We propose possible ways of explaining the net charge event-by-event fluctuations in Au+Au collisions at the Relativistic Heavy Ion Collider within a quark recombination model. We discuss various methods of estimating the number of quarks at recombination and their implications for the predicted net charge fluctuations. We also discuss the possibility of diquark and quark-antiquark clustering above the deconfinement temperature.Fluctuations of the net electric charge of all particles emitted into a specified rapidity window have been proposed as a possible signal for the formation of deconfined quark matter in relativistic heavy ion collisions [1,2]. The argument at the basis of this proposal is that charge fluctuations in a quark-gluon plasma are expected to be significantly smaller (by a factor 3−4) than in a hadronic gas. Because the net charge contained in a given volume is locally conserved and can only be changed by particle diffusion, thermal fluctuations generated within the deconfined phase could survive hadronization and final state interactions. Quantitative estimates of the diffusion of net charge showed that the survival of these fluctuations from an early stage of the collision requires a moderately large rapidity window [3].The most widely used measure for the entropy normalized net charge fluctuations is the D measure [2]:where (∆Q) 2 denotes the event-by-event net charge fluctuation within a given rapidity window ∆y, and N ch is the total number of charged particles emitted in this window. For a free plasma of quarks and gluons D ≈ 1, while for a free pion gas D ≈ 4. For the comparison with experimental data a number of corrections for acceptance and global charge conservation must be applied to the expression for D [4]. The relation of the D-measure to other measures of net charge fluctuations has been discussed by various authors [5,6,7].Several experiments have measured net charge fluctuations in heavy ion collisions at the CERN Super-Proton Synchrotron (SPS) and at the Relativistic Heavy Ion Collider (RHIC) in Brookhaven [8,9,10,11]. The results for D are generally somewhat smaller than 4, but much larger than the value predicted for a free quark-gluon gas. For example, the STAR collaboration has measured D = 2.8 ± 0.05 in central Au+Au collisions at √ s N N = 130 GeV [8], before applying corrections for global charge conservation and other effects [4]. The PHENIX experiment measured net charge fluctuations in a limited azimuthal acceptance window around midrapidity, which extrapolate to a value D ≈ 3 [9]. These results are surprising, because many other observables indicate that a deconfined quark-gluon plasma is formed in these collisions.Bia las has argued that the measured values of D could be compatible with the net charge fluctuations in a deconfined quark phase, if hadronization proceeds according to simple valence quark counting rules [12] and if gluons do not play an active role in the hadronization. Indeed, hadron abundances measured in relativistic heavy ion collisions at the SPS ...