Baskin & Laor 2005 claimed that the density of the NLR clouds that emit the [OIII] 5007 and 4363 lines is of order of log n_e ~ 5.85 cm-3. It is a brilliant use of the [OIII] data but with which I disagree. I consider that Bakin & Laor might instead have given us the first convincing proof that NLR plasmas suffer from the same temperature problem as that found in HII regions (the t2 problem). I will be presenting models of the photoionization code MAPPINGS that is designed to include the effects of temperatura fluctuations t2 or of a non-boltzman-maxwell distribution of electron energies (parameter kappa). The problem with the proposed high densities for the [OIII] emitting region is that other optical lines would be quenched due to collisional deexcitation ([NII], [SIII]...), contradicting what we see.