We gathered data from 40 allozyme loci and DNA-sequence information from a stretch of a rapidly evolving protein coding mitochondrial gene (ND4) to reconstruct the evolutionary history of the T. cristatus superspecies, employing a cross-section of phylogenetic methodologies (MP, NJ, ML and Bayesian). However, neither the allozyme nor the mtDNA data resolve the phylogeny of crested newts at the species level. The preferred topologies are diverse and have in common that internal branches at the species level gain low statistical support (Table 2, Fig. 3). The absence of phylogenetic resolution can not be attributed to either too high or too low a level of genetic differentiation at the average enzyme locus (see Table 1 and genetic distance information in Table 2). Also, homoplasy in the DNA-sequences is not a major issue (see results). Both data sets contradict the NRBV-based phylogenetic hypothesis and because of the correlation between NRBV character-states and the ecology of the species (in terms of the general slender-stout morphological series and the length of the annual aquatic phase) we suggest that NRBV is under natural selection and, hence, not an informative character for building a phylogeny. In the absence of phylogenetic resolution, we suggest that the polytomy recovered in the crested newts is ‘hard’ and that the suggested simultaneous origin of four lineages reflects reality (Whitfield and Lockhart, 2007). However, ‘absence of proof’ shall not be confused with ‘proof of absence’. In other words, this hard polytomy constitutes a hypothesis that may be falsified, e.g., through the sequencing of the full mitochondrial DNA genome (c. 16,000 bp) and the sequencing of nuclear genes.