The work about collective string-like gluon fields, the color ropes [16] , published in 1985, originates in an idea of mine how to describe random color charges of higher SU(3) multiplets. Dr. J.Knoll contributed to the success of this work essentially by drawing the attention to Schwinger's quark production model in classical electric fields and H. B. Nielsen supported us with his expertise in field theory.
Recently (1992 - 1995) a renewed interest in the color rope model can be noted since those microscopic models of relativistic heavy ion collisions which are based on individual strings cannot comply with enhanced strange hadron production observed in CERN SPS heavy ion experiments. The Quark -Gluon String Model [ L.P.Csernai, L.Bravina ] and the Relativistic QuantumMolecular Dynamics [ H.Sorge ] can explain this enhancement by assuming color ropes.
The gaussian effective potential method is a variational approach which treats the ground state wave functional as a gaussian; the reminescent of oscillator ground state wave function. The parameters of this ansatz - in the field theory classical fields and gluon propagators - obey (integro differential) equations derived from minimizing the corresponding thermodynamical potential (Gibbs potential at a given temperature and baryochemical potential) [25, 29, 31, 70] .
Heavy color charges (e.g. quarks) on such a background minimize the energy of chromoelectric field spanned between them. As a result a linear + Coulomb potential energy occurs effectively at zero temperature while a combination of Yukawa and Coulomb terms at overcritical temperatures, where the gluons acquire a dynamical mass in the hot plasma but the gluon condensate vanishes [27, 35, 72] .
A system of heavy quarks can be described by using non-relativistic interquark potentials to a good approximation. Based on a String-Bag Model [23] which unifies the phenomenological confining potential with a color spin - color spin coupling (Ising model like) factor it is possible to consider color non-singlet quark pairs as well. In this case the effects of medium quark gluon plasma are taken into account by mirror charges, so non-singlet color charge clusters automatically carry an energy diverging with the system size (this is the "bag" part).
With W. Cassing we studied in this model multi-quark systems, especially six quarks, in canonical equilibrium at a given temperature by calculating the partition sums for randomly located charges. The system size was inversely proportional to the temperature. An equilibrium distance between the quarks gathering into a singlet state can be observed at low temperature, while above a critical value all interquark distances are the same and scale with the (inverse) temperature.
While for light quarks only hadrons with a minimum number of valence quarks (2 or 3) can exist, the idea occurs that exotic multi quark systems consisting of heavy (charm, bottom or top) quarks may be produced in relativistic heavy ion collisions. This possibility depends on phenomenological parameters which may be different in these physical processes. Together with P. Lévai and J. Zimányi we concluded that unfortunately the parameter window allowing such objects to exist is rather tiny [77, 115] .