What are the early degrees of freedom in ultra-relativistic nucleus-nucleus collisions?
The Parton-Hadron-String-Dynamics (PHSD) transport model is used to study the impact on the choice of initial degrees of freedom on the final hadronic and electromagnetic observables in Au+Au collisions at $\sqrt{s_{NN}}$ = 200 GeV. We find that a non-perturbative system of massive gluons (scenario I) and a system dominated by quarks and antiquarks (scenario II) lead to different hadronic observables when imposing the same initial energy-momentum tensor $T_{μν}(x)$ just after the passage of the impinging nuclei. In case of the gluonic initial condition the formation of $s,{\bar s}$ pairs in the QGP proceeds rather slow such that the anti-strange quarks and accordingly the $K^+$ mesons do not achieve chemical equilibrium even in central Au+Au collisions at $\sqrt{s_{NN}}$ = 200 GeV. Accordingly, the $K^+$ rapidity distribution is suppressed in the gluonic scenario and in conflict with the data from the BRAHMS Collaboration. The proton and antiproton rapidity distributions also disfavor the scenario I. Furthermore, a clear suppression of direct photon and dilepton production is found for the pure gluonic initial conditions which is not so clearly seen in the present photon and dilepton spectra from Au+Au collisions at $\sqrt{s_{NN}}$ = 200 GeV due to a large contribution from other channels. It is argued that dilepton spectra in the invariant mass range 1.2 GeV $< M <$ 3 GeV will provide a definitive answer once the background from correlated $D$-meson decays is subtracted experimentally.