I think the best thing I can do, now that we've found this paper is to read from it. There I explained what the… what the problems that we were worried about were:

Although enthusiastic about the beauty of this theory we hesitated a bit in endorsing it in print for three reasons. One: we were worried about how to generate a non-zero trace of the stress energy momentum tensor in the limit of zero-quark masses. We knew that such a non-zero trace was needed. The mass of a nucleon, unlike the masses of the lower pseudo-scalar mesons had to be non-vanishing in that limit, and scalar invariants had to be broken. Somewhere there was a source of mass that would hold up as quark masses vanished. Even without the explicit dimensional transmutation later demonstrated by Coleman and Erick Weinberg, it was easy to show that in such a theory the trace could be non-vanishing in the limit. John Ellis had been a visitor at Caltech during 1969 to ’70 and he had lectured there on the possible generation of an anomalous trace, yielding what he called—appropriately for that era--'Pot': P - 0 - T, partially zero trace. If I had remembered his work I would not have troubled myself about the generation of mass from no mass. Two: we understood that some form of string theory in terms of which the Veneziano model had just been reformulated and was the embodiment of the bootstrap. And in those days of course the bootstrap idea was thought to apply to hadrons alone, rather than to all the elementary particles. Thus we thought at times that perhaps the Yang-Mills field theory of colored quarks and gluons ought to be replaced by some kind of related string theory. Of course it does turn out that QCD structures, like bags and--when they are elongated--strings, do occur, but they're approximate and they're derived features of the theory, not fundamental ones. Three: (this was what I was discussing a moment ago) we didn't understand what was causing the suppression of color non-singlets, the confinement of color or the mathematical character of quarks and gluons. We didn't know that it would follow from the color SU(3) Yang-Mills theory itself. No, that, sorry, that's not what I was discussing a moment ago, no. What I was discussing a moment ago comes later. I… I was mistaken about that. These were the three, these were the three reasons. The fourth problem was something we took up a bit later.