Summary Recent molecular sequencing results involving multiple genes require evaluation in the light of pre-existing morphological data, particularly as different methodologies and genes produce trees that are incongruent in some respects or have major areas with poorly supported branch resolution. The present paper summarizes the current situation, primarily from a morphologist's perspective. Most of the tabulation-based groups are coherent in small subunit (SSU) and large subunit (LSU) trees; but some, notably the prorocentroids and peridinioids, are not. In prorocentroids this is primarily because of intrinsic inadequacies of the molecules to resolve their phylogeny. In peridinioids it seems to be because of paraphyly of the group. Other artefacts are noted, such as the drastically different positions of Oxyrrhis in phylogenetic trees based on SSU and protein genes, and of Noctiluca in SSU trees that include analyses with different numbers of nucleotides. Polyphyly in non-tabulate or poorly known groups has been confirmed, as has been the presence of cryptic thecae in members of those groups (group misattribution). Whether or not some extant groups of athecate, wholly dinokaryotic forms originated prior to polytabulate groups, like the suessioids, peridinioids and gonyaulacoids, remains unclear. Gymnodinioids with a spiral acrobase seem to have given rise to the more complex athecate forms, whereas morphological features of the genus Gymnodinium are consistent with it being a sister group to polytabulate taxa such as Woloszynskia and the suessioids. Peridinioids and gonyaulacoids appear to have originated after that split. Dinophysoid and prorocentroid dinoflagellates appear to be derived from peridinioid forms. Trees based on protein genes, such as actin or α- and β-tubulin, may help resolve some of the positions of key groups, but they do not include enough taxa to be widely useful as yet.