In 1966 there was a Festschrift volume prepared for my teacher, Viki Weisskopf, and as usual I... what's the word... procrastinated and didn't write my piece and didn't know what to write and didn't feel like writing and so on. And finally the volume appeared without anything from me, and I was so ashamed of not having contributed to this volume in honor of Viki that I didn't read it. But in it was an article by Yoichiro Nambu, or Nambu Yoichiro, a very good friend of mine whom I'd known for years and years. He was the... my first visitor to come to Caltech to see me when I first arrived at Caltech. And Yoichiro was pursuing some work that he had done with a man name Han–the Han-Nambu picture. And what they had done was to suggest that there were actually three quarks; three U quarks, three D quarks and three S quarks, and that the charges were not  two-thirds minus a third and minus a third, but the integral charges, which when averaged over this new three-valued variable, would give two- thirds minus a third and minus a third. So from the point of view of later work what they were doing was introducing the color variable with a terrible mistake; namely that there was a color term in the electric current, in the charge operator or the electric current operator. Well, the first was correct, the second was wrong. He then suggested that these particles were real, that they could actually come out and be seen, but that they would be largely confined. They would be difficult to get out because there would be a potential between them, which would strongly favor the states that were actually seen.

By this time people had proposed a very crude SU(6) symmetry scheme which involved both the SU(3) variable and the spin, and this was quite useful for classification purposes. It gave rise to a fifty six representation which included the octet of spin one-half and the decimet of spin three-halves .
And let's see, what else? That's forty-eight- there must have been another octet and a… and a singlet. No, just another octet I guess. No, wait a minute, no that's right, no, no, no… no, no it's correct. No, fifty-six is just those two, just those two. It's an octet of spin a half which is sixteen, and eight times two and then a decimet of spin three-halves which is forty, and forty and sixteen is fifty-six. Sorry I got mixed up. Anyway, they would form a single multiplet under this combined spin and SU(3) group–SU(6). And that was what we needed because that was what we saw.