Abstract The relationship between the five echinoderm classes has perplexed phylogeneticists for some time. Although each of the crinoids (C), asteroids (A), ophiuroids (O), echinoids (E) and holothuroids (H) are morphologically distinct, evidence from larval and adult morphology, molecular biology, and stratigraphy have failed to provide a single consensus solution. We have reviewed all available morphological and molecular data, added new data and reanalysed independent data sets individually and in combination, in order to resolve echinoderm class relationships. In total, we present 21 larval and 50 adult morphological characters, partial 28S-like large subunit rRNA gene data for 39 taxa and complete 18S-like small subunit rRNA gene data for 37 taxa. For a 5 taxon problem there are 105 possible rooted tree topologies, and yet we were consistently presented with three competing hypotheses when data sets were analysed both individually and in combination. The total evidence solution favoured (outgroup(C(A(O(E,H))))) although the alternative tree topology, (outgroup(C(O(A(E,H))))) was only one step longer and (outgroup(C((A,O),(E,H)))) was two steps longer. Only these three trees are serious contenders and the distribution of morphological characters suggests we should discount the solution placing ophiuroids as sister group to an asteroid+echinoid+holothurian clade. Thus we are left with (outgroup(C(A(O(E,H))))) and (outgroup(C((A,O),(E,H)))) as the two most plausible phylogenetic hypotheses. Our data showed high levels of phylogenetic signal and these trees best fit the available data.