeVorkin:

It's the feeling of comradeship.

Koshiba:

That's right. This senior person kindly offered a job to my father which was not a very respectable appointment. However, my father and mother and younger brothers had to live. But, still there wasn't enough money, so while I was attending the university I did all sorts of odd jobs.

DeVorkin:

So you went to the University of Tokyo.

Koshiba:

Yes. Uh-huh. Like for instance tutoring rich people's children, or sometimes I went to Yokohama to unload the —

DeVorkin:

Oh, to work on the docks unloading ships?

Koshiba:

That was from 8 o'clock at night to 8 o'clock in the morning.

DeVorkin:

Boy.

Koshiba:

[laughs] Yes. I did all sorts of odd jobs to help my family.

DeVorkin:

Did your brothers also work, or they were younger.

Koshiba:

They were too young to work.

DeVorkin:

Much younger. Okay.

Koshiba:

But my elder sister did help, because she was rather good in making dresses. So she also helped the family.

DeVorkin:

Okay. Now you took the physics exams and so you took the physics courses. Did this sustain you in physics? I mean you had made your mind up for physics, or were there still other possibilities?

Koshiba:

Well, as I told you, because I was in financial straights, because I had to take on various odd jobs, I couldn't attend my courses. Therefore, when I was nearing graduation, the records were very, very poor. In Japanese grading, if the top level is called excellent, then good, then in English you would probably you call them fair.

DeVorkin:

Unsatisfactory or something.

Koshiba:

Fair. Almost all the grades I obtained were fair. I had only two excellents. That put me almost at the end of 29 students in the physics department. Therefore, asking for a scholarship in the graduate school or assistantship was out of the question for me. I didn't know what to do. However, luckily, I was vice president of the student body.

DeVorkin:

Well, this was back in high school.

Koshiba:

Back in high school. At that time, the principal of the first higher school was a well-known philosopher called Professor Amano. Professor Amano after being the principal of the first higher school, was appointed the minister of education. And this Professor Amano, while I was working as vice president of the student body, was very, very nice to me. Further, Professor Amano was also a graduate of the higher school. After graduation, he went to Kyoto University rather than Tokyo University, because he wanted to study under Professor Tomonaga, a well-known philosopher at the University of Kyoto, who also happens to be the father of Professor Tomonaga, the physicist. Okay?

DeVorkin:

Oh, I see. Okay.

Koshiba:

And Professor Tomonaga, the philosopher, was a go-between for Professor Amano's marriage. And Professor Amano was a go-between for Professor Shinichiro Tomonaga's marriage. Shinichiro Tomonaga is a physicist.

DeVorkin:

Interesting connections. Okay.

Koshiba:

Yes. And then when I was asked by Professor Armano which department I was going to go into, I told him that luckily I could go into the physics department. Professor Armano said, "I know a physicist; I don't know how good he is in physics, however he is the son of my professor and I made the arrangements for his marriage. He is living in Tokyo, so I'll give you a letter of introduction." So with this letter I went to Professor Tomonaga's place, and that was the beginning of my acquaintance with Professor Tomonaga. And as I said yesterday, I'm very lucky so many important people have been so nice to me. And Professor Tomonaga was very, very nice to me. At the end of my university days he asked me what I wanted to do. I said well even though I learned only the beginning of physics, physics seems to be interesting. However, because of the odd jobs and so forth, my grades were not good enough to continue graduate work. That was the time when Professor Yukawa, while he was a professor at Columbia at that time, met Professor Marshak of Rochester at a New York meeting of the APS. In those days, all good American students, especially in physics, went to Harvard, MIT, Cal Tech, Berkeley, or Princeton. Rochester couldn't get any good students. Therefore, Marshak wanted to get a good student from outside the country. He tried first in India, and then he worked on Yukawa to send a Japanese graduate student to Rochester. However, in order to get this scholarship at Rochester, you have to get a recommendation from a well known physicist like Yukawa or Tomonaga. I was very lucky since I already knew Tomonaga for some time. I hesitatingly asked him if he would write a recommendation letter for me. He said, smiling, "Alright. You write the recommendation yourself, I will make some corrections and so forth, and then I will sign it." So, I wrote that, even though as the attached record shows this man does not have very good grades, I have known him for some years, and he is not completely stupid. That kind of thing. Then he signed it. Fortunately, I was taken on as a research assistant by Rochester.

DeVorkin:

Oh, so you had research money or support that way.

Koshiba:

To support myself.

DeVorkin:

Right.

Koshiba:

Which was $120 a month. However, in those days with the official exchange rate, a professor at the University of Tokyo earned something like $70 a month. So I thought I would become very rich going to Rochester. However, I didn't know that the living expenses were much higher — $120 minus 10 percent deduction left $108 a month which was not quite enough to live on, even for a single graduate student.

DeVorkin:

Before you get too far there, I would like to know, did you actually take courses from Tomonaga?

Koshiba:

No. Not at all. I did attend some seminars he organized.

DeVorkin:

What kind of physics were you getting interested in while you were at the University of Tokyo?

Koshiba:

You know, when I entered graduate school I was taken up by a theoretical physicist, a very well known theoretical physicist, Professor Yamanouchi.

DeVorkin:

At Rochester?

Koshiba:

No, in the University of Tokyo.

DeVorkin:

Oh, okay.

Koshiba:

I spent two years in the graduate school of the University of Tokyo. One of my friends who was a research assistant of the department, who later became a professor of physics at Waseda University, advised me that competition in theoretical physics was very severe. He said, "Why don't you consider experimental physics?" He then described a new type of nuclear emulsion, which was invented by a Bristol group in England. We obtained a small sample of this Ilford nuclear nuclear emulsion, and then started a small experiment by exposing it on top of Mt. Fuji.

DeVorkin:

On top of Mt. Fuji. So you hiked up to Mt. Fuji for this.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

What were your facilities like for physics, your physics labs and—?

Koshiba:

Nothing.

DeVorkin:

It was really very, very meager?

Koshiba:

You know, it was already three or four years, or maybe five years after the war.

DeVorkin:

So this was about 1949, 1950?

Koshiba:

Mm-hmm [affirmative]. For instance, there was only one sodium iodide crystal, about this big.

DeVorkin:

About an inch.

Koshiba:

Yes. Uh-huh. That was held in a very heavy safety box, and a professor was responsible for it. Graduate students couldn't touch it. And, as you know, all the accelerator-related equipment had to be destroyed. When I went to the University of Rochester graduate school, they first gave me an oral exam by four professors in order to decide which courses I should take. Because of this lack of training in experimental techniques at the University of Tokyo, all Japanese students who came to Rochester were required to take a laboratory course.

DeVorkin:

A laboratory course, uh-huh. Did other Japanese students come with you at the same time or around the same time?

Koshiba:

Yes. Two more came in the same year.

DeVorkin:

Who were they?

Koshiba:

One was a Professor Ohnuma, an accelerator physicist. Ohnuma worked many years at Fermilab, and he is a professor at The University of Texas in Houston now. The other one was a theoretical physicist, Professor Takahashi, who is now a professor at the University of Edmonton in Canada.

DeVorkin:

So they both stayed in the West.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Okay. But as far as laboratories, research equipment, that sort of thing, it was very, very meager after the war. There was no immediate reconstruction, no money.

Koshiba:

No technical training to speak of. So when I went to the University of Rochester and was given this one inch sodium iodide crystal to clean, it was really a shock — that a first year graduate student would be permitted to handle so precious a thing. [laughs]

DeVorkin:

I'm sorry for belaboring questions about the University of Tokyo, but I do want to know as much as I can about what conditions were there not only for physics but for astronomy. And I was just wondering one more question. Did you have any contact with anybody interested in astronomy at that time?

Koshiba:

The only contact was through Kozai.

DeVorkin:

So you remained in contact with him.

Koshiba:

Mmm [affirmative], from time to time. And there was Hatanaka, who died some years ago.

DeVorkin:

Okay. Yes. Hirayama was still around, and Hagiwara —

Koshiba:

Hagiwara is still alive, yes.

DeVorkin:

And Hagihara. You didn't have any contact with them at the Tokyo Astronomical Observatory.

Koshiba:

No. After all, it was only four or five years ago that I was elected a member of the Astronomical Society of Japan.

DeVorkin:

Okay. But I'm still curious if you knew any of these names: K. Sotome? Okay, he isn't known to you. Was Professor Hantaro Nagaoka around?

Koshiba:

Before my time.

DeVorkin:

Okay. So he predates you and you had no contact with him.

Koshiba:

We go back to Rochester again?

DeVorkin:

Go to Rochester, right. I'm interested in the route you probably traveled, oh, to Hawaii and then to the west coast and then across to Rochester. Is that how you got there?

Koshiba:

Wait a minute. In those days we could afford only a boat trip.

DeVorkin:

So how did you go?

Koshiba:

From Yokohama to Seattle.

DeVorkin:

Okay.

Koshiba:

Eleven days. And, when I found out that $102 a month was not enough to live on, and that I would get a minimum guaranteed salary of $400 a month as soon as I got my degree, I decided to get my doctorate as soon as possible. I arrived in Rochester in the month of August.

DeVorkin:

Was that 1950?

Koshiba:

'53.

DeVorkin:

'53. That's right. You had two years of graduate school. Okay.

Koshiba:

And then in the month of October, I had to take a German exam. In those days the University of Rochester required two foreign languages.

DeVorkin:

Sure.

Koshiba:

Yes? Besides English, there was nobody who could test my Japanese as a foreign language. So I had to take the German exam in October, and the French exam in November. Then in January I had to take what are called the qualifying exams, which last almost a week.

DeVorkin:

Wow.

Koshiba:

Luckily, I passed all these, and started thesis work.

DeVorkin:

So it was very rapid.

Koshiba:

Yes. Because I had to get the $400 guaranteed income as soon as possible.

DeVorkin:

But it sounds like you were very well qualified and well prepared from the physics that you had at the University of Tokyo, because you passed the qualifying exams within just four or five months of getting there.

Koshiba:

Well, our teachers at the University of Tokyo were really good physicists.

DeVorkin:

Had you had English training?

Koshiba:

Just a little. School English. So when I went to Rochester it was very, very difficult. Each time I met a new person, I had to ask, "Please speak slowly and clearly."

DeVorkin:

Did you live in a dormitory?

Koshiba:

No. I shared an apartment with an Australian graduate student.

DeVorkin:

Who was that and how did that get arranged?

Koshiba:

His name was Jack Noon. He later become a professor at the Renselear Polytechnic Institute.

DeVorkin:

Renselear, yes.

Koshiba:

The reason was because we were both students of Professor Kaplon, initials M. F.

DeVorkin:

So you didn't work directly for Marshak.

Koshiba:

No. Marshak is a theoretical physicist, and I was applying for experimental physics. Kaplon was well-known in those days for nuclear emulsion work.

DeVorkin:

You mentioned that you did a little bit of Ilford nuclear emulsion work on Mt. Fuji.

Koshiba:

Yes.

DeVorkin:

But I didn't ask you if you got any scientific results from it.

Koshiba:

No, not to speak of.

DeVorkin:

But did you develop the emulsions, and scan them with a microscope?

Koshiba:

Yes. In those days I had written one theoretical paper, because at the beginning I was to become a theoretical physicist.

DeVorkin:

At Rochester, you decided to be an experimental physicist.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Now was this more to your liking, or something they suggested that you do?

Koshiba:

No, it was my wish to become the student of Mort Kaplon.

DeVorkin:

His first name was Morton? And he was working in nuclear emulsions?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Was he flying them or sending them down under lakes?

Koshiba:

No. He was flying them on balloons.

DeVorkin:

On balloons. Right. Was this your first contact with ballooning?

Koshiba:

That's right.

DeVorkin:

I'd be very interested to know how your work developed and what your thesis was.

Koshiba:

For my thesis topic I had chosen ultra high energy phenomena in cosmic radiation. The reason I had chosen that title for my thesis work was because Marcel Schein of the University of Chicago published what he called an anti-proton event in cosmic radiation. His analysis seemed fishy to me. So I thought that I would make my own study to clear things up. My conclusion was that nothing extraordinary was occurring in high energy cosmic radiation.

DeVorkin:

This was from your own observations.

Koshiba:

Mm-hmm [affirmative]. I got the doctorate degree in the month of April, which means it took a year and eight months at Rochester for me to get a doctorate.

DeVorkin:

April 1954 then.

Koshiba:

No, '55. A year and eight months.

DeVorkin:

Okay. That is very rapid. Did you actually prepare nuclear emulsion stacks and fly them with Kaplon?

Koshiba:

No, the flown stack was given to me to analyze.

DeVorkin:

So you did not yet have experience flying stacks.

Koshiba:

No direct contact with the balloons. At the beginning of July, I was offered a research associateship by Marcel Schein of the University of Chicago, where I could enjoy a $400 a month salary from the month of April.

DeVorkin:

Now your thesis had shown that something he had suggested was not correct, and —

Koshiba:

Probably Marcel liked it. [laughs]

DeVorkin:

He liked it. I see. Now, at Rochester you already said that you were given a sodium iodide crystal to clean, and so clearly there was more of an infrastructure for experimental research. Was this exciting to you?

Koshiba:

Oh, yes. For instance, I was given the assignment to design an FM receiver and an FM amplifier, which is completely beyond my capability. And the circuit I made was ringing and making noise. [laughs]

DeVorkin:

Why were you asked to do this?

Koshiba:

Well, it was a common assignment, one of the assignments given to a graduate student.

DeVorkin:

And this would be an FM amplifier that would fly on a balloon for some reason?

Koshiba:

Well, I don't know the reason, but you are given some parts and so forth, and then you have to make the FM receiver and FM amplifier work in the 100 megahertz region, which is not easy.

DeVorkin:

You had plenty of laboratory equipment? There was no need to scrounge? You had everything you needed, as far as you could tell?

Koshiba:

Well, I don't remember very much about this laboratory course.

DeVorkin:

Oh, this was part of a course. Is that right?

Koshiba:

Well, I was exempt from taking courses. The committee at the time of my entrance to the graduate school at Rochester, decided that I had already sufficient knowledge to take the qualifying exam.

DeVorkin:

Oh, I see. So this was —

Koshiba:

The laboratory course was the only exception.

DeVorkin:

Right, okay. Now during that time did the Rochester conference take place while you were there?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

And what was that all about?

Koshiba:

Five months after I arrived in Rochester, there was a Rochester conference.

DeVorkin:

So that would be January of '54.

Koshiba:

Mm-hmm [affirmative]. And that was the time when I first met Dick Feynman. Dick Feynman was back from his trip to Japan, and he was considerably interested in the country. Being a graduate student like me, when I asked Professor Feynman to come to my place and have sukiyaki with me, he immediately accepted my invitation, and we had a lovely evening in my very small room.

DeVorkin:

You cooked for him. Do you remember the sorts of things you talked about?

Koshiba:

What impressed me is how he came up with his Feynman Diagram. I think I told you about this yesterday. He said he didn't know anything about the second quantization; he just studied the classical electrodynamics inside out. And since he believed in special relativity, his method of treating the four components on an equal footing should be correct, but he wasn't sure. So using his diagram method, he calculated the physical process of which he knows the exact answer, which is Compton scattering. And then he got the correct answer, so he was sure that his diagram method is a correct method. Those are the things he told us.

DeVorkin:

When did you first encounter the Feynman Diagram, and what did you think of it as a heuristic tool?

Koshiba:

Well, I told you that at the beginning of my graduate work in Tokyo I had written a theoretical paper with the help of a friend which was on the nuclear interaction of muons through electromagnetic interaction. In calculating this, we used the Feynman Diagram method. So I knew about the Feynman Diagram.

DeVorkin:

But did it impress you. I remember when I first encountered Hamiltonian operators, I thought this was a very efficient new way of performing calculations. I was actually very excited about it, even though I didn't do anything with it. But in your case, going on in physics, did you see this as a major tool, as a major way to explore fundamental physics?

Koshiba:

Yes. I found the Feyman diagram a very, very powerful method. Unfortunately, all the processes which are calculable within a reasonable time have already been done by my seniors. So the process I calculated with the Feynman Diagram method was what is called the third order process. Third order in coupling constant. If you want to go to fourth order, there are many uncalculated, interesting problems. However, the difficulty is very much greater. The complexity is very much greater than third order processes. Okay?

DeVorkin:

Yes, sure.

Koshiba:

Not only that. The Feynman Diagram method is a very good method when you are dealing with an electromagnetic process, because the coupling constant is small. But when you have strong coupling involving pions and so forth, then the validity becomes dubious.

DeVorkin:

So in the electro-weak areas.

Koshiba:

For instance, Professor Kinoshita at Cornell is still working on the Feynman Diagram, working on the electromagnetic process. I think he has been working on something like the fifth order, or sixth order, I don't know.

DeVorkin:

Yes. Very complex. Well, you defended your thesis. Is there anything else we should be talking about at Rochester? I asked you about the conference, and I'm wondering what you felt about it. Was this the first large-scale physics meeting that you went to?

Koshiba:

That's right. But in those days Rochester conferences were rather small, altogether about 50 people.

DeVorkin:

But you must have met a lot of people at that time. Were you consciously comparing American physics with Japanese physics, the way physics is done, how physicists behave and interact? Did you find differences?

Koshiba:

Of course.

DeVorkin:

I'd like to know what your observations were.

Koshiba:

The first thing I noticed is that in this open discussion even the well established professor can be asked very embarrassing questions by young people.

DeVorkin:

I see.

Koshiba:

That would not happen in Japan. Even if one notices a mistake in the senior professor's lecture or statement, you are supposed to keep quiet. That is considered to be a part of good manners. But in the United States, even a student can ask very straightforward questions, or express a doubt to very senior people. That was one thing I learned.

DeVorkin:

Okay.

Koshiba:

That, to tell you the truth, made my life a little bit difficult after I came back to this country.

DeVorkin:

Did you participate? Did you learn that you could stand up and question?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

I mean, here it was one thing for you to show that Marcel Schein's work had to be reinterpreted or corrected. Was Schein at the conference, and did you get a chance to talk with him?

Koshiba:

I don't remember whether Marcel was there or not.

DeVorkin:

Okay. Certainly it was a different way of conducting physics. But was it one that you felt comfortable with?

Koshiba:

I found it very nice, because for instance the dean of the graduate school at that time, at Rochester, was a very well known professor, Arthur Roberts. He was not only a physicist, but also a music composer.

DeVorkin:

Really?

Koshiba:

Yes. He was a very big man, with the dignity of being the dean of the graduate school, I was sort of scared of him. In this Rochester conference Professor Roberts presented his experiment and results using the Rochester cyclotron. It was about pion-nucleon scattering. And then, a very small fellow sitting in front of the auditorium, stood up and asked a question about the background contamination to this big Professor Roberts. Literally, Professor Roberts trembled. I later found out this small man was Enrico Fermi. [laughs]

DeVorkin:

Did you meet Fermi then directly? Did you have a chance to talk with him?

Koshiba:

I was introduced by Mort Kaplon, but probably Fermi himself doesn't remember me. After I moved to Chicago, unfortunately, Enrico was already in the hospital. However, through Marcel Schein and his wife Hilde, we were introduced to Lola Fermi, Enrico's wife; they were very nice to us.

DeVorkin:

Once you got your Ph.D. you said your salary increased to $400 a month. Was that at Rochester?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

What was your position there?

Koshiba:

Research associate.

DeVorkin:

So they gave you a research associateship right away. Was this a position that you could have kept? Or were they expecting you to look for another position?

Koshiba:

No. Before that I was already offered a research associateship from Marcel Schein.

DeVorkin:

Oh, I see. Why did you then stay at Rochester after your degree for awhile? How long did you stay there?

Koshiba:

Because my appointment was from the first of July, Rochester university made up this three months.

DeVorkin:

Were you offered this position with Marcel Schein directly, or did you apply for it?

Koshiba:

I applied for it. When I was expecting to receive my degree within a few months, I wrote to Brookhaven and Chicago, because the work there was somewhat related. And then Schein offered me the job.

DeVorkin:

As a research associate. And was it also $400 a month there?

Koshiba:

I think it was better than that. I think it was something like $480.

DeVorkin:

Okay. By now you've been in the United States just about two years, and I'm curious about what your plans were. Did you feel that you would eventually go back to Japan, or did you want to keep your options open?

Koshiba:

I wasn't really thinking of going back to Japan. I had a very tough life there because of the poverty. My father couldn't get a good job, and I had to help my family.

DeVorkin:

Were you sending money home?

Koshiba:

Once a month I sent $100, because my brother was attending Rikkyo University in Tokyo. I gave the money to my brother before I came to Rochester to pay the university tuition. But this young man spent it on other things. About a year later he wrote to me to the effect that unless I paid this much within this period "I will be kicked out of the university." So I borrowed some money from Mort Kaplon and I sent $100 to my young brother.

DeVorkin:

To pay for his tuition.

Koshiba:

So, for the first few months at Rochester, even though I did receive $400, most of the money went to returning debts. [laughs]

DeVorkin:

Oh, sure. Well now you moved to Chicago, and this is —

Koshiba:

'55.

DeVorkin:

In 1955, July, you became a research associate. Were you working in what you called CASA?

Koshiba:

No. That was much later.

DeVorkin:

Oh, that was much later. What was your role then in Marcel Schein's group in 1955?

Koshiba:

Marcel gave me balloon exposed stacks to analyze, and the first work I did was to determine the chemical abundances in the primary cosmic radiation.

DeVorkin:

Right.

Koshiba:

And through this work, I became acquainted with Chandra.

DeVorkin:

Now, let's review the steps that you talked about last night that brought you to the realization that the carbon-nitrogen-oxygen (CNO) to iron ratio was different from what it was as determined by Suess and Urey —

Koshiba:

Urey. The CNO abundance inversion was later corrected in the cosmic abundance data to conform with our cosmic data. So that wasn't a real problem.

DeVorkin:

It was later corrected?

Koshiba:

Yes. Suess and Urey made a correction later.

DeVorkin:

But based on your data.

Koshiba:

I don't know if it was based on my data, but probably they did analyze the spectroscopic data of the sun. You probably know that in order to determine element abundances you have to use the Saha equation in a very complicated process.

DeVorkin:

Correct. And transition probabilities and a lot of things like that are very difficult to determine.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

In analyzing the plate stacks, were the techniques that you used in Chicago any different than the techniques at Rochester and, if so, what were they?

Koshiba:

Well, it was rather simple. In Rochester days, I used a method developed by the Bristol people and the Rochester people, what is called delta-ray counting.

DeVorkin:

Delta-ray, yes. Okay.

Koshiba:

However, this method takes a long time.

DeVorkin:

This is with a microscope?

Koshiba:

Mm-hmm [affirmative]. When I moved to Chicago, because I didn't have any group member and only one undergraduate student to help me in the experiment, delta-ray counting is not the method. So I used a really simple technique of what I call gap counting.

DeVorkin:

Gap counting. Yes.

Koshiba:

When the ionization becomes large, the gap on the track becomes smaller. If you make a good, reliable calibration of this gap density, it works as a charge determination of the track.

DeVorkin:

I see. Is this a technique you developed?

Koshiba:

As far as I know, nobody else has used it, unless in a very low energy proton or helium case. Not on heavy elements.

DeVorkin:

This allowed you to correlate plates in the stack better?

Koshiba:

So you make a very thorough calibration all over the entire stack. Because when you collect many events there are cases in which carbons split into 3 alphas, then you know when you see the 3 alphas almost parallel, then the primary has to be carbon 12. Yeah? So, you count the gap on this primary track, then you get calibration on carbon. Something like this, this calibration event, you collect so many of them, and then you can rely on this gap method. But it isn't a very commendable method, because the gap density depends strongly on the development of the plate, its degree of development. So you have to make the calibration on every plate in the stack.

DeVorkin:

Do you use a sensitometer to produce calibration marks on each plate? Because astronomers did that routinely in spectroscopy.

Koshiba:

Yes. But what I used is just a fragmented alpha particle that was traveling from plate to plate, and then I counted the gaps.

DeVorkin:

Ah. Oh, so you used the particles themselves and identified a particle as a calibration source. I see. Aha. Okay. Now, with this technique, with the help of your one undergraduate, is this when you did the CNO iron abundance? And could you go through the steps that led you to realize that it was different from what Suess and Urey had come up with? And what that meant to you, what the significance was?

Koshiba:

As I said to you yesterday, in those days the interest of primary cosmic ray researchers was concentrated on the relative abundance of light nuclei, lithium, beryllium and boron, as compared with the abundances of CNO. Of course that is an important quantity to make estimates of — the amount of interstellar material the primary cosmic radiation had to traverse before reaching our detectors.

DeVorkin:

The interstellar material. Right.

Koshiba:

I noticed first of all this CNO abundance is reversed as compared with Suess and Urey's cosmic abundances. Also, I noticed another thing, that the heavy elements like calcium, and iron, were considerably more abundant than in the Suess and Urey determination, which I couldn't explain. I relayed this result to Marcel Schein, who kindly introduced me to Professor Chandrasekhar. And Chandrasekhar explained to me that there are different types of stars generally divided into the category of young stars and old stars and so forth, all those fundamental results on stars. I then could conclude my paper by saying that judging from the observed chemical abundances of primary cosmic radiation the likely ion source of the primary radiation can be supernovae and so forth. And I thanked Professor Chandrasekhar at the end of the paper.

DeVorkin:

Yes. Now the idea that supernovae were causes, were one source at least of cosmic rays, high-energy cosmic rays, was reasonably well known, but you had mentioned that Hayakawa had made that suggestion.

Koshiba:

Hayakawa.

DeVorkin:

When did he make that suggestion? Do you know?

Koshiba:

I don't remember exactly, but I think a few years before my result was published.

DeVorkin:

But did you know this generally, or did Chandrasekhar tell you about this possibly? Did you make this suggestion independent of knowing of Hayakawa's suggestion?

Koshiba:

I didn't know of Hayakawa's suggestion. I found out only after I came back to Tokyo.

DeVorkin:

I see. But you knew that people had been thinking that supernovae could be the sources of cosmic rays.

Koshiba:

Yes. Many people speculated about that.

DeVorkin:

Yes, okay.

Koshiba:

But when it comes to showing this possibility in the form of experimental data, I didn't know what other people had done before this.

DeVorkin:

I see. Now you said you published this paper in Nuovo Cimento.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

And this was about in, about what year? 1956?

Koshiba:

'56 I believe.

DeVorkin:

Yes. Did you get any reaction to this paper? Did anybody write you?

Koshiba:

Well, the only reaction was from Hayakawa. Hayakawa was very happy.

DeVorkin:

Because it would support his idea. What about from Suess and Urey? Any reaction from them?

Koshiba:

No, they were above it.

DeVorkin:

Oh, so they didn't pay any attention?

Koshiba:

I don't know. [laughs] I didn't hear anything.

DeVorkin:

Okay. Moving on then, at Chicago how did your career progress there? Did you remain a research associate? Did you get a faculty appointment?

Koshiba:

I was a research associate for nearly three years, two years and eight months, and then I was offered an associate professorship from the University of Tokyo. So I came back.

DeVorkin:

You were offered this unbidden. You didn't apply for it?

Koshiba:

I knew that they were searching for either a research assistant or an associate professor. Probably my friend, a senior person, made a recommendation. I was written while I was in Chicago and asked if I intended to accept this appointment. I said if it were an assistant appointment, I wouldn't go back; if it were an associate professorship, I would come back. I came back as an associate professor.

DeVorkin:

Okay, was this after Schein died?

Koshiba:

No, before Schein died.

DeVorkin:

Okay. So will you help me understand how you eventually replaced Schein after he died?

Koshiba:

When I was coming back to Tokyo University, Schein was very kind and gave me a small stack for me to work on in Japan.

DeVorkin:

Okay. By this time you still had not actually worked with balloons. You had been given stacks to analyze.

Koshiba:

That's right.

DeVorkin:

Okay.

Koshiba:

I think it was next year that Schein wrote me saying that he was now organizing a very large scale international collaboration using a big stack. He wanted me to come and help him in this international enterprise. So I took a leave of absence from the University of Tokyo.

DeVorkin:

This was about 1957?

Koshiba:

No, it was 1959.

DeVorkin:

'59. Okay.

Koshiba:

And I went there with my wife.

DeVorkin:

So you were married when you came back to Tokyo?

Koshiba:

After I came back to Tokyo, several arrangements were made for me to get married, and eventually I took one of them.

DeVorkin:

Right. Okay. I've heard of that process.

Koshiba:

[chuckles]

DeVorkin:

Did you interview your wife?

Koshiba:

Oh, yes.

DeVorkin:

Yes. And this was all standard: a go-between and arrangement?

Koshiba:

Yes.

DeVorkin:

What is your wife's name?

Koshiba:

Kyoko.

DeVorkin:

Okay. And does she have a university education?

Koshiba:

Mm-hmm [affirmative]. She majored in the history of art, especially Buddhist art.

DeVorkin:

Did she practice her profession at any time?

Koshiba:

Before she married me, she was assistant curator of a museum.

DeVorkin:

But after that she didn't work.

Koshiba:

No. She became a housewife.

DeVorkin:

Right. And so when you took a leave of absence she came with you.

Koshiba:

Mm-hmm [affirmative]. And then as I told you, three months after our arrival, Schein died. My wife was already pregnant. My son was born in Chicago.

DeVorkin:

Aha. So he has American citizenship? Or didn't he take it.

Koshiba:

If he wants to, he can have it, mm-hmm.

DeVorkin:

Yes. Now Schein had already flown, or tried to fly the first stack from the Valley Forge, is that right?

Koshiba:

Yes.

DeVorkin:

You were not there for that.

Koshiba:

No. That was before my time.

DeVorkin:

Okay. And that was a failure of the balloon I understand.

Koshiba:

Mmm. Well, let me say it was a partial success.

DeVorkin:

How long did it fly?

Koshiba:

Only seven hours or so, [at a low altitude?].

DeVorkin:

But it was a big payload. It was over a ton, wasn't it?

Koshiba:

Well, altogether it was approaching a ton, but a real payload is not as heavy.

DeVorkin:

Okay. When you arrived in Chicago, what did Schein want you to do? What was your value to the group?

Koshiba:

Well, Schein probably expected me to analyze the primary radiation from this new stack.

DeVorkin:

The one that had flown on the Valley Forge. Okay. And did you actually do that sort of work?

Koshiba:

No, it wasn't developed yet when Schein died.

DeVorkin:

Oh, I see. You said it was developed?

Koshiba:

Yet.

DeVorkin:

And so what was the process after that? When he died that must have been unexpected and a shock to everybody.

Koshiba:

For some time it was turmoil, yes, and the University of Chicago didn't know what to do. So they asked Professor Occhialini, who was at that time spending a year in Bruno Rossi's group at MIT, to come and give advice.

DeVorkin:

And what did he do?

Koshiba:

Occhialini came and interviewed all the group members of Professor Schein, and eventually recommended me as the one to succeed Marcel's role as the principal investigator of this whole thing. So that was the beginning of my busy days.

DeVorkin:

Busy days?

Koshiba:

Mmm.

DeVorkin:

So how did you restart the program?

Koshiba:

It wasn't easy. First of all, Marcel had used up all the research funds. Almost every week, with the help of Professor Occhialini, I had to make a trip to Washington, D.C. asking for more money. We visited the Office of Naval Research, and the National Science Foundation. I didn't go to the Atomic Energy Commission in those days, because Marcel had contracts only with NSF and ONR. So I visited those two offices.

DeVorkin:

And what was your experience?

Koshiba:

Well, that was my first experience to get my own funding. It wasn't easy. I sometimes made our section head angry.

DeVorkin:

Angry?

Koshiba:

Yes.

DeVorkin:

How did you manage that?

Koshiba:

Because I was too outspoken.

DeVorkin:

In what way? What were you saying to him?

Koshiba:

Oh, what had to be done, how to do it, and so forth. In exasperation, the man said to me, "Are you trying to tell me how to run this office?"

DeVorkin:

And you said?

Koshiba:

I had no such intention. It might have sounded like that due to my poor English. It wasn't easy.

DeVorkin:

Were you only trying to be very forceful about the importance of the program?

Koshiba:

I tried to be persuasive. A few years later after all the operations with balloons were over, I received a letter from the Chief of Naval Operations. The letter said that I was very persistent and persuasive and so forth. I don't remember exactly the phrasing. But that was nice of him. Eventually I got the research funding, and then started to go about exposing the remaining stack.

DeVorkin:

This was a large collaborative project.

Koshiba:

Yes.

DeVorkin:

How many different groups were involved, and how were they organized?

Koshiba:

Oh, about 12 countries were involved, including Poland.

DeVorkin:

And did each country send a representative to Chicago?

Koshiba:

In later stages, yes.

DeVorkin:

And how was the work divided up? I mean in the way that Schein had done it, and then how did you change it?

Koshiba:

The idea was that almost everything was to be done by the University of Chicago group, like purchasing the stack and preparing the block and container, contracting with the balloon company, actually exposing and developing and fixing, and cutting the stacks into usable sizes. Only after the plates were exposed and processed, and the only remaining thing was to use a microscope to look at the plate to find the event and analyze it, did the collaborating groups enter the picture. They got a certain portion of the stack, started analyzing it, and then brought the data together.

DeVorkin:

I see. Did you continue that sort of organization?

Koshiba:

Yes. [laughs]

DeVorkin:

But you were now the one leading the project. Were you still employed as an associate professor by the University of Tokyo.

Koshiba:

Yes, I was on leave of absence.

DeVorkin:

An extended leave.

Koshiba:

The University of Chicago gave me the honorary title of visiting associate professor. That was only honorary. Officially, I was the acting director of the laboratory for high-energy physics and cosmic radiation of the University of Chicago.

DeVorkin:

This was a large collaboration.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Okay. Now you were the one who had to do all the organizational work. But you mentioned last night that you had a number of bids from different companies and you decided on a certain combination. Could you review that for the interview now. Was this the way Schein did it, or did you change vendors let's say?

Koshiba:

No. Marcel didn't go into the details of the balloon launching operation. He had the idea of using an aircraft carrier.

DeVorkin:

An aircraft carrier. Yes.

Koshiba:

But that could be done only because he was a big shot with the Navy. But that couldn't be done for me. So I had to think about various ways of launching the stack safely and recovering it with certainty. I discussed many possible solutions with Professor Occhialini.

DeVorkin:

Did he stay in Chicago?

Koshiba:

No, he came to Chicago almost every month, and stayed several days each time.

DeVorkin:

Did he also go to Washington with you?

Koshiba:

Mm-hmm [affirmative], sometimes, yes.

DeVorkin:

So he helped in that regard.

Koshiba:

Yes.

DeVorkin:

Was he officially a part of the collaboration?

Koshiba:

No. He was what you might call the supreme advisor of the whole thing.

DeVorkin:

Okay. So how did you decide to build the second balloon system?

Koshiba:

Well, I knew that the Raven Company had the best launching crew, and I also knew that Winzen produced the best balloons. So I wanted to combine these two. In the end, General Mills came in forcefully with their own plan. They wanted to make a contract to launch a stack, and to expose it using a General Mills balloon and a General Mills launching crew.

DeVorkin:

How were they able to do that? I mean, did they come only to you, or did they go to ONR? Did they exert pressure somehow?

Koshiba:

I did get some pressure from the government side. I don't remember exactly from whom, but General Mills was a very powerful organization; they must have worked some way or the other. I didn't approve of the launching procedures of General Mills and also I didn't have very much confidence in General Mills in general. We wrote down the conditions of a successful flight operation, and further stipulated that only when these conditions were satisfied would we pay the $40,000 launch fee. They accepted these conditions, to remain aloft for a minimum of six hours at an altitude of 70,000 feet. These were rather easy requirements. However, the operation didn't satisfy the conditions, so I didn't pay a penny. Later they came back and wanted their $40,000, but because of the agreement they couldn't do anything.

DeVorkin:

You had saved the document and the Chicago lawyers defended it. Yes. When you wanted to mix together Raven and Winzen, were they willing to do that?

Koshiba:

Yes.

DeVorkin:

They were happy to do that?

Koshiba:

Probably Mr. Winzen wasn't very happy, but I insisted.

DeVorkin:

All of these people sort of grew out of the General Mills group anyway.

Koshiba:

Right.

DeVorkin:

So they're all kind of connected.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

When the General Mills-based flight didn't meet your expectations, what did you do? Did you plan a third flight?

Koshiba:

I haven't had anything to do with General Mills since.

DeVorkin:

But you still wanted to fly a stack. How did you go about arranging for that? You didn't have to pay General Mills, so apparently you —

Koshiba:

I had the full amount in my hand still.

DeVorkin:

Yes. Did you go back to Winzen and Raven?

Koshiba:

I contacted Raven, and authorized them to purchase a Winzen balloon.

DeVorkin:

Okay. And where did they fly from?

Koshiba:

There were two flights. The first one was from the Naval Air Station, Glynco in Georgia.

DeVorkin:

You had an interesting story about that.

Koshiba:

Henry Miller?

DeVorkin:

Right.

Koshiba:

Am I to repeat it?

DeVorkin:

Sure. It's an interesting story.

Koshiba:

[laughs]

DeVorkin:

You were down there, about when was this?

Koshiba:

It was sometime around May or June of '60. In order to launch a big balloon you had to have very good weather conditions, especially the surface winds. So I stayed in the bachelor officers' quarters, and every evening I would go to the bar and have a drink. One day a very handsome looking young ensign came over to me and we started up a conversation. This young man then asked me if I had read any English novels. I thought I would tease him, so I said “Maybe you have never read this man's book, because his books are banned in this country and they are available only in Paris.” “Who is that?” he said. I said, “Henry Miller. Do you know him?” And this young man's eyes beamed at me, and asked “Do you really think Henry Miller is the greatest English author?” I said, “Yes.” He said, “You know what? I grew up calling him Uncle Henry.” “What?” I said. And he explained his mother was a divorcee who lived with Henry Miller for some years. At the end of our conversation, this young man said to me, “Uncle Henry is planning a half a year trip to Japan. Would you give him help?” I said, “Of course, if there is anything I can do I will be glad to help him.” Then the balloon was launched, the fate of which I told you about yesterday. I came back to Chicago, and left immediately on a 2-week trip to Moscow. When I got back, I found an air mail letter waiting for me. In the old days, there was one sheet of paper which you folded.

DeVorkin:

Yes. An aerogramme.

Koshiba:

Yes. An Aerogramme. That was the form of his letter in his own handwriting. My wife still keeps that letter somewhere, but I remember the content went something like this, "My young friend so-and-so tells me that you can help me in Japan. That's very nice of you. Please find a Japanese lady who can take care of me during the six months of my stay in Japan." Luckily I didn't have to answer him, because he had already gone.

DeVorkin:

Yes. That was an interesting side-light. Now about the flight, how did that go?

Koshiba:

Well, as I said, it had to go through a cloud layer and —

DeVorkin:

You launched it through a cloud layer. Yes.

Koshiba:

Mm-hmm [affirmative], and it must have received a lightning bolt which in turn must have broken the circuit of cutting the payload from the balloon. So when the balloon floated to Texas, our crew wanted to cut the payload down using a the small airplane. It didn't work.

DeVorkin:

The airplane couldn't reach the altitude?

Koshiba:

No. In any case a small plane cannot reach the balloon's altitude of some 76,000 feet.

DeVorkin:

Right.

Koshiba:

So you send a radio signal to cut the payload down with a parachute. However, the receiving circuit on the balloon must have been damaged by the lightning, because it didn't respond. It stayed on the continent for a few more days and then drifted to the San Diego area. I had to call the Pacific Fleet people to shoot it down.

DeVorkin:

How did you call them? I mean, what channels of communication did you have that allowed you to do that? Was it through your ONR project manager?

Koshiba:

Yes, because I was the principal investigator of a rather large sum — in those days nearly a million dollars was really a big project.

DeVorkin:

Sure.

Koshiba:

So the Office of Naval Research gave me a sort of corresponding rank in the Navy that allowed me to talk directly to the staffs of the Pacific Fleet in San Diego. At first, they proposed to send fighters and shoot the balloon and burst it. Accordingly, they sent four fighters which had to dive first and then climb, because the balloon was at such a high altitude.

DeVorkin:

Right.

Koshiba:

Three out of the four fighters chickened out, only one of them fired at the balloon. But because it's such a huge balloon, a couple of bullet holes didn't make any difference.

DeVorkin:

Yes. What was the volume of the balloon?

Koshiba:

Ten million cubic feet.

DeVorkin:

Yes, that's a big balloon.

Koshiba:

Yes.

DeVorkin:

So it didn't work.

Koshiba:

No. Then one of the staff suggested the use of a sidewinder missile. But I knew that the sidewinder homes on the infrared.

DeVorkin:

Infrared.

Koshiba:

Yes. I told the staff that our precious payload was painted red, and that the sidewinder may home in on the payload rather than the balloon. So that was out.

DeVorkin:

Right.

Koshiba:

And then I remembered the high flying airplane, the U-2. I asked if a U-2 was available, and they said "Yes, we do have a U-2." I suggested that they attach an anchor to the U-2 and make the anchor knife-edge sharp so that it can rip open a plastic balloon. A couple of hours later they called me back, and said "I'm sorry, but the U-2 is full of photographic equipment, and not only that, the U-2 is built very lightly, so therefore it cannot take the" —

DeVorkin:

Drag?

Koshiba:

Yes. It was not strong enough to hold the anchor and to rip up the balloon. I must say I wasn't kind with those people. I shouted that, "Well, you people spend lots of money and claim to be able to shoot down an ICBM from Russia, and you are telling me that you cannot shoot down a slow-moving balloon which is only at 70,000 feet?"

DeVorkin:

What did they say?

Koshiba:

[laughs] Well, they couldn't really say anything. I wasn't very nice.

DeVorkin:

I guess they could have gotten it with a sidewinder, but the problem there is that you didn't want to destroy the payload. So this was a delicate operation.

Koshiba:

That's right. Yes. So that was that. That payload was finally lost over the Pacific.

DeVorkin:

Yes. You said that somebody later, a Russian physicist —?

Koshiba:

Joked about it, yes.

DeVorkin:

Who was that?

Koshiba:

I don't remember. I think it was Chudakov, who is a well-known cosmic ray man in Russia, an academician. But I'm not sure if it was him or not.

DeVorkin:

But somebody apparently just took you aside at a meeting and said, “We have it”?

Koshiba:

“Some unknown red-painted object came down from sky in the thick forest of Siberia.” [laughs]

DeVorkin:

Did you know anything about the classified balloon reconnaissance programs that the United States had called Moby Dick that would go all the way around the world?

Koshiba:

In a sealed Mylar balloon. Yes.

DeVorkin:

And photograph just at random over the Soviet Union. Did you know?

Koshiba:

No, I didn't know exactly what was going on, but I did know that some people were very much involved in developing a superpressure Mylar balloon.

DeVorkin:

Okay. Now when that one didn't work, did you fly again?

Koshiba:

Yes. The next launching operation was from the State of California. There is a small town near the Mexican border by the name of Brawley.

DeVorkin:

Why was that chosen?

Koshiba:

At the time (autumn, 1960) the upper winds were predominately from the west. So in order to make use of the whole width of the American continent, the launching had better take place in as western a place as possible.

DeVorkin:

Yes. Right.

Koshiba:

This operation was very successful. Raven people did a very good job; the balloon reached an altitude of 100,000 feet, and stayed at that height for more than 36 hours. Of course, the balloon was chased by a small airplane. The payload was dropped over the State of South Carolina, and came down in a very busy highway in rush hour — it stopped all traffic. [laughs] But the police immediately took care of it, and safely returned it back to Chicago. No casualties.

DeVorkin:

No casualties. Right. But did any official get angry or were there any letters of protest?

Koshiba:

No.

DeVorkin:

Okay. So then you finally had a plate stack that you could distribute.

Koshiba:

Not right away. We started the 24-hour operation of unpacking the payload and sealing each nuclear emulsion in a big glass, drying it, developing it, fixing it, and then cutting it.

DeVorkin:

Oh, so you would mount the nuclear emulsions on glass before you developed them. Uh-huh. Okay.

Koshiba:

That took about six weeks of a full 24 hours operation.

DeVorkin:

Yes. How many people did you have working with you at that time?

Koshiba:

I had altogether about 17 or 18 people.

DeVorkin:

Seventeen or 18 people. So it was a good-sized team. And that didn't include collaborators from the other institutions?

Koshiba:

One or two people came to help, but most of the collaborators stayed away.

DeVorkin:

What were the scientific results from this flight? What would you say were the most significant?

Koshiba:

Well, unfortunately the stack was divided up too soon. The ideal situation would be for the collaborators to join the analysis in one place where you can go from one end to the other.

DeVorkin:

One end of the stack to the other end?

Koshiba:

Mm-hmm [affirmative]. But when it is divided in pieces and distributed all over the world, and you reach the end of your part, you have to send a cable, or a telegram, to your neighboring collaborator. So it wasn't very efficient, I must say. However, we did get a good result on the nuclear-nuclear interaction in the 1000 GeV energy range, which was completely unavailable from accelerators. In those days accelerators could produce only 40 GeV or so.

DeVorkin:

And this was 1000 GeV. Yes. Indeed you were working at the forefront of high energy interactions. Now, during this time you came to know other ballooning people. You mentioned last night the Minnesota group. Also, at Rochester there were Bradt and Peters. Were they still there when you were there?

Koshiba:

Yes. Peters was I think at MIT.

DeVorkin:

Oh, okay, but Bradt was at Rochester.

Koshiba:

Bradt died before I arrived.

DeVorkin:

Oh, I see. So you really didn't meet any of them.

Koshiba:

The only person from that old collaboration was Mort Kaplon. That's one of the reasons why I had chosen him as my thesis advisor.

DeVorkin:

How did you get to know the Minnesota group, the people at Minnesota?

Koshiba:

Well, I was invited to give a talk.

DeVorkin:

Ah, uh-huh, at Columbia.

Koshiba:

I went there and met Ed Nye, and immediately we became good friends.

DeVorkin:

He was a wonderful fellow.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

They had been working on experimental balloons for quite some time.

Koshiba:

That's right. That was before my time.

DeVorkin:

Yes. But at this time were you interested at all in the mechanism of ballooning, in developing balloon systems? Because the Minnesota group was very big in that. You mentioned, for example, how Ed Nye invented the duct design to vent helium from balloons. I didn't know he was the one who had done that.

Koshiba:

I'm not sure whether it was actually Ed Nye or not, but I guess it was because of the talks we had.

DeVorkin:

Did you meet Ed Lofgren, or Al Nier, or Phyllis Frier?

Koshiba:

Yes, I met them, but only casually.

DeVorkin:

So it was Ed Nye that you established a relationship with. Okay.

Koshiba:

I had mentioned that I took a colleague of mine, a Japanese colleague to Minneapolis, whose name was Jun Nishimura. Nishimura later became the director of ISAS, was much more interested in the balloon itself than I was.

DeVorkin:

Okay.

Koshiba:

And he is considered the father of ballooning in Japan.

DeVorkin:

Uh-huh. And so this was his introduction when you took him to Minneapolis? This was his introduction to the ballooning systems?

Koshiba:

No, even before that he had some experience with plastic balloons.

DeVorkin:

I see. Okay. That is interesting. Because I always like to find out if scientists who used rockets for scientific research right after the war became interested in the problem of working on rockets, rather than what kind of science they were doing with them. But I take it there was never any question in your mind; it was always the science?

Koshiba:

Yes. I wasn't very much interested in the balloon itself. I just wanted to learn how to make a successful balloon launch for research purposes. But Professor Nishimura was also interested in fabricating balloons and so forth.

DeVorkin:

Was this the final flight that you did?

Koshiba:

Final flight. Yes.

DeVorkin:

Was the contract with ONR basically to get one successful flight and then it would be over?

Koshiba:

Yes. But they provided me a subsistence fund for the laboratory.

DeVorkin:

During that time.

Koshiba:

Afterward.

DeVorkin:

Afterwards. So you stayed at Chicago.

Koshiba:

After the operation, yes. Mm-hmm [affirmative].

DeVorkin:

How long did you stay?

Koshiba:

Well, I stayed until the end of August '62, so a year and nine months.

DeVorkin:

Yes. Now you mentioned last night something about the Chicago air shower array. Does this come later?

Koshiba:

Oh yes. That was much later.

DeVorkin:

So we won't talk about it now. The Auger project was much more —

Koshiba:

That was still more recent.

DeVorkin:

Okay. There was no intention then for you to stay at Chicago. You decided that, in '62 with the project over, —

Koshiba:

To go back to the University of Tokyo.

DeVorkin:

Okay. What were your plans, your professional plans, in returning to Tokyo?

Koshiba:

Well, I wasn't quite sure. When I went back to the Institute for Nuclear Study of the University of Tokyo, it was considered to be open not only to the University of Tokyo faculty, but also to other national university people. There was a steering committee formed by professors of many universities — many powerful people. As I said before, the stack is very useful when it is kept in the same place in order to facilitate efficient analysis. If it is divided into many places and distributed to different places, it slows the process down and a loss in accuracy of results.

DeVorkin:

It's hard to recombine the data?

Koshiba:

Mmm [affirmative].

DeVorkin:

Okay.

Koshiba:

But those big shot people wanted to get their share and take it back to their local universities.

DeVorkin:

Oh, so your share had to be divided among?

Koshiba:

Further.

DeVorkin:

Further. Oh my heavens. Okay.

Koshiba:

So I protested. That made me very unpopular. After about one year I had had enough. I wanted to go back to the United States, because I did have a couple of offers from American universities. Just at that time a different faculty, the faculty of science of the University of Tokyo was looking for an associate professor. I applied. I decided that if they didn't accept me, I'd go back to United States. Luckily I was accepted. So I moved to the physics department of the faculty of science and left the Institute for Nuclear Study.

DeVorkin:

Oh, so you went to the University Tokyo proper.

Koshiba:

Yes.

DeVorkin:

And it was a faculty of science.

Koshiba:

Mmm [affirmative].

DeVorkin:

This Institute for Nuclear Studies was a consortium of many universities. That's what you were saying?

Koshiba:

Essentially yes.

DeVorkin:

Now I see. Okay.

Koshiba:

Even though nominally it was part of the University of Tokyo, and was supposed to be open to many other universities.

DeVorkin:

So you achieved a regular faculty slot at the university rather than an institute slot. What was the relationship as far as research potential was concerned between institutes and universities in Japan, at least at that time? Did institutes have a higher status? Did they have more research funding than universities, or was it the other way around?

Koshiba:

Well, if I may make a frank statement, those open institutions, like the Institute for Nuclear Study, or KEK, which are national open laboratories —

DeVorkin:

National —?

Koshiba:

Laboratory for high energy physics.

DeVorkin:

Right. Yes?

Koshiba:

They do get very large accelerators and plenty of funding for experimentation, but when it comes to the level of people as a scientist I think some of the national universities like the University of Tokyo are higher. For instance, even in the United States, take the case the Brookhaven. Okay? It is run by a university consortium.

DeVorkin:

Right. AUI (Associated Universities, Inc.)

Koshiba:

Mm-hmm [affirmative]. And good experiments are originally the proposed by university people. Right? Like the J-particle discovery, or Jim Cronin's PC violation.

DeVorkin:

Mm-hmm [affirmative], yes, very important.

Koshiba:

So the national laboratories should be willing to accept the original proposals of outside university people. I think that's very important.

DeVorkin:

Now that's true in the United States. Is it true here in Japan?

Koshiba:

I wish I could say yes. When I was a council member at KEK, I did make a lot of noise.

DeVorkin:

A lot of noise. [laughs] Well, okay, going back then to your returning, you stayed in Japan, you became a regular faculty member at the University of Tokyo, and you had your small sample of plates to analyze. What was the end result of that work?

Koshiba:

When I moved to the faculty of science, I was supposed to take a graduate student every year, right?

DeVorkin:

Right.

Koshiba:

Luckily every year two or three graduate students wanted to come to my laboratory. Since I had nothing but just a block of nuclear emulsion, I gave this emulsion to these people to analyze, and some of them did write doctoratal theses. But I worried about the future of those young people if they only learned the nuclear emulsion technique — it would not be easy for them to find jobs in the future. I thought then of switching to the analysis of bubble chamber pictures.

DeVorkin:

Oh, let me ask this. The method of reduction of analysis of the plate stacks, did it remain the same, direct manual observation with microscopes?

Koshiba:

We did try to introduce some computer help, but in those days computers were still in a primitive state, so we didn't do very much.

DeVorkin:

So it was still largely manual.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Okay.

Koshiba:

And then when I looked around the world at the situation in which large bubble chambers were analyzed, I was rather discouraged because even if you had say $4 million to install couple of analyzing machines and hire ten scanners and so forth, still you can be only one of the collaborating groups of a big bubble chamber collaboration.

DeVorkin:

Yes, this is the idea of forming a big group and taking your bubble chamber to KEK and buying time.

Koshiba:

Yes.

DeVorkin:

Okay.

Koshiba:

That is not the type of work I like. To be just one part of a big collaboration, that is not to my liking.

DeVorkin:

Not your style.

Koshiba:

No. So I thought of switching my experimental work to electronics counter experiment.

DeVorkin:

Still doing cosmic rays, but doing it with counters?

Koshiba:

Because there was no good accelerator in Japan in those days. One of the results we obtained from the analysis of the big nuclear stack was that there was good reason to expect parallel high-energy muons, multiple muons, to be observed deep underground. There was already report of this phenomena from the Russian groups and from the Utah group.

DeVorkin:

From Utah?

Koshiba:

Yes. Occasionally, there is a case in which two parallel muons are observed, and very rarely three muons.

DeVorkin:

Was there any theory to back this up?

Koshiba:

No theories. But from the analysis of the 1000 GeV nuclear interactions, I thought there was reason to expect a more abundant occurrence of such multiple muon phenomena. So as a first electronics attempt, I set up an underground experiment in 1969 or so, almost 30 years ago. The same company, the Kamioka Mine Company, has a main office in Kamioka Town, which is about 10 kilometer south of this place.

DeVorkin:

Where we are now.

Koshiba:

Mm-hmm [affirmative]. And there is another mine. I was introduced to the president of this company by a very powerful man. In those days, this mine prospered.

DeVorkin:

Mining tin and zinc primarily?

Koshiba:

That's right. By order of the president himself, our experimental cave was prepared, and telephone and power lines were installed at the company's expense. What I did was to set up there a rather large, 3 meter by 3 meter, spark chamber, which can be used to follow the tracks of charged particles. We triggered this spark chamber by two layers of plastic scintillators. I was happy with the results of this experiment, because we did observe multiple muons up to a multiplicity of 18. That was the experiment by which Professor Totsuka received his Ph.D.

DeVorkin:

Aha. So he was working with you at that time.

Koshiba:

Yes, he was a second year student.

DeVorkin:

This is something then that theoreticians had to explain.

Koshiba:

Well, it is a rather complicated phenomena, and we did give our own interpretation in our paper, but since it cannot be analyzed in a clear-cut way, not many people were interested in it, unfortunately.

DeVorkin:

To what degree did you maintain your knowledge of theory, as particle physics continued to change? This was before the grand unified model. But there were a lot of problems with it, and a lot of things just didn't fit, quarks had been hypothesized but not yet seen, it sort of was in a messy state. What did you think of theory at that time, the late sixties?

Koshiba:

Well, as I said before, I had a very good friend in theory, Nambu in Chicago.

DeVorkin:

Yes. Did you stay in contact with Nambu after?

Koshiba:

Well, I called by phone and —

DeVorkin:

Oh, he spent half a year here, isn't that right, in Tokyo?

Koshiba:

Yes, after his retirement, yes.

DeVorkin:

Oh, I see, yes. We didn't talk about Nambu. But you came —

Koshiba:

He was a great physicist.

DeVorkin:

And you came to know him in Chicago.

Koshiba:

No, before that.

DeVorkin:

Oh. When?

Koshiba:

When I was a first year graduate student, he was already a professor at Osaka City University.

DeVorkin:

Oh, I see.

Koshiba:

I spent one month in his laboratory.

DeVorkin:

I see. As a theorist he had a laboratory?

Koshiba:

No, I shouldn't call it laboratory. His seminar room.

DeVorkin:

Fair enough. Okay.

Koshiba:

I spread a futon on the table and stayed there for one month. [laughs]

DeVorkin:

So you would talk with Nambu about the state of particle physics, the theory?

Koshiba:

Mm-hmm [affirmative], and also I had another good friend in theory named Nishijima.

DeVorkin:

You relied on others to give you the latest insight in theory?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Okay. So you were now moving into electronic detectors, and you had detected parallel muons up to 18, which sounds like quite an achievement. Has that ever been explained?

Koshiba:

Not completely. For instance, Barry Barish at Cal Tech —

DeVorkin:

Who?

Koshiba:

Barry Barish of Cal Tech.

DeVorkin:

Okay. I don't know that name.

Koshiba:

He's a big shot in physics. He started a similar experiment, very much bigger in size than our's in Gran Sasso, and they have been observing similar multiple muon events.

DeVorkin:

This gets us now to about 1970, 1971. What were your plans for the future for your experiments?

Koshiba:

Well, it just happened that there was a conference in Moscow around 1960 or so, which I attended. A Russian cosmic ray physicist came up to me, and said there is a Professor Budker who wants to speak to me. So Budker arranged a dinner party, and that was the first time I met him.Budker was a very good physicist and experimentalist, one of the pioneers in electron-positron colliding machines. He was building such a machine in Novosibirsk. Budker wanted me to come to Novosibirsk and carry out a joint experiment with his e+e- colliding machine. I was interested, but it required a number of difficulties to be overcome. First of all, it required a considerable amount of research money. Another thing was that it was prohibited by international agreement called COCOM, which prohibits western countries from exporting front-line electronics and computers to Russia.

DeVorkin:

So this is a Cold War problem.

Koshiba:

That's right. But I could manage to overcome the obstacles, and I was seriously thinking of doing the joint experiment there in Novosibirsk.

DeVorkin:

So he was interested in the electronics that you had, your electronic technology.

Koshiba:

The participation of my group in the experiment.

DeVorkin:

Yes.

Koshiba:

However, you know there is a saying in Russia, when you become an academician you can have everything. You get a big house, you have a chauffeured car, you get a dacha —

DeVorkin:

A what? Oh dacha, yes, summer house.

Koshiba:

But, they cannot obtain hard currency, foreign currency. All that an academician gets is rubles. What do they do? Because there is nothing else to do, they start changing their lives. They divorce their wife and get a new one. After about four wives they have a heart attack.

DeVorkin:

A what?

Koshiba:

Heart attack. That was exactly what happened to Budker.

DeVorkin:

You mean they have nothing else to do? What promotes this kind of behavior?

Koshiba:

Enjoyment. In the western world you can go out, take a boat trip to the Caribbean and so forth, and buy good things with your own money. But the Russian people with rubles cannot do that.

DeVorkin:

They can't go outside of Russia is what you mean, so things get very confined.

Koshiba:

Yes. That's why they start changing their wives.

DeVorkin:

I see. That's their form of recreation?

Koshiba:

Yes.

DeVorkin:

I see.

Koshiba:

Well, Budker got a heart attack. Even though I persuaded our ministry of foreign affairs to send an official document to the Russian government concerning our joint experiment in Novasibersk, the Russian government didn't give a straightforward affirmative answer. So I went to Novasibersk and discussed the situation and future possibilities. I said to Budker that unfortunately he was sick, and that there was no foreseeable possibility for a joint experiment in Novasibersk. I suggested that we call it off until he recovered his health. And then I went to Frascati, Italy where e+e- collider, Adone, was working.

DeVorkin:

Who was that?

Koshiba:

Adone is the name of the colliding machine. It's a nickname.

DeVorkin:

You went there to work actually?

Koshiba:

No, to see if there was anything we could do there. And then I went to CERN, Geneva, where a proton-proton colliding machine was working.

DeVorkin:

Right. Now there were all these possibilities. How much time did you spend at CERN, and did you perform any experiments while you were there?

Koshiba:

No. I spent only a week or so.

DeVorkin:

So these were visits to find out what was being done.

Koshiba:

Yes, what could be done. What would be the possibility of participation.

DeVorkin:

I take it this was about 1970-71 and you were a full professor? You were about 44 years old?

Koshiba:

Wait a minute. I think I became a full professor in 1970 or so.

DeVorkin:

And you were in your mid-late forties?

Koshiba:

Middle forties. I was 44.

DeVorkin:

Middle forties, forty-four, right. Is it correct to say that at this time you were searching for a new direction?

Koshiba:

Mm-hmm [affirmative], that's correct.

DeVorkin:

Okay. You weren't quite sure what you wanted to do.

Koshiba:

Yes. First of all there was Adone at Frascati. I didn't see very much of a future there, because a number of experiments had already been done, and upper limit of the energy achieved was only 3 GeV. Second, at CERN, as I said, there was a proton-proton colliding machine, but I wasn't very much interested, because the collisions are very much more complex than e+e- collisions. I like the simplicity. So after CERN I went to DESY, Hamburg, Deutsche Electron em Synchrotron. Luckily there was a professor there who was a member of my Chicago group, we became very good friends, and he helped me and introduced me to a number of people at DESY. At that time DESY was building a new e+e- colliding machine called DORIS. There were two experiments being prepared, and we joined one of the two, called DASP, for Double Arm Spectrometer.

DeVorkin:

Okay. So you had the backing of your university, you knew that you had some equity there, and you were looking, you were now realizing that you had to join a larger group at an accelerator to continue on in the area, and you were just looking for the best place.

Koshiba:

Yes.

DeVorkin:

Okay. You didn't include the United States in this. Is there any reason?

Koshiba:

Well, because I was very much interested in an e+e- collider.

DeVorkin:

And that work was not going on in the United States.

Koshiba:

Well, there was activity, of course, at Stanford.

DeVorkin:

Oh, at the linear accelerator. Yes.

Koshiba:

Yes. There was a colliding machine there.

DeVorkin:

Yes, that's right, okay.

Koshiba:

But because of Professor Lohrmann, who was my group member in Chicago —

DeVorkin:

Lohrmann. Okay. Because of him you went to DESY and then decided that that was a good collaboration. Okay.

Koshiba:

And then we started. I did obtain the research funds for this collaboration, made all the arrangements at DESY, but since I did not contribute to this experiment I didn't put my name on the paper at all. I sent my younger people, like Totsuka.

DeVorkin:

Oh, I see, okay. So people in the group were there, just not your name. Did your people find that unusual or other people find that unusual to —?

Koshiba:

Well, many people expected that I would include my name in the collaborator list, but I specifically said no, my name was out.

DeVorkin:

And why was that? Why did you decide —?

Koshiba:

Because what I did was only to make the arrangements for the experiment. If I designed part of the detectors, then I wouldn't mind putting my name on it.

DeVorkin:

But in arranging for the experiment, you had a particular goal in mind.

Koshiba:

Yes. But, maybe a later example will explain my attitude.

DeVorkin:

Well, I'm very interested in it, especially because of the famous example of Jocelyn Bell and Anthony Hewish and the discovery of pulsars. Many people said that she should have been given the credit, but other people argued no, because she hadn't built the entire facility and created the infrastructure. So I'll be very interested as you describe your attitude and your philosophy.

Koshiba:

Probably it will become gradually clear if I go on to explain the attitude I took in the following experiment.

DeVorkin:

Good. Please do.

Koshiba:

Now, this DASP experiment was lucky in the sense that not many people were interested in the e+e- experiment. Because many people thought that: "oh, they just have electromagnetic interaction, nothing new will come out." However, only a few months after we started this DASP operation, there was a discovery by the SPEAR people at Stanford and also the discovery of the J particle at Brookhaven, which was really a new era of particle physics. So DASP immediately changed the beam energy, and followed up. They discovered a new state too. So I was very happy about the DASP experiment, even though I sent only three of my young people there.

DeVorkin:

How many did you have in total? How many students and assistants by then?

Koshiba:

About six.

DeVorkin:

And you sent half of them there.

Koshiba:

Mm-hmm [affirmative]. In international collaboration we can send only our most advanced students. It takes too much time to take care of a first year graduate student.

DeVorkin:

Sure. About the international collaborations, your university of course was providing manpower and funding —

Koshiba:

Mm-hmm [affirmative]. And part of the equipment.

DeVorkin:

And part of the equipment. But did it also make it possible for you to apply to international sources of funding?

Koshiba:

What international sources?

DeVorkin:

UNESCO, for example?

Koshiba:

Never thought of it.

DeVorkin:

Never did that. Okay. Just a query.

Koshiba:

The next development was that DESY was building a bigger colliding machine, called PETRA.

DeVorkin:

PETRA. Yes, I've heard of that. There are so many of these.

Koshiba:

Oh yes. Through the experiment, DASP, we made a number of very good German friends, like Professor Heintze of Heidelberg University.

DeVorkin:

Heintze. Okay.

Koshiba:

He is a good physicist; his group was very good in buildng what are called drift chambers, big drift chambers. So in proposing an experiment to PETRA, we joined Heintze's group. Together we designed the experiment detector, using what is called modular lead glass, an electromagnetic detector. Lead glass, Pb.

DeVorkin:

Lead glass.

Koshiba:

This type of detector came from the student experiment I gave to one of the second year graduate students back in Tokyo. I asked him to make a modular type lead glass bar next to each other to form a lattice so that you can get the position and the energy in every direction. And then this idea was developed to form this lead glass barrel.

DeVorkin:

Barrel?

Koshiba:

Barrel. And this detector was very useful and very accurate. This same type of detector, a lead glass modular type detector, was also used in the next experiment at CERN using what is called LEP. So this PETRA experiment called JADE, which is short for Japan-Deutschland-England. In later years an American group participated, it gave the "A".

DeVorkin:

The "A" was in there. So it was originally JDE.

Koshiba:

No, J-A-D-E, JApan-Deutschland-England. JADE did produce a good result in observing what is called a 3-jet event.

DeVorkin:

3-jet?

Koshiba:

Mmm [affirmative]. Implying the existence of a gluon. I think it was last year that the European Physical Society gave a special prize to those experimental groups at PETRA responsible for finding gluons.

DeVorkin:

And this experiment, JADE, was the first to find evidence that something with the properties of gluons existed.

Koshiba:

Yes.

DeVorkin:

Because there are many kinds of gluons.

Koshiba:

Yes.

DeVorkin:

I mean, there are so many it's —

Koshiba:

Nobody has ever seen a naked gluon.

DeVorkin:

Right.

Koshiba:

But I was happy with JADE. Since this lead glass detector originally was my idea, I allowed my name to be put on the papers. But only the first five or six.

DeVorkin:

There were quite a few papers then coming from this.

Koshiba:

Yes, exactly. In the beginning the first five or six papers contain my name.

DeVorkin:

This was a very important discovery. I would imagine that your reputation by then must have been pretty considerable.

Koshiba:

Well, I don't know myself, but after about six papers I told my people to delete my name.

DeVorkin:

Did they protest?

Koshiba:

To some extent yes, but I was stubborn. [laughs] And the same thing happened at the LEP experiment which is called OPAL.

DeVorkin:

OPAL, yes. All of these are acronyms I understand.

Koshiba:

Yes.

DeVorkin:

Yes. But I've heard of them, and they're in the literature, are they not? I mean, I know you've talked about them in your, some of them, in your history. The article you gave me yesterday on the observational neutrino in astrophysics, your beginning preliminaries on elementary particle physics leading up to the grand unified theory, there were a lot of elements in there that I saw explained for the first time. Very good. So then you joined OPAL?

Koshiba:

Yes. I allowed my name to be put on the collaborators list on the first six or seven papers, but not after.

DeVorkin:

How many authors were there on some of these papers?

Koshiba:

In the case of OPAL, altogether about 350 or something like that.

DeVorkin:

Three hundred and fifty. [laughs]

Koshiba:

Well, that is not my type of experiment. As I said, when a student is well advanced to the late years of graduate school, you can send those students to write doctoral thesis at one of the international corporations. Assistant professors, associate professors and assistants I sent to the site can take care of the thesis-making students. However, what do we do with the undergraduate physics student or the student in the first three or four years of graduate study? Just give lectures, or do the laboratory work? I like to have something I can bring to the frontiers of physics — first or second year graduate students or even undergraduates. I want to have something besides this big international collaboration.

DeVorkin:

Mm-hmm [affirmative]. So you needed something that you had control over. So KEK wouldn't have helped, Tokuba and any of the national facilities, they wouldn't have worked for you. That's correct.

Koshiba:

First of all those accelerator experiments became too routine. In order to attract the attention of younger people I thought we needed something different.

DeVorkin:

Okay. And as you said, this is when you started thinking about Kamioka. Now —

Koshiba:

I wasn't specifically thinking of the Kamioka experiment, but I was feeling the need for some experiment in the country where I could use and train young graduate students and undergraduates too. And that was the time when I received a telephone call from Sugawara on the possibility of doing a proton decay experiment.

DeVorkin:

Sugawara was at KEK?

Koshiba:

Yes. He was the head of the theoretical division there.

DeVorkin:

Yes. And he asked you to think about how you could test the grand unified theory, the GUT. Okay. At that time, did you know about the IMB collaboration —?

Koshiba:

Not at that time. You see, that was, oh, I don't remember exactly, it was something like December of 1978. I didn't know about the existence of the IMB project at that time, but when I received the telephone call from Sugawara I immediately remembered the chat with Occhialini on the use of a deep underground lake in shielding the photomultiplier pointing downward.

DeVorkin:

You didn't talk about that in the interview; you talked about that last night. Maybe it would be good for you to re-cover that now.

Koshiba:

I see. Well, I will make it short.

DeVorkin:

No, give me the whole thing.

Koshiba:

Well, as I said, when Marcel Schein died, the University of Chicago asked Professor Occhialini to be an advisor, and he recommended me as the successor of Marcel Schein for the whole project. He used to come to my place every month, staying several days, and we used to chat over beers on almost everything. And since we were storing the big stack in a salt mine cave near Cleveland —

DeVorkin:

This was a salt mine you knew about, or —?

Koshiba:

We investigated all the possibilities near Chicago.

DeVorkin:

Yes. This was how to keep the stack unexposed.

Koshiba:

Shielded against cosmic rays.

DeVorkin:

Right.

Koshiba:

And then I said to Beppo Occhialinni, this place is deep enough, and since it is salt mine, if you make a pool-size hole in the ground and pour water in, it will make the saturated salt water; after a few days all the dirt will settle to the bottom and you get very clear, transparent salt water. Since it is saturated, there will be no bacteria or no plant growth. It will stay very clean and transparent for years. What would we observe if we could install a large number of phototubes facing downward? You know that cosmic ray muons do come down to that depth, but the Cherenkov light they produce would be pointing downward. So if you had photomultipliers on the surface facing downward, you wouldn't see those downward-going muons.

DeVorkin:

The light is directed downward.

Koshiba:

Yes.

DeVorkin:

You wouldn't see it from the back of the envelope?

Koshiba:

Well, you would observe some scattered light. Also, at that time, the photomultiplier was very small, expensive, and not very dependable.

DeVorkin:

This was in the early 1960s. Right.

Koshiba:

So, we couldn't do it, but when I received this call from Sugawara since the most favored mode of proton decay by this SU(5) in question was into positron and pi-zero [0], and pi-zero immediately decays into two gammas, and those gammas and positrons produce an electromagnetic shower in the water. What you observe is back-to-back generated showers. Therefore, if you had a photosensitive device on the inner surface, you would observe this Cherenkov light immediately, and there will be no question about the nature of the event. So within about one hour after the phone call from Sugawara I took a piece of paper, drew the detector and photomultiplier installation, and since I had another appointment on the symposium day, I sent my assistant to present this idea to the KEK symposium.

DeVorkin:

At the KEK, where Sugawara would be.

Koshiba:

Yes. Sugawara was presiding over the symposium. By the way, the assistant is now a full professor at the Tokyo Institute of Technology. There were competing proposals from other groups in Japan, but finally our proposal was approved, and we immediately started. That was in December '78. The next month, in January, I received a copy of the IMB proposal.

DeVorkin:

I see.

Koshiba:

Then I had to start thinking, very seriously, since they were planning a much bigger detector. And they had much more money.

DeVorkin:

Did your money come from KEK?

Koshiba:

No, it came through the ministry of education.

DeVorkin:

But was this a proposal approved by KEK or what?

Koshiba:

Not by KEK, but through the University of Tokyo, faculty of science.

DeVorkin:

Okay. Did you have to do any politicking or pushing?

Koshiba:

Well, the most difficult part was to get a higher ranking in the faculty of science budget process. There are projects which have been proposed for years occupying the first rank, the second, and so forth. If I had to wait at the bottom of the line, it would easily take ten years. So I proposed to the committee of the faculty of science to put my proposal outside the regular ranking. There's a condition that even if my project was approved, it wouldn't affect the acceptance of the usual list. I wouldn't disturb the other people.

DeVorkin:

So you are asking to make the pie bigger rather than to get the first piece of the piece.

Koshiba:

Yes, that's right.

DeVorkin:

I see. Who did you have to convince to do that?

Koshiba:

Well, first the faculty of science committee members and the dean of the faculty of science.

DeVorkin:

And this was knowing that IMB was going to be a competitor?

Koshiba:

Yes.

DeVorkin:

Did that help?

Koshiba:

Well, those people were not very much interested in IMB. They were only interested in whether their list would be affected by the addition of my proposal. And luckily I could convince Mombusho, the Ministry of Education, that all I needed was something on the order of 200 million yen — which in the exchange rate of those days was about $600,000. I knew that IMB had $3½ million. So I had to think very seriously, how could we compete in the proton decay work with IMB? Okay? If our aim was only to detect a proton decaying into a positron and a π0, we didn't have a chance, because they had a much bigger volume and there was no difficulty, even for IMB people, to identify the decay into a positron and a π0. Then I thought, well, SU(5) is not the only candidate for the grand unified theory. There are other possibilities too. What would be the useful result for identifying the future direction beyond the standard theory? I thought, not only the detection of π0e+, if we could also observe other decay modes like K+ anti-neutrino, or mu+ + π0, and so forth. And if we could measure the branching ratios into those various decay modes, that would constrain the type of grand unified theory very much. Then, in order to detect such other possible decay modes, your detector has to be very sensitive in the following sense: we can observe only the Cherenkov light, the secondary particle produced in the water. So ideally if all the inner surface is covered by photocathodes, that would be ideal. So how can we approach that limit within the limited amount of funding? That was the reason why I came up with the idea of developing very large phototubes. Then, using the same number of channels, like one thousand channels, this big phototube will give very much more sensitivity than the smaller phototube.

DeVorkin:

Now a lot must have happened in Japanese electronics between 1960 and 1978 to allow you to think in these ways. Is this correct?

Koshiba:

Well, I don't know if you know this, but this Hamamatsu photonics —

DeVorkin:

Hamamatsu, yes.

Koshiba:

— used to produced phototubes for RCA.

DeVorkin:

Oh really?

Koshiba:

Yes. In those days they just produced them according to the specifications of RCA. And they were sold in the name of RCA.

DeVorkin:

I didn't know that.

Koshiba:

But when we started the DASP and JADE experiments, we needed thousands of phototubes. So we contacted this Hamamatsu photonics outfit, and we specified what characteristics were important in this particular type of experiment, and so forth. They responded very nicely, and that was the reason why I asked this company to try producing 20-inch phototubes.

DeVorkin:

So you had previous experience and contact with Hamamatsu.

Koshiba:

Yes.

DeVorkin:

Was there also an element there that a company like Hamamatsu wanted to make its name in something that was really dramatic?

Koshiba:

Yes. Luckily, the president of this company wanted to do something different, something new. He was very aggressive in his field.

DeVorkin:

What was his name?

Koshiba:

Hiruma. A very nice man.

DeVorkin:

It's very important to understand how this technology developed to the point where you could think of such large phototubes. Can you give me some sense of how you felt the technology was ready for it? What other types of tubes did you know existed that compared with this, what you had in mind?

Koshiba:

Well, I knew that this company had experience with tubes up to something like 12-inches in diameter. We had a rather long discussion with the president and the chief of the technical division of that company in my office at Tokyo University. The chief of the technical division was quite reluctant to go beyond anything bigger than 14 inches in diameter, because that was the size he felt was safely approachable. But I insisted on 20 inches, that was the size I really wanted.

DeVorkin:

What were the technical limitations?

Koshiba:

One was to blow that big a glass tube. Making one or two is not a problem, but to make the same shape, same thickness, and the same strength in this size phototube is not. The glass tubes were not easy to make.

DeVorkin:

Standardization is very important.

Koshiba:

Yes.

DeVorkin:

Well, let me just ask this out of pure ignorance. Television tubes are large vacuum tubes. Are we talking about a tube where the tolerances on the glass, or the degree of the vacuum, is much greater than a regular television tube?

Koshiba:

One thing was of course the degree of vacuum. More important, however, is the fact that our big phototube had to be immersed in deep water, which means additional pressure. In the case of a television tube, the pressure difference is only one atmosphere.

DeVorkin:

Right.

Koshiba:

But in the case of say a 40-meter depth in water, there would be a five atmosphere difference. So the glass has to be very strong.

DeVorkin:

Was the thickness of the glass also a factor? Did it have to be —?

Koshiba:

Of course. You have to choose a correct thickness to withstand the pressure difference.

DeVorkin:

Okay. So the glass thickness was a function of the mechanical stability of the tube. There was no problem making it too thick for the nature of the experiment. No problem there, because it's visible light.

Koshiba:

Mm-hmm [affirmative]. However, if you make it too thick, several problems arise. In the case of thick glass, it's not easy to make things uniform. That's one of the problems. And another thing is that even though it is transparent to visible light, the Cherenkov light contains a short wave component too. A more important difficulty with the large size is the fact that unless you design the shape of the photocathode and the shape of dynode very carefully, the arrival time of photoelectrons from various parts of the photocathode will be quite different, making the time resolution of the tube very bad. Because at this low voltage difference, the slow electron has to travel a long distance. So if the electric field is not very well conditioned, different electrons will take different transit times, and you would lose the time resolution.

DeVorkin:

We're talking about extremely precise timing.

Koshiba:

Mm-hmm [affirmative]. In order to help design this large tube, including the dynode structure and so forth, I sent my research assistant and one graduate student to Hamamatsu photonics to help design and test the tube.

DeVorkin:

And who were they?

Koshiba:

The assistant was Atsuto Suzuki, who is now a professor at Tohoko University. He is now starting a new project at the old Kamioka Kamiokande site using large amounts of liquid scintillator.

DeVorkin:

The water or the oil?

Koshiba:

Oil. The graduate student is now a full professor at UCLA.

DeVorkin:

Oh really? Who is he?

Koshiba:

Arisaka.

DeVorkin:

Okay.

Koshiba:

He had just entered graduate school, and when I told him about the Kamiokande project, he was very interested and became enthusiastic. The first thing he did was to help design these big phototubes, was the first collaborator of mine on the Kamiokande experiment. Totsuka and others joined later.

DeVorkin:

Oh, I see. But they were still part of your group?

Koshiba:

Totsuka at that time was working on JADE in Germany.

DeVorkin:

Oh, I see, okay. And he came back to work —

Koshiba:

Mmm [affirmative].

DeVorkin:

Was that actually a general characteristic? I mean you had a number of students who were out working on different projects, but when this one, when you built Kamiokande you brought them back?

Koshiba:

Mm-hmm [affirmative]. Well, I never told them to do this or that. I explained what the possibilities were, and then let them decide.

DeVorkin:

That's good.

Koshiba:

I never forced them.

DeVorkin:

Okay. Well, the phototubes were the first big technical hurdle. When were you quite sure that you had the phototubes in hand, that they would work? How were they tested?

Koshiba:

Well, first we made a 1 meter cube water container, filled it with water, immersed the phototube in it, and used cosmic rays muons to see the response.

DeVorkin:

And so that was the method of testing.

Koshiba:

Mmm [affirmative]. And also we used the secondary beam of the KEK accelerator.

DeVorkin:

So you took them there. Okay.

Koshiba:

Yes. Specifically to check the timing characteristics.

DeVorkin:

Did anybody else get interested in your phototubes while you were testing them? Other people must have seen them, seen how big they were.

Koshiba:

Yes. For instance IMB people, some of the IMB people did show interest, but they had already committed to the 5-inch tubes.

DeVorkin:

The 5-inch tubes, yes. What contact did you have with Fred Reines at this time?

Koshiba:

I invited him to come and visit Kamioka, and he invited me to Irvine. Fred and I got along very well.

DeVorkin:

Okay, good. But you were essentially in competition.

Koshiba:

Yes.

DeVorkin:

Now there's many aspects to this project that I would like to better understand. We've talked a bit about the funding, how you did that, but there is the question of the logistics of the experiment, how you organized it. We've talked a bit about the detectors, but not about the system of collecting the data. I guess I need some word from you on what technical and logistical milestones did you have to meet to build it, what constraints were you under. You didn't have that much money, as you indicated. So the phototubes were the first big technical hurdle. What was the next one that you had to overcome?

Koshiba:

Well, the second problem I had to overcome was the fact that our funding was not quite sufficient. Since I decided to maximize photon sensitivity by developing big phototubes, I used much of my funding in this area. But, when it came to the electronics I knew that the timing electronics, what is called TDC, Time to Digital Converter, would be nice to have. I couldn't afford it. So I used only ADC, Analog to Digital Converter. In order to measure pulse height. So, because the total number of channels was about one thousand.

DeVorkin:

You had a thousand phototubes in the first instrument.

Koshiba:

Not only that, I made a very detailed cost breakdown of 20" phototubes. The company wanted to charge something like 200,000 yen per tube. But I worked it out that the cost of making one tube would be something like 120,000 yen. I gave them 130,000 yen per tube. They didn't like it, but since we were the only customer of these tubes they had to swallow that price.

DeVorkin:

This is after they had made the prototypes?

Koshiba:

Mm-hmm [affirmative]. So I saved a considerable amount there, but still I could afford only ADC, not TDC.

DeVorkin:

Getting back just for a second to the tubes themselves, I understand from Professor Suzuki that the critical factor was the ability to blow the glass. I know that you said you weren't too much involved with the glass part of it, but did you know that there were only two glass blowers, or were there only two glass blowers, who could do this work?

Koshiba:

In Japan you mean?

DeVorkin:

Or in Hamamatsu. Were they already in Hamamatsu, or did Hamamatsu have to hire them?

Koshiba:

I don't know exactly. You'd better ask Atsuto Suzuki, because he is the one directly involved.

DeVorkin:

Okay.

Koshiba:

There was another thing I worried about, namely, the the background contamination of the glass.

DeVorkin:

The glass itself having material in it that was radioactive. Yes.

Koshiba:

That I had checked by my student and assistant. I told him, "Don't forget to check this."

DeVorkin:

Okay. So you were planning out the costing. You had to make a compromise in the electronics for the TDC. What about the siting, where to put the detector? You didn't have any salt mines. So where were you going to put it?

Koshiba:

We had to dig a new cave, and for this I had to obtain additional funding.

DeVorkin:

Oh, I see.

Koshiba:

I got it through KEK.

DeVorkin:

Did they know that you didn't ask for enough money in the beginning?

Koshiba:

Initially, the minister of education did give me the instrumentation money, but not the digging money.

DeVorkin:

Oh, so that was —

Koshiba:

Later, the minister of education gave me the digging money upon the recommendation of the KEK director general.

DeVorkin:

Was the director general a friend of yours?

Koshiba:

Not the director general himself. At that time, the director general was Nishikawa, who was three years senior to me. But Sugawara, head of the theoretical division, made a strong recommendation, because we already had produced successful 20-inch phototubes.

DeVorkin:

What kind of survey did you do to find the best site?

Koshiba:

Oh, the people in the mining company were very experienced, they made a test hole, and examined the composition of the rock and also the shape of the cave had to be taken into account. We consulted with a professor in the engineering faculty who is an expert on this kind of thing. Upon his advice, we decided on the shape of the cave. Luckily the cave is located in the oldest rock formation on the entire Japan island. It is very old, very solid rock, necessitating the use of explosives, not a jack-hammer or anything.

DeVorkin:

So this is a relatively stable area, geologically speaking.

Koshiba:

Yes.

DeVorkin:

Did you consider other sites other than the Kamioka area?

Koshiba:

Yes, I did, but I picked this place. One reason was that I had experience with this company over many years — dating from my first underground experiment.

DeVorkin:

The muon experiment. Right.

Koshiba:

Of the other places I investigated, one was a copper mine, but the water inside the mine contained a lot of sulphur which is deteriorating to the iron container wall. That was one reason for rejecting this site. I also investigated an under-ocean tunnel between Hokkaido and the mainland, but it was too narrow and too shallow.

DeVorkin:

You mean there wasn't enough water.

Koshiba:

The mine I chose not only has very solid, stable rock, but also the miners there are very experienced. There can be a lot of dangerous situations in a deep mine.

DeVorkin:

Oh yes.

Koshiba:

And the company has been friendly to me from the very beginning. When I carried out my first experiment here, the company was booming, was very prosperous, and making lots of money. During the last 20 years, however, they ran into some difficulties. People downstream in the Toyama area claimed that because of the poisonous water, polluted by the mine, people there suffered from a special type of disease.

DeVorkin:

It was cadmium, wasn't it?

Koshiba:

Cadmium, yes.

DeVorkin:

Cadmium poisoning, yes.

Koshiba:

So finally the company had to pay something like $1 billion every year to those people. Combined with the fact that the world price of lead has been low, the company has been in difficulty for many years. Further, the number of employees has been reduced by a very large factor.

DeVorkin:

What is the name of the particular company here?

Koshiba:

The Kamioka Mine.

DeVorkin:

The Kamioka Mine. Okay. And they're part of Mitsui?

Koshiba:

Yes.

DeVorkin:

So the company's name is the Kamioka Mine. They've been here for years and years. Actually how long have they been here as the Kamioka Mine?

Koshiba:

I think Mitsui obtained this mine at the beginning of the 1850's or the 1860's.

DeVorkin:

So the 1860s, or the 1850s-60s. Okay. But as we discussed at lunch, there has been mining in this area for over you said —

Koshiba:

One thousand years.

DeVorkin:

Over a thousand years. Okay. So then you decided on the Kamioka Mine, but then it was a question of finding the best spot. And you also had an engineer, or a professor of engineering at —

Koshiba:

To give me advice.

DeVorkin:

Yes, to give advice, with the mining engineers. And what was the final decision? How was it finally made?

Koshiba:

Well, we found a region of very solid rock which would be stable for some time to come. We made a number of test drillings to check the properties of the rock, and then started digging.

DeVorkin:

Was the chamber you were digging bigger than anything these miners had dug up to this point?

Koshiba:

Not quite. They did have a vacant cave which was bigger in size. They used it as a place for dumping sand and rock.

DeVorkin:

Oh, the stuff they didn't use. I know, yes, there's a word for it. Tailings.

Koshiba:

I didn't know that.

DeVorkin:

Yes. I think that's it. The material left over after the minerals have been harvested. But that was not an acceptable place?

Koshiba:

No, because first thing was safety. And then what came next? As I explained, after the cave was dug, we installed phototubes layer by layer using graduate students.

DeVorkin:

Mmm. But in making the tank itself, did that require any special technology?

Koshiba:

Not very much, but we had to worry about residual radioactivity. We lined the inner side of the tank with a coating to prevent water from leaking. Originally, our aim was to search for proton decays.

DeVorkin:

Yes.

Koshiba:

For that purpose we didn't need any anti-coincidence system, because the signal was large and very typical, that is back-to-back. Yeah? So we didn't worry about the background. So we installed the tank and we laid down the phototubes on the wall and on the bottom and top and filled it with filtered water.

DeVorkin:

Oh, and just to be clear, because I think I cut you off in the middle of your statement, you put detectors on the floor first, and then your graduate students went around connecting them as you filled more and more water.

Koshiba:

Mm-hmm [affirmative]. Storing one layer after another, gradually raising the level of water.

DeVorkin:

How long did this process take?

Koshiba:

I think we started the installation operation in January, and it was completed by the end of June. About half a year operation.

DeVorkin:

That would be in 19 —?

Koshiba:

'83.

DeVorkin:

'83, right. From the chronology I have found from your records, you started observations on the 4th of July, 1983. Right?

Koshiba:

Not exactly, because the 4th of July was the time when we started pouring water.

DeVorkin:

Oh, I see.

Koshiba:

So at the beginning there was only this depth, then this depth. We gradually increased it.

DeVorkin:

So the water that you'd put in there for your graduate students to float around in, you then got rid of it?

Koshiba:

Mm-hmm [affirmative]. Because that water would have been dirty.

DeVorkin:

Did they then have to clean all of the tubes?

Koshiba:

Yes, wash them.

DeVorkin:

I have this picture that you brought the water level down as they cleaned all the tubes, leaving everything above crystal clear.

Koshiba:

Yes. Mm-hmm [affirmative].

DeVorkin:

What a job.

Koshiba:

Yes. It was a big job.

DeVorkin:

Now you also knew of Rines's IMB.

Koshiba:

Mmm [affirmative].

DeVorkin:

And you didn't know how quickly they would be in operation, or did you?

Koshiba:

Well, I didn't have detailed information, but they were about one year ahead of me, because of the time we used to develop the 20-inch phototube. So we were one year behind. This I knew.

DeVorkin:

Were you afraid of being scooped?

Koshiba:

At the very beginning when it was only a matter of finding e+ + π0, IMB has very much of an advantage over Kamioka.

DeVorkin:

That's right. You explained, that you had different capabilities. And so you knew that you would be able to make a wider —

Koshiba:

Even if they discovered the e+ + π0 decay, eventually our detector would supplement the information by observing other types of decay modes. That would be rather important.

DeVorkin:

Yes. Were you working also in the same energy range as they were?

Koshiba:

Yes, proton decay produces the same amount of energy.

DeVorkin:

Okay then, from the chronology, you started pouring water back on the 4th of July. At this point there was something that I had read in one of your reviews that indicated that you began to realize that you could detect solar neutrinos about this time. Is that —?

Koshiba:

I think all the water was back in sometime in early August, and then we started full data taking. I wanted to make the energy calibration. What you observe is the number of photons in each phototube. You have to convert this to the actual energy of the event. For this purpose, I used the cosmic ray muons stopping in the water detector, and then decaying into an electron. We know the energy spectrum of decay electrons accurately. Therefore if you observe the number of photons from decay electrons, you can calibrate the number of photons to the energy of the event. After about three months of operation, we had accumulated a considerable number of decay electrons.

DeVorkin:

This is by the end of 1983?

Koshiba:

Sometime around the end of October. We could see a beautiful distribution of decay electron energy distribution down to an energy of about 12 MeV. Below that, there is a very sharp rise due to environmental background. It may be interesting to tell you that during the three months of operation we found a very attractive event which looked like just proton decaying into mu+ and [eta]0, decaying into 2 gammas. This is a neutral particle, just like π0, only eta0 is more massive than π0.

DeVorkin:

Is that the symbol?

Koshiba:

Yes, that's the symbol.

DeVorkin:

Yes. Eta zero. Okay.

Koshiba:

Which then decays into 2 gamma rays.

DeVorkin:

Into 2 gamma rays?

Koshiba:

A photon decaying into mu+ and eta zero. We were very much excited about this event. You probably know the expression "beginner's luck." That was exactly the case. We had been searching for a similar event for more than ten years with Kamiokande and also with the Super-Kamiokande. We didn't find it.

DeVorkin:

So you don't really know —?

Koshiba:

We don't claim it was a proton decay, not only because it was the only event, but also if it were a proton decay it must have occurred inside an oxygen nucleus. The total momentum didn't exactly balance.

DeVorkin:

This is what you understand from theory.

Koshiba:

No. From the observation.

DeVorkin:

From the observation.

Koshiba:

Yes. The total momentum was not quite on a flat surface, but it was a little bit like that.

DeVorkin:

It was asymmetric.

Koshiba:

It was not on a plane. If it were a 3 body decay, it should be on the same plane.

DeVorkin:

Oh, all three tracks have to be on the same plane.

Koshiba:

Yes. Coplanar.

DeVorkin:

Coplanar. But it wasn't.

Koshiba:

It was not quite coplanar.

DeVorkin:

Indicating that something else was going on.

Koshiba:

Yes. There was something which absorbed the excess momentum. Forget about this, because we are not claiming it as a proton decay.

DeVorkin:

But it was exciting.

Koshiba:

Yes. And then I thought, if we can somehow reduce the background at the low energy end below 12 MeV, then there is a possibility of observing solar neutrinos by means of neutrino-electron scattering. Because we know that the boron component of solar neutrinos does have an energy in excess of 14 MeV.

DeVorkin:

Boron 8?

Koshiba:

The Boron 8 decay. Yes.

DeVorkin:

Were you reading the literature at this time and realized that there was a problem with detecting solar neutrinos? Did you know about Davis' work at this time?

Koshiba:

Of course.

DeVorkin:

I had to ask. [laughs] And you knew that there was a deviation from prediction.

Koshiba:

Mm-hmm [affirmative]. Only one-third is observed of the predicted value.

DeVorkin:

Right.

Koshiba:

Of course I knew that. I admire Davis' work very much. However, it all depends on how much you can reduce the background. To do this we had to install additional layers of anti-coincidence detectors.

DeVorkin:

When did you actually decide to search for solar neutrinos? Was it before you met with Alfred Mann at Utah in January of '84?

Koshiba:

Somehow we had to do it. But how. Because I had already exhausted my funding. When I thought of how to go about it, I had to install additional layers of anti-counters as a first step. Also, because those events do not give very many photons, the timing device, TDC, was necessary to reconstruct the event.

DeVorkin:

That's what you needed.

Koshiba:

Yes. But as I said, I didn't have money to buy a TDC. The trouble is the minister of education. Once an experiment is approved, it is completed as proposed. In our case, the experimentalists found new possibilities, and we wanted to change the setup. But, the minister of education wouldn't give us the money.

DeVorkin:

Why?

Koshiba:

Because a project has to be thought out thoroughly before proposing it. Changing the structure of the detector and so forth, after only three months of operation, would not be permitted.

DeVorkin:

The predictions of GUT, the grand unified theories, had originally said that proton decay should occur, and that it should have a lifetime of 10 to the 28th or 29th years.

Koshiba:

29th, yes.

DeVorkin:

But at some point, just as you were building Kamiakande, the first Kamiokande, that estimate was shifted up to 10 to the 33rd.

Koshiba:

Not quite.

DeVorkin:

When did that happen or —?

Koshiba:

Well, it was only after IMB and Kamiokande didn't find the event. The experimentalists were the ones who pushed up the lifetime.

DeVorkin:

How long had IMB been operating by the time that was the case? They were operating before you were.

Koshiba:

Mmm [affirmative].

DeVorkin:

And then you were only operating for three months.

Koshiba:

Mmm [affirmative].

DeVorkin:

Had you thought that you should have found something within the first three months?

Koshiba:

Yes. At that time, because of this particular event, I wasn't sure whether proton decay was occurring or not.

DeVorkin:

Aha. Okay.

Koshiba:

We didn't make any statement.

DeVorkin:

Oh, you didn't tell anybody.

Koshiba:

We showed the event, but we didn't claim that this was a proton decay.

DeVorkin:

Okay. But, I guess what I'm trying to clarify is this, if the experimentalists showed that there were no proton decays, that would require a certain amount of observing time to be definitive, and yet you were only observing for three months. So was that enough time?

Koshiba:

No. In order to set the low limit for the lifetime, three months operation can give a very poor lower limit, while IMB can give a better lower limit. Yeah? But when it comes to a positive result, that is finding something, it doesn't matter whether it is only three months operation or three years operation.

DeVorkin:

Right. So you had not given up on detecting proton decay?

Koshiba:

No. I haven't given up the possibility of detecting proton decay, even now. I know that the theoretical people have been pushing up the limit after we failed to find anything, but at some point the proton has to decay. This is my personal belief the lifetime is 1035 years or 1036 years.

DeVorkin:

So is it correct to say that even though you may be not too concerned about the details of the theory, you basically are working within the general framework of the theory that says protons should be decaying?

Koshiba:

Yes.

DeVorkin:

Okay. So, within that three months, you decided that there was another use for your instrument.

Koshiba:

Mm-hmm [affirmative]. I figured that I had to install additional anti-coincidence layers, and I had to obtain TDC’s, for thousand channels. As for installing the additional anti-coincidence layers, I could use graduate student labor and some scraped-up funds for an additional small number of phototubes. But when it came to TDC’s, for thousand channels, I didn't have any idea of how to get it.

DeVorkin:

Mm-hmm [affirmative]. It was a question of money. Were these things commercially available?

Koshiba:

Yes.

DeVorkin:

Yes, alright. Because I know ADCs are commercially available.

Koshiba:

And then I went to the Park City meeting in January next year.

DeVorkin:

That was January of 1984.

Koshiba:

Mmm [affirmative].

DeVorkin:

Okay. And that's in Utah.

Koshiba:

Yes.

DeVorkin:

Okay.

Koshiba:

At that time three underground experiments were ongoing. One was of course IMB. The second involved another American vintage detector by Carlo Rubbia and David Cline. I don't remember the name of that experiment.

DeVorkin:

But it's in your review paper.

Koshiba:

Cline is now at UCLA.

DeVorkin:

So there were three underwater detectors and probably representatives from all three were at Park City? You mentioned that you were also planning to present three papers at the symposium.

Koshiba:

One was to report on the preliminary results of the Kamiokande detector. I showed the 3-ring event, the 2-ring event, the 5-ring event, and so forth.

DeVorkin:

Ring event?

Koshiba:

Yes. One particle produces a Cherenkov ring.

DeVorkin:

Okay. So by then you'd had a family of events that you had seen, a series of rings.

Koshiba:

Yes.

DeVorkin:

Which said that something was happening, your detector was working. Okay.

Koshiba:

I was glad that the audiences were very much impressed by actually looking at the Cherenkov rings. I showed this expanded view which you saw on the real time display of the event.

DeVorkin:

Right. It's an equatorial projection of the cylinder plus the polar.

Koshiba:

That's right. Exactly the same display.

DeVorkin:

You didn't have the ability in real time, but you had the ability to do it in analysis.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Yes. Of the three underwater detectors, was yours the only one that could really do imaging like that?

Koshiba:

I think so, yes.

DeVorkin:

Or at least with a resolution and show the rings?

Koshiba:

There is an IMB report which didn't show this actual ring image and so forth. They were just talking about the statistically analyzed upper limit and lower limit and so forth. And when I showed this actual event display, people were very much interested and started shouting that IMB should show a similar event display. And they of course had to take some time to produce it to show their version of the event. But unfortunately, because their sensitivity is only one-sixteenth of the Kamioka experiment, what you saw was scattered crosses.

DeVorkin:

Oh, their resolution was only one-sixteenth. Not sensitivity, but resolution.

Koshiba:

Well, not resolution. Resolution is one fourth, sensitivity is one-sixteenth.

DeVorkin:

So you have 16 times the resolution of the original IMB?

Koshiba:

No.

DeVorkin:

Okay. Do I have that wrong?

Koshiba:

The resolution would be the accuracy.

DeVorkin:

Spatial resolution.

Koshiba:

Not only the spatial, but also the amount of energy. The energy estimation would be proportional to the square root of the total number of photon events. Right?

DeVorkin:

Yes.

Koshiba:

In this sense we were observing 16 times more photons as compared to IMB. Therefore, our accuracy was four times greater — the square root of 1 over 16, or 1 over 4. In my second paper, I wanted to get my American collaborators to work on the problem of solar neutrino detection, and possibly bring TDC’s to the Kamiokande site.

DeVorkin:

Yes.

Koshiba:

Unfortunately, however, I had a very bad case of the flu.

DeVorkin:

Flu.

Koshiba:

Yes. My throat was swollen, it hurt like hell, and I couldn't speak.

DeVorkin:

Oh. So you had laryngitis.

Koshiba:

I didn't know what it was. But I talked to Al Mann because after I had shown him the Cherenkov ring patterns, he complimented me on my work. So I said to him, "Are you interested in trying to observe solar neutrinos by this type of detector and through electron scattering?" He said, "yes", and and then we started planning on how we could go into it together.

DeVorkin:

But with your flu —

Koshiba:

I could speak only with a low voice when we were talking —there were only two of us — but not in a manner to make a speech.

DeVorkin:

So what happened?

Koshiba:

So I made a transparency, three transparencies for each topic, and I asked Al Mann to present them for me. One was the feasibility study of solar neutrinos.

DeVorkin:

You had already written that out.

Koshiba:

Mmm [affirmative], with transparency paper. I had written the transparency and gave it to Al Mann to show it to the audience.

DeVorkin:

Okay.

Koshiba:

What I wanted to clarify was that there was a definite possibility of observing solar neutrinos with the scattering of electrons in water. In this way we can not only observe the total flux, get directional information, and also obtain energy information. That means we can make astrophysical observations by neutrinos. I said a few minutes ago that I have great respect for Davis' work.

DeVorkin:

For Davis' work. Yes.

Koshiba:

However, I wouldn't call it a breakthrough in neutrino astrophysics observation, because with his method you don't have time information, directional information, nor energy information.

DeVorkin:

Did you make that statement at Park City?

Koshiba:

Mmm [affirmative].

DeVorkin:

How did people react to it?

Koshiba:

Well, I'll tell you. To observe solar neutrinos by this method, I described how we needed to install additional anti-counter layers, and also to improve the event reconstruction accuracy we needed one thousand channels of TDC. Many people were interested; however, Al Mann was the one who actually was very eager to try out this possibility. For instance, Dave Cline and the IMB people were all very much involved in their own experiments. Another paper I asked Al to present for me was the possibility of a new U.S.-Japan collaboration — a Super-Kamiokande.

DeVorkin:

Even at that time.

Koshiba:

Uh-huh [affirmative]. And I considered a nickname for this. I called it JACK, J-A-C-K, which means Japan-America Collaboration at Kamioka. I figured it to cost about $100 million, but nobody showed interest.

DeVorkin:

What was your argument for the value of building the Super-K at that time?

Koshiba:

Well, first of all the possibility of observing solar neutrinos through electron scattering was very good. This was actually opening up a new field of astrophysics, neutrino astrophysics. So I felt that this would justify $100 million. That's what I thought. But others didn't think so.

DeVorkin:

Were there people like John Bahcall at the meeting?

Koshiba:

I don't remember.

DeVorkin:

I was just wondering who first —

Koshiba:

The organizer of the conference was David Cline.

DeVorkin:

Okay. But —

Koshiba:

David Cline promised to produce a proceeding, but he never did.

DeVorkin:

So this was never published.

Koshiba:

That's right.

DeVorkin:

Do you still have the manuscripts that you gave?

Koshiba:

I have the transparencies somewhere.

DeVorkin:

Good. I hope that your papers are being preserved.

Koshiba:

Yes, but it's a handwritten transparency.

DeVorkin:

That's okay. Those things are very important. But as well as your letters and your correspondence are the different records that help future historians understand your career. Have you deposited your papers at the University of Tokyo or anywhere?

Koshiba:

No. Because I believed that David Cline would produce the proceedings.

DeVorkin:

No, I'm talking about all the papers produced throughout your entire career. Not published papers, but your private correspondence, letters, and things like that.

Koshiba:

Well, I don't accumulate such papers. From time to time I throw them away.

DeVorkin:

Oh, you do? Oh, that's too bad. Because that's the basic data for historians.

Koshiba:

I'm sorry. But I am not very good in keeping things.

DeVorkin:

Mmm. But what about when you have grants from the minister of education, do they keep records, do you keep records of your proposals to them, and reports to them, things like that?

Koshiba:

Aahhhh, the minister of education should have kept the proposals somewhere.

DeVorkin:

Okay. Well, we'll talk about this later. I don't want to derail you. So you gave these three talks at Park City, and from there Alfred Mann became very interested.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

How did he gain support for the timing, electronics and things like that?

Koshiba:

I don't know which agency he contacted. I don't know.

DeVorkin:

Okay.

Koshiba:

But he did say he wanted to join, and he brought Gene Beier with him, who was I think at that time an associate professor or something.

DeVorkin:

Okay. So the job was to close down Kamiokande, add the anti-counters, and the timing devices. Did you have a plan to do all this?

Koshiba:

Well, it took more than I hoped. First we emptied the water tank, and then dismantled the bottom part of the phototube array.

DeVorkin:

You had to take out all the phototubes?

Koshiba:

Not all. We left the barrel part intact. We took the bottom part. And then raised the floor by about 1.5 meters and made a new floor and then made this bottom 1.5 meters as a anti-counter layer. Yeah? So the tubes are installed on the new floor, and then at the bottom there were fewer number of phototubes facing outside. Yeah?

DeVorkin:

Right.

Koshiba:

On the side there is a space of about 1.5 meters to 2 meters between the iron wall of the container and the rock.

DeVorkin:

Oh, so you didn't have to build another wall.

Koshiba:

No.

DeVorkin:

The water is contained by the rock. So this water didn't have to be as pure as the water inside?

Koshiba:

Well, it has to be pure, but not as pure as the inner water. Because, after all, light travels the distances of only a few meters. But we did have to do some sort of water leakage prevention.

DeVorkin:

So you coated the rock with something.

Koshiba:

And then also we made an additional layer on top of the detector by lowering the top layer of phototubes. And then we restored 1 meter of additional water there.

DeVorkin:

Okay.

Koshiba:

In the meantime, Al Mann and his group were preparing TDC’s, and a front-end computer device. It took longer than I anticipated. They brought their own electronics and started installing. There were a number of problems and so forth, but finally everything looked fine at the very end of 1985. From early January '86, we started taking solar neutrino data.

DeVorkin:

Okay. Let's see, I have a question or two here. From Alfred Mann's book, he said that the first measurement for electron neutrinos started in December of '85.

Koshiba:

December of '85?

DeVorkin:

Yes, through '86. That it was sort of a shakedown period, that you found the background was highly variable, that the rates were higher than the predicted signal, and that there were a lot of problems. Now, I'm curious, at this point did you think that you were actually detecting a solar neutrino flux or did you know that this was a spurious background radiation?

Koshiba:

We knew that there was a tremendous amount of spurious background radiation. He said that he started the first operation in early '85, but he doesn't mention that his electronics didn't function and that he had to bring them back to Pennsylvania to have them fixed and bring them back and so forth.

DeVorkin:

Oh really?

Koshiba:

Yes.

DeVorkin:

He said that the high count rate problem tested the relations between the two groups. Quoting, he said, "tested the patience," and "patience with the other group began to wear thin."

Koshiba:

Yes.

DeVorkin:

Could you explain what he means there?

Koshiba:

What happened is this. My people, like Totsuka, became so irritated by the performance of the electronics, that they started arguing.

DeVorkin:

What kind of arguing?

Koshiba:

[laughs] Shouting at each other. Part of my job in these days was to reconcile the Penn people and my people.

DeVorkin:

How did you do that?

Koshiba:

Well, I talked to each of them quietly. [laughs]

DeVorkin:

Alfred Mann gave the impression that some of your people were reluctant to share the experience that they had in operating the detector.

Koshiba:

I don't think that was the case.

DeVorkin:

That was not the issue then. Okay.

Koshiba:

Our people were very open-minded. For instance, Al Mann brought a brand new, fresh Korean graduate student, who worked here for a couple of years and finally produced a Ph.D. thesis. He said he was very happy dealing with our people, he enjoyed everything, and found everyone very open-minded.

DeVorkin:

So the problem really was with the electronics not working right. Was this the cause of the variable background rate, or was that something else?

Koshiba:

Yes, that was part of the reason.

DeVorkin:

What were the other sources and reasons?

Koshiba:

Our understanding of the background came gradually. For instance, when you added new water to the tank, the background increased. We know that this was due to radon.

DeVorkin:

Radon, yes.

Koshiba:

The control of radon is the most difficult part of background control. We learned the hard way. We made the whole system of the detector airtight, so that the environmental atmospheric radon didn't come into the detector.

DeVorkin:

Was this difficult to find out or was this difficult to decide?

Koshiba:

Yes. We knew that radon was giving the trouble, but we brought in a specialist in radon measurement, and gradually controlled the radon contamination.

DeVorkin:

Did the Pennsylvania group and your group have different styles of research that maybe didn't mesh too well? Did you have different ways of solving technical problems?

Koshiba:

I think the main difference probably was due to the fact that Al Mann mostly worked on accelerator experiments. In the case of accelerator experiments, the environmental background is nothing.

DeVorkin:

Yes, because the signal is so huge.

Koshiba:

Yes. But in an experiment like this, the environmental background is the one you have to fight against. And since there was no established procedure to get rid of it, you had to resort to trial and error. That is kind of an irritating process.

DeVorkin:

It can make patience wear thin, as he says.

Koshiba:

Yes.

DeVorkin:

But your people understood this, I take it.

Koshiba:

We have been doing underground experiments for many years.

DeVorkin:

Yes. Was there any feeling among your younger colleagues that you could do the whole project alone, that they resented the people coming in?

Koshiba:

Probably at some time these younger people might have thought that if the boss could get the funding from the Japanese government we would have been much better off.

DeVorkin:

Did you convince them otherwise, or did you just let them think what they wanted?

Koshiba:

As the result of a rather long history of international collaboration experience, I realized the important thing was to understand the other side's feelings, their approach to science, and so forth. So I took time to explain this philosophy not only to my people, but also to Al Mann.

DeVorkin:

That's a very important diplomatic function.

Koshiba:

Well, sometimes one has to be diplomatic.

DeVorkin:

Yes, exactly. I'm not sure where to bring this in, but you know last night you talked about the deep water —

Koshiba:

DUMAND?

DeVorkin:

The DUMAND project. Who asked you to head that? Did that come after this or before this?

Koshiba:

I don't remember exactly, but it was in the very early days of the project DUMAND, which must have been about 20 or more years ago.

DeVorkin:

So it was before your collaboration with Alfred Mann.

Koshiba:

Yes, before Kamioka started.

DeVorkin:

Okay. We should not forget to talk about that later on. But let's stay with Kamioka. You finally were able to reduce the background by the end of 1985. When did observations then actually begin with what you call Kamiokande-II?

Koshiba:

In the beginning of '86.

DeVorkin:

Had you actually found evidence for solar neutrinos or —?

Koshiba:

No. In order to see the signal of the solar neutrino, you had to go through various stages of background reduction, and in order to do that you had to accumulate a certain number of events. As always, the signal was masked by the background. We didn't try to extract the signal when the supernovae neutrino came.

DeVorkin:

I see. So you were in the pure data collecting mode.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

And you were probably going to go for several months yet.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Okay. Well, that brings us to February of 1987 then.

Koshiba:

Yes.

DeVorkin:

And February 23rd. I would like very much for you to walk me through the experience. I know that you had written it out in a review paper, but it would be nice if you could flesh it out for me here, give me sort of the third dimension of what it was like when you received the news, and how you put everybody to action.

Koshiba:

Well, I will try to recollect the situation, but my memory is fading, and probably you had better check with the article I have given you.

DeVorkin:

Right. Certainly.

Koshiba:

The occurrence of a supernovae in the southern sky was first brought to our attention by a Pennsylvania colleague who sent a fax to Gene Beier, I believe.

DeVorkin:

It was a fax from S. Bludman.

Koshiba:

Bludman, yes.

DeVorkin:

Okay. And you received that on February 25th, which was two days after the event.

Koshiba:

The next day a friend of our's, a theoretical astrophysicist, Sato, also brought the news to us.

DeVorkin:

Yes. K. Sato.

Koshiba:

K. Sato. He's a professor at the University of Tokyo now. I immediately contacted the Kamioka shift to send the tapes immediately. As you already know, in those days we recorded the data on tape, and when about 20 tapes had accumulated over a week or so, they were shipped it to the University of Tokyo, where we had a big computer to analyze the data. But of course we couldn't wait until the next shipping, so I ordered the immediate shipping of the existing tapes.

DeVorkin:

I just have a very simple question. When you shipped the tapes, did you make a backup?

Koshiba:

I don't think we did. After all, tapes are not cheap.

DeVorkin:

Yes.

Koshiba:

The tapes arrived on Saturday, I believe.

DeVorkin:

Yes, let's see, it says they arrived Friday afternoon.

Koshiba:

Oh yes, Friday afternoon. There was a girl graduate student, whose name appeared on top of all the Kamioka publications.

DeVorkin:

Okay.

Koshiba:

She was a lucky girl; she became famous for that. We taught her how to analyze the data, and what the signal would look like; she had written a set of criteria for picking up the event. I had to leave Tokyo that weekend, so I was away on Saturday and Sunday. Monday morning when I came back to laboratory, Totsuka brought me this clear neutrino signal.

DeVorkin:

What's it like —?

Koshiba:

That's it, uh-huh.

DeVorkin:

Figure 5.1 in your article on observational neutrino astrophysics?

Koshiba:

Mm-hmm [affirmative]. Wait a minute.

DeVorkin:

It's page 294.

Koshiba:

Yes. This is the signal.

DeVorkin:

Right.

Koshiba:

Since our background stays at this level, this does mean something.

DeVorkin:

I would say so. You have at least 11 points there. And that occurred at, oh, I can't quite read it, but 1743 304.

Koshiba:

That's the time marker.

DeVorkin:

The time marker, yes. These are the times in seconds. Now, I couldn't help but noticing, and you do talk about it here, you have a gain check blank that occurred just a very short time before.

Koshiba:

[laughs] We were lucky, huh? [laughs]

DeVorkin:

That's pretty amazing.

Koshiba:

Almost every day we do a gain check.

DeVorkin:

And that's pure luck, I take it.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

You do describe it in here as something like that.

Koshiba:

Another piece of luck was the following. You see this event occurred on Monday I believe, and if it were a usual Monday our people would have to take the train at around 4:30 in the evening so they would have left the cave.

DeVorkin:

Oh, a usual Monday.

Koshiba:

The event happened at 4 hours and 35 minutes. So if it were a usual Monday, our people would have stopped the operation and changed the tape and then come out to get the train. So the data would have been lost for 35 minutes.

DeVorkin:

If it had been a regular Monday. But this was a hol —

Koshiba:

Yes. Luckily, this was a special Monday in which the people stayed until later. So we had a number of pieces of good luck. I told you I am a lucky man. [laughs]

DeVorkin:

By March 2nd you knew that the signal was genuine. That was the March 1st entry, and that's on page 295 of your record. But then you waited until Monday, March 2nd to call representatives of the collaborating institutions. Was it at that point that you made a public announcement?

Koshiba:

No.

DeVorkin:

Oh. What happened?

Koshiba:

I had written there that the first announcement of a supernovae neutrino observation was made by the Italian group.

DeVorkin:

The Mont Blanc group, yes.

Koshiba:

Yes. Which is hours away from our signal band. I felt that we had to be very, very careful. If by any chance we made a mistake, nobody would believe our results in the future. So I told all the members of my group to make every conceivable check.

DeVorkin:

Oh, I see.

Koshiba:

And I told everybody to keep their mouths shut.

DeVorkin:

Yes. That's exactly what this record says.

Koshiba:

When I felt safe that this was not due to any unlucky background coincidence, or anything of the sort, but was really a genuine signal, I started writing up a paper for the Physical Review Letters; Al Mann and Gene Beier helped me with the English wording. But I wrote the draft myself.

DeVorkin:

I just have to ask a question though. A lot of the others were sending their announcements to the Central Bureau for Astronomical Telegrams at the Smithsonian so that they could be announced very quickly. That's one of the ways the priority in astrophysics is established. But you chose to send yours instead to —

Koshiba:

Physical Review Letters.

DeVorkin:

Yes. And you did not send an Announcement to the Smithsonian.

Koshiba:

Well, after I had sent a paper to the American Physical Society, I contacted

DeVorkin:

Brian Marsden. Is that right?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

So everybody was checking for errors.

Koshiba:

Yes. When I finally sent the manuscript to Physical Review Letters, I told my collaborators, "Now you can talk about it." Of course, Al Mann wanted to make several long distance calls addressed to John Bahcall and other people. I told our collaborators that, "You can announce the content of the abstract." The abstract appears at the top of our paper, and gives the essential parts — signal times, the number of events, and so forth. Now comes a rather touchy point. Just about this time I received a long distance call from one of the IMB collaborators.

DeVorkin:

This was either March 9th or 10th.

Koshiba:

Before that. I cannot tell you exactly what day it was, because my diary was stolen in Rome.

DeVorkin:

What? Ohhh!

Koshiba:

I had a little pocketbook in which I jotted down significant events.

DeVorkin:

A day book sort of. And it was stolen?

Koshiba:

In the subway in Rome.

DeVorkin:

Oh no! That would have been a wonderful thing to have.

Koshiba:

Yes. I missed it too. But what happened was this. The young man from the IMB collaboration wanted to make it out that IMB was, in fact, the first discoverer of a supernovae neutrino, before Kamiokande.

DeVorkin:

And before Mont Blanc as well.

Koshiba:

No, Mont Blanc we don't worry about, because by that time we knew that their claim was spurious.

DeVorkin:

Really?! Oh, I see.

Koshiba:

Mont Blanc.

DeVorkin:

I knew the time didn't agree.

Koshiba:

But they didn't consider the possibility of a correlated background, which I analyzed in our paper and also in that paper.

DeVorkin:

Yes. So that was pretty much discredited completely.

Koshiba:

It was a very long conversation, but finally I became angry. Because Sugawara, head of the theoretical division at KEK, spent two months every year in Hawaii with a good friend in the University of Hawai, an Indian-American theoretical physicist, by the name of Pakvasa. And John Learned of that university belonged to the IMB collaboration.

DeVorkin:

Okay.

Koshiba:

Since KEK is one of the sponsoring institutions, Sugawara knew about our finding the signal and the signal type. Before I could release the news of the abstract of our paper, he telephoned Pakvasa with the signal time and so forth. This was immediately brought to the attention of John Learned, who was a collaborator of IMB. John Learned started searching for the signal, using the Kamioka signal time, and found it in the IMB data.

DeVorkin:

So the IMB people weren't looking for the signal?

Koshiba:

No. They were looking for a signal at the time of the Mont Blanc signal. Of course they couldn't find signal. Okay? However, unless they were given some specific short time period to look for, their detector background level is 20 to 30 MeV, because of the small number of photons they can observe. Okay? Therefore there are lots of background in the 20 to 30 MeV area. And unless you specify this is the time period, you just cannot look all the way.

DeVorkin:

So yours was easy to see because your background rate —

Koshiba:

Was very low.

DeVorkin:

In their case, the signal was there, but it was swamped by the background, so you had to know exactly where to look.

Koshiba:

That's right.

DeVorkin:

So both groups were frantically looking, but you could find it right away, I mean comparatively.

Koshiba:

This I knew because Sugawara later confessed that he leaked the information to the University of Hawaii.

DeVorkin:

But was it really a leak? I mean, he just was being collegial.

Koshiba:

But when the principal investigator of the experiment said to keep quiet until everything was made certain.

DeVorkin:

Yes. How did you feel about that?

Koshiba:

Well, I felt bad. But Sugawara immediately contacted John Learned, and received a copy of John Learned's manuscript to the other members of the IMB collaborators. Learned's manuscript stated that knowing the Kamioka signal time, he searched part of the IMB data and found the signal.

DeVorkin:

Well, that's honest. But who was it who —? What was the nature of the phone call, though, to you? What were they trying to get you to do?

Koshiba:

This man wanted to convince me that IMB found the signal independently of Kamioka, and that happened before Kamioka's discovery of the signal. That's what he tried to convince me.

DeVorkin:

And did you believe him?

Koshiba:

Of course not!

DeVorkin:

But you didn't know at the time that it had leaked, did you?

Koshiba:

Well, I was suspicious. The next day Fred Reines made a long distance call to me, because they were having an IMB group meeting, and this man must have been saying let's claim that we discovered the signal independently, and first, and so forth.

DeVorkin:

You're not telling me who this other person was.

Koshiba:

No. I'm not going to tell you, because that was a dirty trick he played on me.

DeVorkin:

But Fred Reines was not the person, I mean obviously.

Koshiba:

No. When I received this long distance call from the young man at the IMB collaboration, I was furious. I called up Sugawara, and he finally admitted that he did give away the time information, but promised me that if any problems arise, he would get John Learned to testify. Later I received a copy of Learned's internal memo. The next day Fred Reines called me on the phone, while in the middle of an IMB group meeting, and said that he knew this young man had called me on the subject.

DeVorkin:

He was a junior member of their group, sounds like.

Koshiba:

Yes. Fred was quite apologetic. He said he was sorry that this has happened. He said further, "Would you be satisfied if we included the following sentence at the end of our paper?" The sentence went like this, "First they searched for a signal at the time of the Mont Blanc experiment. They didn't find any. Then, knowing the signal time of Kamioka experiment, they searched for a signal and indeed they found one." "Does that satisfy you?" Fred said. Yes, that's fair. I'm happy about that.

DeVorkin:

Yeah. So that was all cleared up.

Koshiba:

Mm-hmm [affirmative]. Fred was very fair.

DeVorkin:

Yes. Well, he already had his Nobel Prize.

Koshiba:

Not yet.

DeVorkin:

Oh no! '95, okay, I'm sorry. Okay.

Koshiba:

No, he was a fair person.

DeVorkin:

That was what actually happened.

Koshiba:

Mm-hmm [affirmative]. Not many people noticed the last couple of lines of their paper. So many American scientists think that IMB and Kamiokande made independent discoveries.

DeVorkin:

That's why. But it's in the literature. I can find it in —

Koshiba:

It's in the original paper. Yes.

DeVorkin:

Okay. That's fine. It makes sense that with your lower background you can find things a lot faster. Did either group have the ability to search at that time automatically, in other words by computer? Or —?

Koshiba:

We didn't have an automatic supernovae watching device. We now have one.

DeVorkin:

What happens now if a signal is —?

Koshiba:

If it occurs say about this scale, then it will be known to the shift person within a minute. It gives a warning.

DeVorkin:

Do red lights start flashing?

Koshiba:

Yes.

DeVorkin:

Do they really?

Koshiba:

Well, not exactly, but it does give a warning.

DeVorkin:

Uh-huh, that something has happened.

Koshiba:

Then the information would be sent to this astronomical union, the what you call it —

DeVorkin:

Oh, the Central Bureau of Astronomical Telegrams. Yes, which is of course one man, Brian Marsden. Right. There was a press release made by the ministry of education?

Koshiba:

Mm-hmm [affirmative]

DeVorkin:

And simultaneously by the University of Pennsylvania?

Koshiba:

Mm-hmm, mm-hmm [affirmative].

DeVorkin:

And you say this was the end of the hectic days.

Koshiba:

Yes. [laughs]

DeVorkin:

There was not a big public response? Didn't people, news and —?

Koshiba:

Well, yes. TV and newspapers wrote about it. First of all, I didn't intend to have a press release, but the collaborators, especially Al Mann, wanted to have one.

DeVorkin:

Oh, I see.

Koshiba:

Originally I thought I would do it in the University of Tokyo. But the minister of education people heard about it, and wanted to have it in their own building.

DeVorkin:

Well, it was a big deal, I imagine. It certainly was in the United States. It was on the 6 o'clock news constantly, what was being revealed, what was new. Now, how did this supernovae event change life around here? Your primary program was solar neutrinos, and proton decay? Is this all correct? So was this discovery disruptive of the normal routine, or did it take a while for you to get back to the normal routine?

Koshiba:

Well, I told all of my collaborators to concentrate on solar neutrino observation. They went back immediately to the solar neutrino analysis.

DeVorkin:

When did you start getting solar neutrino data?

Koshiba:

Let's see. It took more than half a year before we could make the background subtraction and see the sign of the signal.

DeVorkin:

And what did you conclude from that?

Koshiba:

Well of course you see when the statistics are low you cannot say very much, but the first paper on solar neutrinos was published I believe at the end of '87. I don't remember. You better check that with a reference.

DeVorkin:

Yes, I can do that.

Koshiba:

But in the first paper we already knew that indeed the flux was lower than theoretical expectations by a factor of more than two.

DeVorkin:

You found a little more than Davis did.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Yes. But it was still closer to Davis than to Bahcall's theoretical predictions. And has that stayed pretty much the same all this time?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Even with the Super-K?

Koshiba:

Yes. Forty-six percent or something.

DeVorkin:

Mm-hmm [affirmative]. Did you start hiring your own astrophysical theorists here to try to sort out what the differences were and what the problems were?

Koshiba:

No, we didn't hire theoretical people. We were just experimentalists. But we had colleagues in the physics department. K. Sato is a good theoretical physicist.

DeVorkin:

And did you start trying to encourage people like Sato to look at the standard model, this time in terms of the proton-proton reaction?

Koshiba:

K. Sato is more interested in cosmology rather than in the solar model.

DeVorkin:

Has he found any of your data of use for cosmological speculations?

Koshiba:

I don't know, but he did use our result, for instance, in establishing an upper limit on the massive particles decaying inside the sun in relation to the missing mass. He is also very much interested in the latest result on the non-zero mass of the neutrinos.

DeVorkin:

How did that come about? I'd like you to walk me through that observation.

Koshiba:

Well, the first anomaly I noticed was at the time of old Kamiokande. After my retirement at the end of March '87, I was invited to spend four months in Hamburg and one year at CERN. When I came back, I looked at the accumulated Kamiokande data, and I noticed a funny thing, that is, a nu-mu event. The neutrino which produces muons on interaction.

DeVorkin:

Okay.

Koshiba:

A Nu-e event is what the model predicted, but the flux of nu-mu seems to be considerably lower than expected.

DeVorkin:

So you had come back from Hamburg and —?

Koshiba:

No, I had just come back from CERN.

DeVorkin:

Oh, CERN. Okay.

Koshiba:

Back in Tokyo I looked at the accumulated data of Kamiokande and I noticed this, but it is not my type of experiment to compare the absolute flux with the model of prediction, because I don't trust those model predictions. So I took the ratio of the two types of neutrinos, nu-mu flux and over nu-e flux. If you take this ratio, many uncertainties cancel out in the model. One can then compare it with experimentation, which can be derived rather easily and simply. I like to consider the problem in the simple, clear-cut way. How are those neutrinos produced in the atmosphere? The cosmic rays, first of all, have to produce mesons, like pi-mesons or k-mesons. Those particles then decay in thin air into mu + nu-mu. Okay? Decay product muon, when the air is still thin, would have time to decay into an electron + e-nu + nu-mu. Alright, you count the number of neutrinos. In the first decay, you get one nu-mu. Okay? In the second decay, you get one nu-mu and one nu-e. Alright? So originally, nu-e flux and nu-mu flux has to be in the ratio of 1:2. As simple as that. Of course, there is a modification due to a difference in the energy spectrum in the first decay and second decay, but they are minor corrections. And also the contribution of k-mesons makes a minor correction. But this ratio cannot be very much away from 1:2. As I mentioned, the ratio we observed in Kamiokande was 1:1. So there was a factor of 2, which I thought this was real. So I checked the various causes, and had written a draft and circulated it among the Kamioka collaborators, inviting objections and comments, and so forth — especially because one of the graduate students of IMB had just published a paper which showed no anomaly in the ratio of the nu-mu flux to the nu-e atmospheric neutrino flux.

DeVorkin:

They found no anomaly.

Koshiba:

No anomaly, the man claimed. But when I looked at the IMB data and took the ratio, it was again 1:1, not 1:2.

DeVorkin:

Using their data?

Koshiba:

Yes. But this young man, in writing up his thesis, didn't want to make the statement which brings a big argument among the professors who judge his thesis. So he concluded that nothing extraordinary was happening, and that everything was in order, that's what he wanted to show probably. But later I heard that this young man was kicked out of IMB collaboration, and later I indeed did admit that there is a ratio discrepancy which is enormous.

DeVorkin:

I take it you published this result?

Koshiba:

Yes.

DeVorkin:

Yes. That's quite interesting. You said you were retired by now? You retired in '87?

Koshiba:

Mm-hmm [affirmative]

DeVorkin:

And that was right after the supernovae.

Koshiba:

That's right.

DeVorkin:

Did it have anything to do with the supernovae?

Koshiba:

No. It's a compulsory retirement at the age of 60.

DeVorkin:

Sixty?

Koshiba:

Yes. The University of Tokyo has such a rule.

DeVorkin:

Okay. So supernovae or not, you retired.

Koshiba:

That's right. The lucky thing was that it happened one month before my retirement. [laughs]

DeVorkin:

How has life changed for you since you retired?

Koshiba:

Well, a retired Japanese professor's life is quite different from a retired American professor. American professors can keep their office at the university, can keep their secretary, and can also keep receiving research funds. None of these perks are available to a Japanese retired professor.

DeVorkin:

Really? But you have been very active.

Koshiba:

The usual thing for a University of Tokyo retired professor to do, is to find a job in some private university.

DeVorkin:

Ah, that's it.

Koshiba:

I was employed by a private university called Tokai University, and I was given an office and part-time secretarial help. I also took a few graduate students. But the working environment is entirely different from my days in the University of Tokyo. For instance, my research funding was only 300,000 yen.

DeVorkin:

About three thousand dollars?

Koshiba:

Yes.

DeVorkin:

Per year?

Koshiba:

Per year.

DeVorkin:

So that's what the university gave you. It wasn't much.

Koshiba:

No, not at all.

DeVorkin:

Did that include travel money too?

Koshiba:

Yes. Everything. This appointment also came to an end this last March, so I am now a completely free man now — jobless. [laughs]

DeVorkin:

That's what you said, you were free as a bird, when you sent a fax back to me. Yes. But you have been traveling around, visiting different major facilities in Europe I understand, and you had a 2-year appointment in the United States.

Koshiba:

Yes. That was not quite an academic appointment. There is a society, the Japan Society for the Promotion of Science, which is under the jurisdiction of the ministry of education. This organization has a liaison office in Washington, D.C., and I was assigned to that office for two years.

DeVorkin:

Is there a reason? Did you deliberate over whether you wanted to take the job or not? What were the duties?

Koshiba:

Duties? There weren't any, but the general idea was to keep up good relations with the NSF people, the DOE people, and the Academy people. And as I said, I organized the science forum once a year.

DeVorkin:

Right. Mm-hmm [affirmative]. The nature of the forum, you brought in people to speak. What was the character of it?

Koshiba:

Well, I wanted to present the status of Japanese science and technology and I wanted to make the DOE and the NSF people very informed of Japanese trends in these areas. So I invited from Japan very distinguished scholars, like Esaki, Leo. Esaki, who is now the president of Tsukuba University. He got the Nobel Prize many years ago for finding the tunnel effect in semiconductor junctions. — also Oda.

DeVorkin:

Yes, Minora Oda.

Koshiba:

I invited four distinguished scientists from Japan, and as a guest speaker I invited Leon Lederman. You know him?

DeVorkin:

I know the name, yes.

Koshiba:

And also another big name, another Nobel Laureate in biophysics who was the president of Johns Hopkins at that time. His name was Nathans. This year I invited Burt Lichter of SLAC as a guest speaker.

DeVorkin:

At the forum you would then have an audience, you would have an invited audience?

Koshiba:

We invited about a hundred people.

DeVorkin:

Yes.

Koshiba:

And the bioscience speaker was a deputy director of NIH. I don't remember his name.

DeVorkin:

Oh, that's okay. Alright. But I get a sense of what it was like. Let's talk about the Kamiokande part, and of course your plans for Super-Kamiokande.

Koshiba:

I told you how I helped to get the funds.

DeVorkin:

Yes, I would like you to repeat that discussion. That's very important. Now Super-Kamiokande was operational in April of '96, and my interest is to find out what arguments were made for scaling up to 32,000 tons of water —

Koshiba:

The correct figure is 50,000 tons.

DeVorkin:

Fifty thousand tons of water. And how you worked to get the support for it. Because you were retired, were you not?

Koshiba:

[laughs] Yes. I was at that time a guest professor at DESY and the University of Hamburg. As I said, I retired at the end of March that year, and then after attending the international conference in Kyoto at the end of April, I went on to Hamburg. I appointed Totsuka as the spokesperson of the Kamiokande experiment. At that time, the Institute for Cosmic Ray Research was headed by a man called Arafune, a theoretical physicist. He's a good theoretical person. It was a time when the minister of education was seriously thinking of abolishing this cosmic ray institute because —

DeVorkin:

The ICRR.

Koshiba:

Mmm [affirmative].

DeVorkin:

And why was that?

Koshiba:

Because they didn't produce any academic results, and that institution was slated to be abolished. It just happened that Totsuka and Arafune were good friends from student days. I was also objecting. The Kamiokande experiment was sponsored by the ministry of education through the channel of the faculty of science of the University of Tokyo.

DeVorkin:

Not through the ICRR.

Koshiba:

It had nothing to do with ICRR.

DeVorkin:

Right.

Koshiba:

Because I didn't like the way things were done at ICRR. When I was in active service, I didn't want to have anything to do with the ICRR. Totsuka knew that. But after I handed the baton to him he wrote to me in Hamburg to the effect that he was seriously thinking of transferring the project to ICRR.

DeVorkin:

The Kamiokande project.

Koshiba:

Mmm [affirmative].

DeVorkin:

Okay.

Koshiba:

Also he wanted to prepare for Super-Kamiokande through the ICRR. I replied to him, "You know that I didn't want to have anything to do with ICRR. However, now you are the boss of the experiment. If you think it's good for the future of the experiments, go ahead." I didn't make any objection. So it was transferred to ICRR. Next spring, I received a letter from Arafune and Totsuka.

DeVorkin:

This was the spring of '88?

Koshiba:

I think so. They told me that they wanted to push the Super-Kamiokande proposal, but because of the amount involved they worried. They asked me for advice. I wrote back to them, "In dealing with bureaucrats, one has to be very careful, because if you ask some big shot to use his influence over the ministry of education, they would consider it as interference from the outside." I happened to know a number of scientists who won Nobel Prizes. So I wrote to them asking them to write a letter of recommendation for the Super-Kamiokande, and I asked them to address their letters to Arima, to the president of the University of Tokyo, who happened to be my junior in the physics department. About five or six distinguished physicists all over the world did write letters to Arima. Arima immediately made copies of those letters, and forwarded them to the attention of a big shot in the ministry of education. In this way, nobody felt pushed around, you see.

DeVorkin:

But didn't you also have friends in the ministry of finance?

Koshiba:

Yes, I had a good friend.

DeVorkin:

And wasn't there a step there too, wasn't there there?

Koshiba:

Yes, but I didn't have to make contact with my good friend in the ministry of finance at that time.

DeVorkin:

Uh-huh. But you did at one point, didn't you, for something?

Koshiba:

Some time ago.

DeVorkin:

And what was that for?

Koshiba:

Not for the Super-Kamiokande. I didn't have to contact the minister of finance for this project, because my good friend was already out of the ministry of finance. He was at that time the president of the Sakura Bank.

DeVorkin:

Okay.

Koshiba:

After a stint at the Sakura Bank, he became president of the Bank of Japan. Also the top bureaucrat of the ministry of finance was a classmate of mine, but I didn't like him, so I intentionally avoided contacting him.

DeVorkin:

Okay. Let's talk a little broader about the other different detectors and projects that are underway. By the late 1980s there were many new projects on the board. There was the Gran Sasso in Italy, that was a large argon drift chamber?

Koshiba:

Yes.

DeVorkin:

The LVD, the large liquid scintillator; GALLEX, which is the germanium detector; and the Sudbury project, which was a Canada-U.S. sort of a Kamiokande type detector using heavy water, deuterium oxide.

Koshiba:

Heavy water. Yes.

DeVorkin:

With all of these different projects, how were they complementary or competitive to Kamiokande?

Koshiba:

Let's talk about them one after the other. GALLEX is in a way similar to Ray Davis' experiment in the sense that they use the radio chemical method. The difference between the two experiments is that the detection threshold of GALLEX is considerably lower than Chlorine-37 of Ray Davis. Now, the operation of this Gallium-71 experiment is really a good thing, because it is the only experiment which can measure the flux of pp neutrinos, which is the main source of energy. Even though from the neutrino astrophysics point of view it lacks the directional information, energy information, and time information. But the fact that it can observe the flux of pp neutrinos is very important. And it did function as expected; I think it has already been closed, GALLEX. When it comes to SNO [Sudbury] it is a very nice detector if it works. In a sense it's complementary to the Super-Kamiokande. SNO stands for the solar neutrino observatory, and it's located in Sudbury, Ontario. My former collaborator, Gene Beier of Pennsylvania, is one of the main figures there.

DeVorkin:

Yes, you mentioned former collaborators. I know that the University of Pennsylvania-Japan collaboration has ended. Was this a planned ending, or did something happen to cause it to end? They're not here now, right?

Koshiba:

One of the reasons I suspect is that Totsuka and Al Mann didn't get along. That's my guess. I was away in Europe, I don't know exactly what happened.

DeVorkin:

Because Mann doesn't say a thing about it in his book. Okay, but that's all you know at this point.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

What about the fate of the IMB detector? I understand that's closed?

Koshiba:

Yes, it's closed.

DeVorkin:

And why is that?

Koshiba:

I don't know. [laughs]

DeVorkin:

Who should I ask?

Koshiba:

There is a professor at UC-Irvine who is at present one of the Super-Kamiokande collaborators. His name doesn't come to my mind. He is a successor of Fred Reines at UC-Irvine. This man may be able to tell you about the fate of IMB.

DeVorkin:

Okay. I wonder how I would find out his name. Fred Reines is no longer —?

Koshiba:

Fred Reines is still alive, but the actual work is done by this man. He is already a full professor there.

DeVorkin:

We'll try to get his name. The important thing is that the IMB did close, but there are other types of detectors now worldwide. At one point you called for a worldwide network of Super-Kamiokandes. But it seems as though —

Koshiba:

It's not practical.

DeVorkin:

It's what?

Koshiba:

It's rather difficult to form such a network.

DeVorkin:

People seem to be more interested in building different types of detectors.

Koshiba:

That's right.

DeVorkin:

Is there a reason?

Koshiba:

I don't blame them. [laughs]

DeVorkin:

What was the advantage of a worldwide network of Super-Kamiokandes?

Koshiba:

When you have only one Super-Kamiokande, even if a supernova occurred at the center of our Milky Way galaxy, we can use the directional information only from the first few millisecond burst of electron neutrino, which makes an electron scattering, which gives you the directional information. Even so, the accuracy of detection is something like 2 or 3 degrees from the number of expected events.

DeVorkin:

Yes.

Koshiba:

If we had a world network of Super-Kamiokande type detectors, you could use the timing of first burst to make a triangulation, which gives you much better accuracy in directional information.

DeVorkin:

Finding out which star had done it.

Koshiba:

Yes.

DeVorkin:

Assuming that it's not obvious.

Koshiba:

Yes.

DeVorkin:

In the case of 1987 A, you had 14 hours before the optical?

Koshiba:

I don't remember. You better check with the article.

DeVorkin:

It was a good numbers of hours.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Now before we end up with just some general questions, I think a while ago we said we would go back and talk about some of the earlier projects that you were asked to head and that you refused to, thinking that it wasn't a good idea, and one of them was DUMAND. Is that right?

Koshiba:

[laughs] Yes.

DeVorkin:

Could you describe again what you had said last night to me about DUMAND, Deep Underwater Muon —

Koshiba:

Muon and Neutrino Detector. Well, it was well over 20 years ago. One day, Fred Reines called me long distance while I was in my University of Tokyo office. He said that he was now discussing with possible colleagues a new project which would be a very large detector deep underwater. He said he wanted me to join the discussion. I replied, "Fred, I'm sorry, I used up all the foreign travel money this year. I cannot possibly go." The next week I received a telephone call from the ministry of foreign affairs in Tokyo, and was told that they had the travel money in hand. My guess is that Fred worked on the Japanese Embassy in Washington, D.C. so that the minister of foreign affairs got involved. Now that travel funds were available, I went there. It was at the time APS meeting in Utah, and there were about 20 people gathered in a small room, discussing what they could do with this underground detector, how to build it, and so forth. I sat quietly listening to the discussion all around, and my intention was to say goodbye at the end. I wasn't interested in the project.

DeVorkin:

Could you tell me why you weren't interested?

Koshiba:

Because their idea of the DUMAND detector was so crude. How should I describe my feelings? When you want to accomplish something by an experiment you define what you want to identify, what you want to measure. First of all, your detector has to be able to detect, and then measure the characteristics of the particular type of event in your experiment. I got the impression that the people in that room just decided to place a photomultiplier, at say 150 meters apart, to see what would happen. It seems to me that before you decide on a separation of 150 meters between phototubes, you have to specify the type of event you are searching for — whether this 150 meter separation is appropriate or not. It was like a group of amateurs trying to do something different from existing experiments. It's good to be different, but one has also to be able to accomplish something.

DeVorkin:

I see.

Koshiba:

I wasn't impressed at all.

DeVorkin:

I see. You look for specific characteristics in an experiment, where the parameters are well defined.

Koshiba:

Yes, if I were the one who has proposed this big experiment, I would first make a number of simulations to determine the appropriate relative distance of phototubes and what will be the expected event rate, and so forth.

DeVorkin:

Did you do that for Kamiokande?

Koshiba:

Oh yes, of course.

DeVorkin:

This had been your general experimental style. Why do you think they did not do this?

Koshiba:

I don't know. My impression was this. Fred Reines is a great physicist. He always has good ideas. However, he's not an experimentalist. So when he had a good collaborator like Cowan, who didn't have any theoretical knowledge, but was a really good experimentalist. Cowan, mm-hmm. He's the one who did the atomic power neutrino experiment with Fred. At that time Fred had a good experimental collaborator, and he got good results. Unfortunately, since then Fred hasn't had a good experimental collaborator.

DeVorkin:

I see.

Koshiba:

Probably Fred himself didn't trust those 20 people as experimental collaborators, and that was the reason why he wanted me to get involved.

DeVorkin:

But it must have meant something to him when you didn't want to get involved. Did that sort of send a message to him that there was something wrong with —?

Koshiba:

Because I kept quiet all during the discussions which lasted more than an hour, all of a sudden he asked me, "Toshi, you had the experience of directing a very large scale international collaboration experiment."

DeVorkin:

This was the Marcel Schein project. Now I understand.

Koshiba:

Fred said, "Why don't you take, direct this DUMAND project?" I replied, "Fred, I'm sorry, I cannot possibly direct this big project." Fred asked, "Why?" I answered, "In order to bring this big project to a successful conclusion, the director has to be able to pick up the telephone and talk directly to the Queen of England or the President of United States, which I cannot do." Then Fred asked, "Can you think of somebody who could direct this project?" I thought it over for a few minutes and said, "Yes, I can think of one person." "Who is that?" I said, "Professor Blackett." And then I told Fred, "I'm sorry I cannot join this project, but if you need a liaison person in Japan, I would recommend Professor Miyake."

DeVorkin:

Now which Professor Miyake is that?

Koshiba:

Professor S. Miyake, who was the director of ICRR for many years.

DeVorkin:

You were not recommending yourself.

Koshiba:

No.

DeVorkin:

Right. Okay.

Koshiba:

So, I thought I washed my hands of the project. But there is an after story. A few years later I received a letter from the DOE asking me to referee the DUMAND proposal. But the proposed experiment looked to me just the same crude plan as before, so I graded it at the lowest level. Still, the DOE supported it.

DeVorkin:

And why was that do you think?

Koshiba:

I later found out there were two more people in Japan who were asked to give a reference. One was S. Miyake, whom I just mentioned. The other was Jun Nishimura. I mentioned his name in relation with ballooning.

DeVorkin:

Right. And they both gave it good marks?

Koshiba:

Yes.

DeVorkin:

Did you ever have arguments with them later?

Koshiba:

I later talked with them. They said, "Oh, we just wanted to be nice." [laughs]

DeVorkin:

Was this a particularly expensive project?

Koshiba:

DUMAND?

DeVorkin:

Yes.

Koshiba:

If you wanted to do it properly, it was a very, very expensive experiment. Not only that, because of the large pressures involved at depths of 4,000 meters, four hundred atmospheres, there were a lot of technical difficulties involved.

DeVorkin:

What happened to DUMAND? Did it work?

Koshiba:

They did try some experiments. A couple of times they lost the detectors, found them and then re-installed them. Oh, I don't know, it has a rather sad history.

DeVorkin:

Okay.

Koshiba:

John Learned of Hawaii has been involved in it from the very beginning. He can tell you all about it.

DeVorkin:

Fine. If I continue with that.

Koshiba:

You might remember this. One of the original collaborators came from the State of Washington. He was a good friend of the senator of the State of Washington. This man was a little bit queer in the sense that he wanted to propose a neutrino as a signal to communicate with an atomic submarine deep underwater.

DeVorkin:

This is the senator.

Koshiba:

No, the physicist.

DeVorkin:

The physicist!

Koshiba:

Yeah. [laughs] And he proposed a test experiment using the neutrino beam of Fermilab, but the director at that time, Leon Lederman, thought it was a crazy idea and didn't give him machine time. So, [laughs], he went to the senator of the State of Washington, and the senator, knowing nothing about such things, raised hell. And [laughs] Lederman had to allow this man to build a detector of his own just outside the fence of Fermilab. [laughs]

DeVorkin:

That's amazing. I know that, after talking to Professor Suzuki the neutrino beam from KEK is going to be used to test for neutrino oscillations here at Kamiokande. That's something very different. I mean, the impression I get is that you need something like it to make the submarine thing work, I mean how do you modulate a neutrino to get any information out of it?

Koshiba:

When I had published my paper on the atmospheric neutrino anomaly from the old Kamiokande, I made a number of personal guesses which I talked about five years ago at Cal Tech for the first time. One of my guesses was about the neutrino masses. I'm very happy that the new Super-Kamiokande confirmed, with much better statistics, the conclusion of old Kamiokande result on the atmospheric neutrino anomaly: the ratio 1:2 or 1:1. I was for a total of about four years, a member of the research council of KEK. I told Sugawara, who was director general of KEK at the time, "You better quit the job. Go back to physics." [laughs] Just before I left for the United States, about three years ago, I was attending this council, and made the following statement, "Take my advice. Now that we are quite sure that the neutrino has a mass of such-and-such range, Fermilab is now contemplating a long baseline neutrino oscillation experiment. It's called MINOS. "They already have a good high-energy neutrino beam, but they don't have the detector yet. Also, in Europe CERN is seriously considering the neutrino oscillation experiment with the Gran Sasso detector. There again they don't have a good detector. Here in Japan we have an excellent detector, the Super-Kamiokande. The only problem is that we don't have a high enough energy proton accelerator to produce a neutrino beam. So this is my advice. "Increase the energy of the proton accelerator to at least 35 GeV so that the resulting neutrino beam can produce tau-leptons." I made it very strong. Next year, when I was already in the Washington office, Sugawara came and stopped by. He then told me that his institution was now proposing a new proton accelerator which can go up to 50 GeV.

DeVorkin:

Oh, good.

Koshiba:

"We are following your advice." That was what he said.

DeVorkin:

That must have been very gratifying.

Koshiba:

Yes, I was happy.

DeVorkin:

Yes. But would it not have been possible to cooperate with Fermilab, or with CERN, and have them direct the neutrino beam here? Or is the distance too great?

Koshiba:

When I was back from CERN and installed again in Tokyo. I received a letter from Carlo Rubbia, who was the director general of CERN at that time. Carlo said in his letter, "Toshi, we are studying two possibilities. One is to shoot our neutrino beam to Gran Sasso, and the other is to shoot our beam to the Super-Kamiokande. What do you think of these two possibilities?" My reply was this: "By all means concentrate on the possibility of using Gran Sasso. Because, you know, to shoot the neutrino beam to Kamioka, it has to go deep down, which is not easy, because the beam energy is not small. You have to bend vertical down.

DeVorkin:

You mean it isn't a linear beam?

Koshiba:

No. You see, when it is a horizonal bending, there is lots of space. But when you want to bend it deep underground, then not much space is available for bending, unless you dig a deep tunnel.

DeVorkin:

Oh, I see. Yes, as far as the wave guide is concerned.

Koshiba:

Mostly it is, yes. Mostly it's difficult because at this large distance, neutrino beams spread out. So even with a gigantic detector like Super-Kamiokande, the number of events you can detect is very, very small.

DeVorkin:

Yes. Okay. We've gone a good while, and I just have a few more kinds of questions to ask you about particle physics and about observational neutrino astrophysics. In 1996, Bahcall and others, I think in an article that was co-authored possibly by Totsuka, stated that most physicists and astronomers now believe that the observational discrepancies are not — you know, between observed neutrino flux and predicted — are not in the standard model, but in, and I'm quoting, "our overly simplistic view of what neutrinos can do after they have been created." By that he says we need better knowledge of the electro-weak interaction between the neutrino and the environment between the sun and the earth, or neutrino mixing in the sun or in space. Do you basically agree with that?

Koshiba:

Yes.

DeVorkin:

Is it a question that can be experimentally tested? Is neutrino oscillation one of the ways to go to better understand this?

Koshiba:

One of my personal guesses is related to this problem you just asked. Additional neutrino oscillation experimental result of not observing the expected amount of neutrino, this is called "disappearance experiment." You don't observe what you expected.

DeVorkin:

Right.

Koshiba:

Disappearance. The real clarification has to come from what is termed "appearance experiment." Okay? You shoot the neutrino you are sure of consisting entirely of nu-mu. After a certain distance, you detect a neutrino interaction in which tau-lepton is produced. That means during this passage some of the nu-mu were converted into nu-tau to eventually produce tau-lepton. This is called "appearance experiment." You detect something which was not in the initial condition. Yes?

DeVorkin:

Right, yes.

Koshiba:

So this appearance type experiment gives a finer picture of the neutrino oscillation. And that is a experiment I urged KEK to perform by increasing the proton energy so that it can produce neutrino beam of energy high enough to be able to produce a real tau-lepton in Super-Kamiokande.

DeVorkin:

Right. Okay, next question. You already mentioned this at the beginning of the interview, and also in your review paper, that observational neutrino astrophysics began in the years 1987 to 1990, despite Davis' work.

Koshiba:

That statement made John Bahcall very angry.

DeVorkin:

Well, that's what I wanted to know.

Koshiba:

Yes.

DeVorkin:

Yes. The arguments you gave were that you had to have data of astrophysical value, arrival time, direction —

Koshiba:

Time, direction and energy.

DeVorkin:

— and the energy spectrum. Yes. Why was he angry with this statement?

Koshiba:

Oh, he was attached to Ray Davis' experiment I guess.

DeVorkin:

But neutrinos were part of theoretical astrophysics before that time.

Koshiba:

Yes. But I said at the beginning observational neutrino astrophysics. Yes?

DeVorkin:

And did John Bahcall have any good arguments stating that at least Davis had found neutrinos?

Koshiba:

Well, there is a proceeding of an international conference held in Takayama around 1994. I don't remember exact year. That was the time when John Bahcall blew his top, and he made a rather nasty statement about my report.

DeVorkin:

Is it in print?

Koshiba:

Yes, it should be in print. Also, my reply is in print.

DeVorkin:

And where is that again do you think?

Koshiba:

In the proceeding of this international conference on — I don't remember the exact title — but the young people here would know. It was held in Takayama.

DeVorkin:

Oh, in Takayama. Okay, maybe I could actually even get the citation from John Bahcall.

Koshiba:

Yes. [laughs]

DeVorkin:

He was the one that recommended I come and talk to you.

Koshiba:

Oh yes? [laughs] I don't have any hard feeling against him. [laughs]

DeVorkin:

No. I mean, if you make a bold statement like that, you expect that somebody will say something.

Koshiba:

Oh yes. [laughs]

DeVorkin:

Do you know Ray Davis?

Koshiba:

Yes, he is a very nice man. I like him.

DeVorkin:

What did he say about this?

Koshiba:

He doesn't say. He just smiles.

DeVorkin:

Okay. Maybe it means less to him. As I mentioned, I'm building this exhibit, and I want to be able to link the neutrino detector work to the many ways that the universe is observed. We'll have a whole array of detectors from gamma ray and X-ray detectors and neutrino detectors, ultraviolet sensors, things like that. But what would your recommendation be for linking your neutrino work here to anything of cosmological significance? How would you describe the cosmological significance of the neutrino detection in supernovae in 1987-88 for instance? What did it confirm that would be of cosmological significance?

Koshiba:

Well, I have to go back to my personal guesses, again.

DeVorkin:

Sure.

Koshiba:

The third item included in my personal guesses is related to what is called the background neutrino flux, which corresponds to cosmic microwave background of 2.7 degrees Kevin. There also has to be a cosmic neutrino background at a temperature of 1.9 Kelvin degrees. We want to detect it.

DeVorkin:

How could you detect that?

Koshiba:

Good question. I was talking to my last crop, my students who are now associate professors, both of them are my students.

DeVorkin:

This is your last crop, you said? Okay. Again, this is for the transcriber. Now what you're talking about is observing a very low energy neutrino, a red-shifted neutrino.

Koshiba:

Mm-hmm. [affirmative].

DeVorkin:

Okay.

Koshiba:

Which you cannot possibly dream of detecting by the Super-Kamiokande. You have to use an entirely different technique. Yesterday I was talking to some young people. The next big project, you yourself have to conceive. They asked me for my advice. I talked about the possibility of detecting this background neutrino. I have been giving some thought for years to this problem, and I think I could make a mirror which reflects low-energy neutrinos.

DeVorkin:

A mirror.

Koshiba:

Mmm [affirmative]. I wasn't sure about this possibility, so I wrote a letter to Professor Nambu, whom I respect very much as a theoretical physicist.

DeVorkin:

Yes.

Koshiba:

And I asked if this possibility is substantiated theoretically. He replied in one and a half pages stating that this was indeed possible.

DeVorkin:

He thinks it's possible.

Koshiba:

Yes. However, Nambu's way of describing why this is possible was beyond my comprehension. So I asked Professor Takeda to make a more detailed explanation. I also respect Professor Takeda very much. He's a rather quiet type, but certainly a most reliable theoretical physicist. He also gave a proof that this is indeed possible. You can totally reflect neutrinos. This is one step forward. Yes? The next remaining problem is much more difficult: how to detect such low-energy neutrinos. I gave some ideas to these young people, but for this we have to wait for advances, and a very high frequency technology like terra-hertz devices.

DeVorkin:

Terra-hertz?

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Yes, I would say so. But, getting back to the detector that we would like to display in the museum, its significance really is—and I'm sort of looking for your advice here—in displaying it in a gallery that is essentially about cosmology, its significance is in verifying theories of supernovae reactions.

Koshiba:

That's one thing, yes.

DeVorkin:

And the neutrino flux that was measured, as I understand it, did pretty much agree with models for the nature of neutron capture and other processes in the explosion of supernovae. Is that correct?

Koshiba:

Yes. The standard theory of supernovae explosion especially is length of the signal time of about 10 seconds. Our signal time had a spread of 10 seconds. That was rather important in understanding, in the sense that neutrino, even with its small cross-section, because of the extreme high density, it has to diffuse out. Yeah? If it comes straight, the signal time must be very, very short.

DeVorkin:

Right, right. So this is the diffusion, the scatter time, and that verified the physical conditions within the supernovae.

Koshiba:

Yes.

DeVorkin:

So the Irvine contact about the IMB is Hank Sobel?

Koshiba:

Mm-hmm [affirmative]. He's a nice man, very good-natured.

DeVorkin:

I think I've heard the name. Yes. I have a friend who is a historian at Irvine, and I might ask him to help out and contact him for me. So we could add that to the record.

Koshiba:

Another result of cosmological significance stemming from the Kamioka experiment had to do with the remnants of past supernovae explosions. Those first cases must have occurred in the past, and a remnant of those explosions may be detectable. We searched for them, and we set some upper limits. It wasn't very much in the way of information, but it did confirm certain aspects of present cosmology.

DeVorkin:

Alfred Mann talks about the missing mass and how the missing mass in the universe might be in the form of neutrinos. Do you believe in that?

Koshiba:

There is a very popular theoretical physicist whom I believe is the director of the Weizman Institute. His name is Harari. This gentleman publicized a guess that if a tau-neutrino had a mass around 10 electron volts, it could account for the missing mass. And he also added some of his theoretical speculation and so forth, and became an attractive target for some of the experimentalists. And indeed, there are experiments going on at CERN specifically searching for a tau-neutrino of mass around 10 electron volts. Yeah?

DeVorkin:

That is not anything that Kamiokande could detect. No.

Koshiba:

Instead we are setting the mass region much lower.

DeVorkin:

You believe the mass is a lot lower.

Koshiba:

Mm-hmm [affirmative]. From our result.

DeVorkin:

Uh-huh. Because your background is around that range, is it not, 10 electron volts?

Koshiba:

Oh, no. Our expected neutrino mass is less than one tenth of an electron volt.

DeVorkin:

Oh, a tenth of an electron volt. That would still be a cosmologically important finding.

Koshiba:

But that wouldn't account for the missing mass though.

DeVorkin:

No. But that would be an answer at least, even so.

Koshiba:

Mm-hmm [affirmative].

DeVorkin:

Okay. A final question about particle physics. After reading through your introduction and also my general feelings after reading popular literature about particle physics, I am always struck with the complexity of all of the interactions that are suggested. The field is so complex. Doesn't it sometimes seem kind of amazing that it all fits together as well as it does? Does that every bother you — that maybe you're making it fit together a little too much? "You" being the physics community—not you personally. I'm just thinking broadly here.

Koshiba:

Well, it is really amazing thing that people did come to this standard model, which is so good, explains almost everything we observe. But still, there are a number of unsatisfactory aspects to this model. For instance, too many unknown parameters, and also a gravitational force. These interactions are not explained at all. And also the way quarks and leptons are represented in this standard model is very artificial. So, people in the world try to find out where a standard model doesn't work so that you could get a hint of what other type of theory is to be sought for. And proton decay is also one of these things. And neutrino masses another thing, which definitely this present standard model cannot account for. You know, we are very happy about confirming our previous result on atmospheric neutrino anomaly.

DeVorkin:

Okay. Is there anything that we have not covered about your life, your career, something that you feel should be recorded that we have not managed to discuss in these three or four hours?

Koshiba:

No. [laughs] I must say you scrutinized my career thoroughly. [laughs]

DeVorkin:

Well, that's my job. Okay. Well then, I want to thank you very much for this long and exhausting, but very valuable, session. Thank you very much.

Koshiba:

Thank you for your interest.