By exploting the data of three large surveys (WINGS, NFPS and SDSS), we present here a comparative analysis of the Fundamental Plane (FP hereafter) of the early-type galaxies (ETGs) belonging to 59 galaxy clusters in the redshift range $0.04<z<0.07$. We show that the variances of the distributions of the FP coefficients derived for the clusters in our sample are not consistent with the hypothesis of universality of the FP relation. By investigating the origin of such remarkable variances we find that, besides a couple of obvious factors, such as the adopted fitting technique and the method used to measure the photometric and kinematic variables, the coefficients of the FP are strongly influenced by a number of things, mainly related to the distribution of photometric/kinematic properties of galaxies in the particular sample under analysis. For instance, the $a$ coefficient derived for the whole sample of ETGs, turns out to decrease when faint galaxies are progressively removed from the sample, suggesting that bright and faint ETGs have systematically different FPs, likely because of different mechanisms of galaxy formation. In general, by comparing mock cluster samples with the real one, we conclude that the observed variances of the FP coefficients cannot be attributed just to statistical uncertainties. We speculate that the FP is actually a bent surface, which is approximated by different planes when different selection criteria, either chosen or induced by observations, are acting to define galaxies samples. We also find ...