Charge screening, large-N, and the abelian projection model of confinement

We point out that the abelian projection theory of quark confinement is in conflict with certain large-N predictions. According to both large-N and lattice strong-coupling arguments, the perimeter law behavior of adjoint Wilson loops at large scales is due to charge screening, and is suppressed relative to the area term by a factor of 1/N2. In the abelian projection theory, however, the perimeter law is due to the fact that N - 1 out of N2 - 1 adjoint quark degrees of freedom are (abelian) neutral and unconfined; the suppression factor relative to the area law is thus only 1/N. We study numerically the behavior of Wilson loops and Polyakov lines with insertions of (abelian) charge projection operators, in maximal abelian gauge. It appears from our data that the forces between abelian charged, and abelian neutral adjoint quarks are not significantly different. We also show via the lattice strong-coupling expansion that, at least at strong couplings, QCD flux tubes attract one another, whereas vortices in type-II superconductors repel. (orig.)